ProgrammableMarketing/nianticspatial.com (opens in a new tab)
PDF

Views

  • Plan
  • Roadmap
  • Research

On this page

  • The brief
  • Start here
  • Idea portfolio
  • Effort vs impact
  • This week · 6
  • This quarter · 10
  • Big bets · 3
  • Wildcards · 3
  • How we can help
Generated Oct 23, 2026
  1. Programmable
  2. Playbooks
  3. Niantic Spatial
Niantic Spatial logo

growth playbook · Oct 23, 2026

nianticspatial.com (opens in a new tab)

Niantic Spatial growth playbook

Diagnosis

Niantic Spatial has unusual assets for embodied AI: 30 billion posed images, a real-to-sim pipeline with a measured transfer result, a 100-place library and a popular capture app.

Its hiring shows the next step: a product marketer for pricing, packaging and product-led growth, a GTM lead and founding SDR for robotics pipeline, a product engineer for the developer experience, and a test lab to produce evidence. The site has not caught up.

The "self-serve" Evaluate plan has no price, Pricing shows only Scaniverse, most of the library is hidden, there is no Isaac Sim quickstart, and no customer story involves a robot. This playbook opens a priced Evaluate path, publishes the library and the evidence, and builds pipeline from samples, signals, partners and the Scaniverse base.

Growth Score

7/10

Biggest Opportunity

Let robotics engineers buy and use Evaluate the same day, with the full library visible and a quickstart that gets a scene into Isaac Sim in an hour.

Biggest Risk

Synthetic data and generated world models are improving fast and are easy to try, so a sales-only path can lose engineers before they see the real-to-sim result.

01

Start here

The three moves to make first. Each one is written up in full below.

  1. 1

    Put a price and a checkout on Evaluate, the plan the site already calls "self-serve"

    Agree an Evaluate price with the General Manager and show it on the Plans page with a "Buy Evaluate" button that, for now, opens a short card form.

    High impact First result in 1 month
  2. 2

    An Isaac Sim and Isaac Lab quickstart for Places Library scenes in the docs

    Record the steps one engineer takes to load the Leake Street sample in Isaac Sim.

    High impact First result in 2 weeks
  3. 3

    Follow up every sample download within a day, opening with the scene they chose

    Pull last month's sample form fills from HubSpot and write the first follow-up email.

    High impact First result in 2 weeks

02

Your idea portfolio

22 ideas, grouped by when to run them. Open any idea for how to run it: the steps, who to approach, copy you can use, and how to measure it.

  • 8 ideas are marked Unexpected: plays you are unlikely to have tried.

Every idea placed by the effort it takes against the impact it could have. Quick wins sit top left; select a number to open that idea. The filters below narrow the list by status, timing, category and what Programmable can build.

Effort vs impactEach number is an idea. Select one to open it.
Impact →
Quick wins · 2Big bets · 12Fill-ins · 4Reconsider · 4
Effort →

Showing 22 of 22 ideas

This week 6

Quick wins with a clear first action.

  • 01

    The Plans page's own description promises "self-serve Evaluate", and the Embodied AI page says a deployment environment "starts at a few hundred dollars". But no plan shows a price, and all four, including the three-month Evaluate trial, end in the same "Connect with our team" form. Meanwhile Scaniverse on /pricing already sells Plus at $20 and Pro at $50 a month with "Get Started for Free" and credit top-ups, so the billing pattern exists. A robotics engineer who wants to test one warehouse scene in Isaac Sim this week has to wait for a sales conversation first.

    High impactMedium effort First result in 1 month Senior Product Marketing Mana…

    First action

    Agree an Evaluate price with the General Manager and show it on the Plans page with a "Buy Evaluate" button that, for now, opens a short card form.

    How to measure it

    Self-serve Evaluate starts

    Now
    –
    Target
    15 a month
    By
    90 days

    Steps

    1. 1

      Set the price

      One Evaluate price, plus "from" prices for Train and Deploy.

      General Manager, Embodied AI 3 days

    2. 2

      Reuse billing

      Evaluate as a product in the Scaniverse checkout and credit system.

      Senior/Staff Product Engineer, Embodied AI 1 week

    3. 3

      Automate delivery

      Bucket, training rights and the quickstart sent on purchase.

      Senior/Staff Product Engineer, Embodied AI 1 week

    4. 4

      Update the page

      Prices, a "Buy Evaluate" button and "Talk to us" for larger plans.

      Web team 2 days

    5. 5

      Alert sales

      HubSpot alert to the GTM lead and SDR for every purchase.

      Sales Development Representative, GTM 1 day

    6. 6

      Check in at day 30

      A call offering help and the step up to Train.

      GTM, Embodied AI ongoing

    7. 7

      Measure

      Evaluate starts, downloads used and Evaluate-to-Train conversion.

      Senior Product Marketing Manager, Embodied AI monthly

    First result
    1 month
    Cost
    Engineering time on billing and delivery
    Owner
    Senior Product Marketing Manager, Embodied AI

    Smallest test

    Add the Evaluate price to the Plans page and count clicks on "Buy Evaluate" for two weeks.

    Why it works

    Publish a price for Evaluate (one quarter, 15 library downloads, 15 capture minutes a month) and let a team buy it by card, reusing the Scaniverse billing and credit model. Keep Train, Deploy and Enterprise on "Talk to us", but show "from" prices so buyers can size the jump. On purchase, deliver the GCS bucket and a short Isaac Sim quickstart automatically, and alert the GTM lead and SDR with the company name so a human follows up while the team is already training. Track how many Evaluate buyers move to Train.

    Developer-led buyers in robotics expect to try before they talk, and the PMM role is hired to "shape pricing, packaging, and initial product-led growth motions". A paid self-serve trial filters for real intent, gives the founding SDR warm accounts instead of cold ones, and turns the site's "few hundred dollars" claim into something a buyer can act on today.

    Ready to use

    ChecklistSelf-serve Evaluate launch list

    [ ] Evaluate price on /embodied-ai/plans
    [ ] "from" prices for Train and Deploy
    [ ] Card checkout reusing Scaniverse billing
    [ ] Bucket and training rights delivered automatically
    [ ] Isaac Sim quickstart in the welcome email
    [ ] HubSpot alert to GTM and SDR on purchase
    [ ] Link from /pricing to Embodied AI plans
    [ ] Evaluate-to-Train report

    Risks and how to handle them

    • Underpricing anchors larger deals

      Price Evaluate as a paid proof, and show that Train lowers the cost per place.

    • Buyers who need a contract

      Keep the form for teams that need procurement terms.

    Evidence

    • Your websiteEmbodied AI Plans page (checked October 22, 2026): the meta description says "from self-serve Evaluate to Enterprise", but no plan shows a price. Evaluate ("3-month minimum", "15 Places Library downloads across the term", "15 capture minutes a month (~1 place)"), Train ("Most teams start here", 12-month minimum), Deploy and Enterprise all end in the same "Select plan" form: "Connect with our team to discuss the shape of your plan". The FAQ calls Evaluate "a paid three-month plan".Source
    • Your websiteEmbodied AI page: "Make real places simulatable"; "A deployment environment starts at a few hundred dollars vs. tens of thousands to professionally survey or hand-build synthetic twins"; "hours, not weeks"; delivered as "one USDZ per place" (Gaussian splat plus aligned collision mesh) that "Runs in: Isaac Sim · Isaac Lab · any OpenUSD-compatible simulator"; upload is "Drag and drop into a secure GCS bucket. Enterprise APIs coming soon." In a comparison with Flexion and NVIDIA, an RGB policy trained on the reconstruction cut the fall rate from 27.1% (synthetic office) to 2.1%; the livestream sits behind a form. Roadmap items "Vary" and "Deploy and feed the loop" invite teams to "Tell Us What You Need".Source
    • Your websiteThe /pricing page covers Scaniverse only: Free ($0, 20,000 monthly credits), Plus ($20 a month, "360° camera Gaussian splats and USDZ export"), Pro ($50 a month, commercial rights) and Enterprise, with "Get Started for Free", monthly/annual billing and credit top-ups; it has no link to the Embodied AI plans. A separate /plan-comparison-calculator recommends a Scaniverse plan only.Source
    • HiringSenior Product Marketing Manager, Embodied AI (posted September 22, 2026): "Define the Market Narrative" for embodied AI across "our website, campaigns, sales materials, and partner communications"; "Shape pricing, packaging, and initial product-led growth motions, turning technical capabilities into clear offerings and activation paths"; "Build the target-account list for robotics and physical AI"; "Run competitive intelligence"; build "co-marketing and case-study partnerships with robotics OEMs, simulation ecosystems, and strategic customers"; and "Ensure every message clearly distinguishes what ships today, what is being validated, and what is on the roadmap".Source
    • HiringSenior/Staff Product Engineer, Embodied AI (posted September 30, 2026): will "Own how autonomy teams experience our work: the APIs, UIs, packages, and documentation" and "Turn scans and reconstructions into reliable, interoperable assets that work in customers' simulation tools"; "a capability isn't a product until someone outside the building has used it without help". Nice to have: Isaac Sim, Isaac Lab, MuJoCo or OpenUSD. A Staff Engineer, Robot Learning (posted October 2) will "lead engineering for our real-to-sim product".Source
    • HiringSales Development Representative, GTM (posted September 14, 2026): the "founding Sales Development Representative", "Reporting into Marketing", will "Respond quickly to inbound interest", "Develop outbound campaigns with Marketing and Business Development, testing audiences, messages, channels, and calls to action", research accounts "across industries such as oil and gas, defense and the public sector, and embodied AI", and "Capture patterns from prospect conversations: recurring pain points, objections, terminology, competitive context". The CRM is HubSpot.Source
  • 02

    The Embodied AI page says each USDZ "Runs in: Isaac Sim · Isaac Lab", but none of the 2,547 docs pages explains how to import a Places Library scene, check its scale and collision mesh, or start a training run. The new product engineer is hired to own "the APIs, UIs, packages, and documentation".

    High impactLow effort First result in 2 weeks Senior/Staff Product Engineer

    First action

    Record the steps one engineer takes to load the Leake Street sample in Isaac Sim.

    How to measure it

    Sample-to-inquiry rate

    Now
    –
    Target
    +30%
    By
    60 days

    Steps

    1. 1

      Record the path

      Steps and issues loading the free sample.

      Senior/Staff Product Engineer, Embodied AI 2 days

    2. 2

      Write the quickstart

      Import, scale check, colliders and a first run.

      Senior/Staff Product Engineer, Embodied AI 1 week

    3. 3

      Add sample code

      A small Isaac Lab example in a public repository.

      Staff Engineer - Robot Learning, Embodied AI 1 week

    4. 4

      Link it

      From the samples, Plans page and welcome email.

      Web team 1 day

    5. 5

      Collect questions

      From support and the developer forum.

      Sales Development Representative, GTM ongoing

    6. 6

      Update

      With each release.

      Senior/Staff Product Engineer, Embodied AI monthly

    First result
    2 weeks
    Cost
    Engineering time
    Owner
    Senior/Staff Product Engineer, Embodied AI

    Smallest test

    Publish a one-page import guide for the two free samples.

    Why it works

    Write a quickstart: download the free sample, open it in Isaac Sim, verify metric scale and colliders, run a sample navigation policy in Isaac Lab, and common fixes. Add a version for MuJoCo or other OpenUSD tools if supported. Link it from the sample download, the Plans page and the welcome email.

    Engineers judge a data product by their first hour with it. A working quickstart turns sample downloads into real evaluations, and docs pages rank for the exact searches robotics teams make.

    Risks and how to handle them

    • Simulator versions change

      State the tested version and update monthly.

    Evidence

    • Your websiteEmbodied AI page: "Make real places simulatable"; "A deployment environment starts at a few hundred dollars vs. tens of thousands to professionally survey or hand-build synthetic twins"; "hours, not weeks"; delivered as "one USDZ per place" (Gaussian splat plus aligned collision mesh) that "Runs in: Isaac Sim · Isaac Lab · any OpenUSD-compatible simulator"; upload is "Drag and drop into a secure GCS bucket. Enterprise APIs coming soon." In a comparison with Flexion and NVIDIA, an RGB policy trained on the reconstruction cut the fall rate from 27.1% (synthetic office) to 2.1%; the livestream sits behind a form. Roadmap items "Vary" and "Deploy and feed the loop" invite teams to "Tell Us What You Need".Source
    • Your websiteDeveloper docs (/docs, 2,547 pages in the docs sitemap): the NSDK for Unity, Swift and Kotlin and a few Scaniverse guides; no page covers importing a Places Library USDZ into Isaac Sim or Isaac Lab, or the embodied AI pipeline.Source
    • HiringSenior/Staff Product Engineer, Embodied AI (posted September 30, 2026): will "Own how autonomy teams experience our work: the APIs, UIs, packages, and documentation" and "Turn scans and reconstructions into reliable, interoperable assets that work in customers' simulation tools"; "a capability isn't a product until someone outside the building has used it without help". Nice to have: Isaac Sim, Isaac Lab, MuJoCo or OpenUSD. A Staff Engineer, Robot Learning (posted October 2) will "lead engineering for our real-to-sim product".Source
    • Your websitePlaces Library page: "a collection of 100 off-the-shelf warehouse, residential, and outdoor environments", but the page lists 16 places, with filters only for "Last Mile" and "Warehouse"; two are "Free Sample" downloads, the rest say "Select Plan to Access", and "+84 more places in the library; reach out for access!". A "Request A Place" vote (Warehouse, Factory, Refinery, Data center, Construction, Natural, Urban, Residential) shows 0 on every option. The launch post says the catalog spans "industrial environments, last-mile and logistics settings, commercial spaces, and residential environments".Source
  • 03

    The free samples and the "7 more scenes" behind the USDZ form are the clearest sign that a robotics team is testing. The founding SDR is hired to "Respond quickly to inbound interest", and HubSpot is already the CRM.

    High impactLow effort First result in 2 weeks Sales Development Representat…

    First action

    Pull last month's sample form fills from HubSpot and write the first follow-up email.

    How to measure it

    Sample-to-meeting rate

    Now
    –
    Target
    20%
    By
    60 days

    Steps

    1. 1

      Pull the list

      Last 60 days of sample forms.

      Sales Development Representative, GTM 1 day

    2. 2

      Write templates

      One per scene type.

      Sales Development Representative, GTM 1 day

    3. 3

      Set routing

      Same-day HubSpot tasks by scene.

      Sales Development Representative, GTM 1 day

    4. 4

      Send

      Personal notes, not sequences.

      Sales Development Representative, GTM daily

    5. 5

      Log use cases

      Robot type, sites and simulator.

      Sales Development Representative, GTM ongoing

    6. 6

      Share patterns

      Weekly notes to PMM and GTM.

      Sales Development Representative, GTM weekly

    First result
    2 weeks
    Cost
    $0
    Owner
    Sales Development Representative, GTM

    Smallest test

    Email last month's ten company-domain downloads.

    Why it works

    Route every sample form to the SDR the same day with the scene downloaded. Open with a useful note: how to load it, what the matching paid places are (a warehouse download gets the warehouse list), and an offer to scan one of their own sites on Evaluate. Log the use case each team describes.

    A sample download plus a fast, specific follow-up is the warmest signal the site produces, and the notes feed the PMM's target-account list.

    Risks and how to handle them

    • Students and hobbyists

      Answer helpfully, prioritize company domains.

    Evidence

    • Your websiteSitemap (checked October 22, 2026): three customer stories (Sphere, Forgers Rittai, The Running Man), none about robots; 39 /campaigns landing pages (including Gartner reports, "physical-ai-readiness-checklist", "simulation-vs-digital-twin" and "agentic-ai-vs-physical-ai"); no comparison pages against other approaches or vendors; a /forms/usdz-samples page where "7 more scenes unlock with the form"; a public map-free relocalization benchmark with a leaderboard under /research/map-free; blog posts on Coco Robotics and the Flexion humanoid real-to-sim work.Source
    • Your websitePlaces Library page: "a collection of 100 off-the-shelf warehouse, residential, and outdoor environments", but the page lists 16 places, with filters only for "Last Mile" and "Warehouse"; two are "Free Sample" downloads, the rest say "Select Plan to Access", and "+84 more places in the library; reach out for access!". A "Request A Place" vote (Warehouse, Factory, Refinery, Data center, Construction, Natural, Urban, Residential) shows 0 on every option. The launch post says the catalog spans "industrial environments, last-mile and logistics settings, commercial spaces, and residential environments".Source
    • HiringSales Development Representative, GTM (posted September 14, 2026): the "founding Sales Development Representative", "Reporting into Marketing", will "Respond quickly to inbound interest", "Develop outbound campaigns with Marketing and Business Development, testing audiences, messages, channels, and calls to action", research accounts "across industries such as oil and gas, defense and the public sector, and embodied AI", and "Capture patterns from prospect conversations: recurring pain points, objections, terminology, competitive context". The CRM is HubSpot.Source
    • HiringSenior Product Marketing Manager, Embodied AI (posted September 22, 2026): "Define the Market Narrative" for embodied AI across "our website, campaigns, sales materials, and partner communications"; "Shape pricing, packaging, and initial product-led growth motions, turning technical capabilities into clear offerings and activation paths"; "Build the target-account list for robotics and physical AI"; "Run competitive intelligence"; build "co-marketing and case-study partnerships with robotics OEMs, simulation ecosystems, and strategic customers"; and "Ensure every message clearly distinguishes what ships today, what is being validated, and what is on the roadmap".Source
  • 04

    The main Pricing link in the menu leads to Scaniverse plans only and never mentions the Embodied AI plans, and the plan calculator recommends Scaniverse plans only. A robotics buyer who clicks Pricing sees $20 phone-scanning plans and may conclude the company is a consumer scanning app.

    Medium impactLow effort First result in 2 weeks Web team

    First action

    Add a link to the Embodied AI plans at the top of /pricing.

    How to measure it

    Pricing-to-plans clicks

    Now
    –
    Target
    10% of pricing visits
    By
    30 days

    Steps

    1. 1

      Add the switch

      Three tabs at the top.

      Web team 1 day

    2. 2

      Write the chooser line

      Which plan family fits whom.

      Senior Product Marketing Manager, Embodied AI 1 hour

    3. 3

      Update the calculator

      Recommend Embodied AI plans for robotics.

      Web team 2 days

    4. 4

      Add events

      Track tab clicks.

      Web team 1 hour

    5. 5

      Review

      Click-through after a month.

      Senior Product Marketing Manager, Embodied AI 1 month

    6. 6

      Repeat for industries

      Defense and energy entry points.

      Senior Product Marketing Manager, Embodied AI 1 quarter

    First result
    2 weeks
    Cost
    $0
    Owner
    Web team

    Smallest test

    Add one link and watch clicks for a week.

    Why it works

    Add a switch at the top of /pricing (Scaniverse, Embodied AI, Enterprise), link the Embodied AI plans, and add one line explaining which to choose. Add the Embodied AI plans to the calculator's recommendations when the visitor picks robotics or logistics.

    Pricing pages are one of the most visited pages for B2B buyers, and showing the right offer avoids losing the high-value buyer at the first click.

    Risks and how to handle them

    • Confusing two price lists

      Keep one clear chooser line above both.

    Evidence

    • Your websiteThe /pricing page covers Scaniverse only: Free ($0, 20,000 monthly credits), Plus ($20 a month, "360° camera Gaussian splats and USDZ export"), Pro ($50 a month, commercial rights) and Enterprise, with "Get Started for Free", monthly/annual billing and credit top-ups; it has no link to the Embodied AI plans. A separate /plan-comparison-calculator recommends a Scaniverse plan only.Source
    • Your websiteEmbodied AI Plans page (checked October 22, 2026): the meta description says "from self-serve Evaluate to Enterprise", but no plan shows a price. Evaluate ("3-month minimum", "15 Places Library downloads across the term", "15 capture minutes a month (~1 place)"), Train ("Most teams start here", 12-month minimum), Deploy and Enterprise all end in the same "Select plan" form: "Connect with our team to discuss the shape of your plan". The FAQ calls Evaluate "a paid three-month plan".Source
    • Your websiteHomepage: "Real-world foundation models for physical AI"; "30B+ Posed images from real-world locations", "50M+ Neural networks trained", "140+ Patents filed or pending"; "We created SPZ, the open-source file format for Gaussian splats now used across the industry"; three industries in the menu (Embodied AI, Defense and Intelligence, Oil and Gas).Source
  • 05

    The library is described as 100 warehouse, residential and outdoor places across industrial, last-mile, commercial and residential settings, but the page shows 16, filters only Last Mile and Warehouse, and says "+84 more places; reach out for access". Buyers cannot see whether the library covers their environment before they talk to sales.

    High impactMedium effort First result in 1 month Senior Product Marketing Mana…

    First action

    Export the full 100-place list with categories from the catalog.

    How to measure it

    Library-page inquiries

    Now
    –
    Target
    +40%
    By
    60 days

    Steps

    1. 1

      Export the catalog

      All 100 places and fields.

      Senior/Staff Product Engineer, Embodied AI 2 days

    2. 2

      Make previews

      One image or short splat clip each.

      Capture team 1 week

    3. 3

      Build the grid

      Four category filters and robot-type tags.

      Web team 1 week

    4. 4

      Place pages

      One indexable page per place.

      Web team 1 week

    5. 5

      Link to plans

      How many places each plan covers.

      Senior Product Marketing Manager, Embodied AI 1 day

    6. 6

      Measure

      Views, filters used and inquiries.

      Senior Product Marketing Manager, Embodied AI monthly

    First result
    1 month
    Cost
    Web and capture time
    Owner
    Senior Product Marketing Manager, Embodied AI

    Smallest test

    List all 100 names on the page this week.

    Why it works

    List every place with a preview image, category, city, size and what it suits (AMR, humanoid, drone). Add filters for all four categories, keep downloads behind a plan, and give each place its own indexable page. Show "new" and "most requested" badges.

    Breadth is the library's selling point; showing it lets buyers self-qualify and makes Train's 100 downloads tangible. Individual pages also rank for searches like "warehouse simulation environment".

    Risks and how to handle them

    • Customers see gaps

      Pair gaps with "Request a place" and capture times.

    Evidence

    • Your websitePlaces Library page: "a collection of 100 off-the-shelf warehouse, residential, and outdoor environments", but the page lists 16 places, with filters only for "Last Mile" and "Warehouse"; two are "Free Sample" downloads, the rest say "Select Plan to Access", and "+84 more places in the library; reach out for access!". A "Request A Place" vote (Warehouse, Factory, Refinery, Data center, Construction, Natural, Urban, Residential) shows 0 on every option. The launch post says the catalog spans "industrial environments, last-mile and logistics settings, commercial spaces, and residential environments".Source
    • Your websiteEmbodied AI Plans page (checked October 22, 2026): the meta description says "from self-serve Evaluate to Enterprise", but no plan shows a price. Evaluate ("3-month minimum", "15 Places Library downloads across the term", "15 capture minutes a month (~1 place)"), Train ("Most teams start here", 12-month minimum), Deploy and Enterprise all end in the same "Select plan" form: "Connect with our team to discuss the shape of your plan". The FAQ calls Evaluate "a paid three-month plan".Source
    • HiringSenior Product Marketing Manager, Embodied AI (posted September 22, 2026): "Define the Market Narrative" for embodied AI across "our website, campaigns, sales materials, and partner communications"; "Shape pricing, packaging, and initial product-led growth motions, turning technical capabilities into clear offerings and activation paths"; "Build the target-account list for robotics and physical AI"; "Run competitive intelligence"; build "co-marketing and case-study partnerships with robotics OEMs, simulation ecosystems, and strategic customers"; and "Ensure every message clearly distinguishes what ships today, what is being validated, and what is on the roadmap".Source
  • 06

    Real Chrome visitors fail Core Web Vitals on the homepage: content appears in 4.9 s and taps take 492 ms to respond. In Google's phone test the largest element is the cookie notice text, which appears at 9.4 s, and the page downloads 2.9 MB, mostly images.

    Medium impactMedium effort First result in 1 month Web team

    First action

    Move the cookie notice into the server-rendered page and re-run PageSpeed.

    How to measure it

    Field LCP at the 75th percentile

    Now
    –
    Target
    under 2.5 s
    By
    90 days

    Steps

    1. 1

      Fix the cookie notice

      Server-rendered, no late font swap.

      Web team 1 day

    2. 2

      Compress images

      Modern formats at display size.

      Web team 2 days

    3. 3

      Lazy-load demos

      Below-the-fold media on scroll.

      Web team 2 days

    4. 4

      Trim tags

      Remove unused tags and scripts.

      Web team 2 days

    5. 5

      Improve tap response

      Break up long scripts on menus.

      Web team 1 week

    6. 6

      Re-test

      PageSpeed now and field data after 28 days.

      Web team 1 month

    First result
    1 month
    Cost
    Developer time
    Owner
    Web team

    Smallest test

    Fix the cookie notice and re-test today.

    Why it works

    Render the cookie notice in the first HTML instead of after scripts, compress and resize the hero and demo images (about 1.6 MB of savings), lazy-load demos below the fold, and trim unused JavaScript loaded through the tag manager. Re-test, then watch the Core Web Vitals report.

    Technical buyers judge a spatial AI company by how fast its own site feels, and faster pages convert more visitors from campaigns and events.

    Risks and how to handle them

    • Consent rules

      Keep the notice compliant while making it lighter.

    Evidence

    • PageSpeed testGoogle PageSpeed for the nianticspatial.com homepage (October 22, 2026). Real Chrome visitors (Chrome UX Report, 75th percentile) fail Core Web Vitals: the main content appears in 4.9 s (good is 2.5 s) and taps respond in 492 ms (good is 200 ms); layout shift is 0. The mobile lab test scores 67 out of 100 (desktop 87) with Largest Contentful Paint 9.4 s; the largest element is the cookie notice text; the page downloads 2.9 MB, about 2 MB of it images, with 1.6 MB of possible image savings and 233 KB of unused JavaScript.Source
    • Your websiteHomepage: "Real-world foundation models for physical AI"; "30B+ Posed images from real-world locations", "50M+ Neural networks trained", "140+ Patents filed or pending"; "We created SPZ, the open-source file format for Gaussian splats now used across the industry"; three industries in the menu (Embodied AI, Defense and Intelligence, Oil and Gas).Source

This quarter 10

The core plays for the next three months.

  • 07

    The strongest proof on the site is the Flexion and NVIDIA result (fall rate from 27.1% to 2.1%), but it sits behind a livestream form. The Real-World Test Lab is hiring to "build the system that produces our evidence", with scorecards comparable "across product versions, devices, and capture conditions".

    High impactMedium effort First result in 1 quarter AI Automation Engineer

    First action

    Publish the Flexion study as an ungated two-page summary.

    How to measure it

    Visits to the evidence page that become inquiries

    Now
    –
    Target
    5%
    By
    2 quarters

    Steps

    1. 1

      Write the summary

      Method, numbers and limits.

      Research Scientist, Embodied AI 2 weeks

    2. 2

      Design the scorecard

      Fields with lineage.

      AI Automation Engineer, Real-World Test Lab 2 weeks

    3. 3

      Automate updates

      From Lab runs each release.

      AI Automation Engineer, Real-World Test Lab 1 quarter

    4. 4

      Publish

      Ungated page linked from Embodied AI.

      Web team 1 week

    5. 5

      Use in sales

      In every evaluation plan.

      GTM, Embodied AI ongoing

    6. 6

      Review

      Quarterly with research.

      Senior Product Marketing Manager, Embodied AI quarterly

    First result
    1 quarter
    Cost
    Lab and writing time
    Owner
    AI Automation Engineer, Real World Test Lab

    Smallest test

    Ungate the Flexion summary this month.

    Why it works

    Turn the Lab's scorecards into a public page: each benchmark with scene, policy, simulator version, capture device and result, updated with each release. Publish a short technical report on the Flexion study, ungated, with method and limits stated.

    Robotics buyers trust reproducible numbers over claims, and the PMM role requires every message to separate what ships from what is being validated. A public scorecard does both.

    Risks and how to handle them

    • Results that look weak

      Publish them anyway with context; credibility is the goal.

    Evidence

    • Your websiteEmbodied AI page: "Make real places simulatable"; "A deployment environment starts at a few hundred dollars vs. tens of thousands to professionally survey or hand-build synthetic twins"; "hours, not weeks"; delivered as "one USDZ per place" (Gaussian splat plus aligned collision mesh) that "Runs in: Isaac Sim · Isaac Lab · any OpenUSD-compatible simulator"; upload is "Drag and drop into a secure GCS bucket. Enterprise APIs coming soon." In a comparison with Flexion and NVIDIA, an RGB policy trained on the reconstruction cut the fall rate from 27.1% (synthetic office) to 2.1%; the livestream sits behind a form. Roadmap items "Vary" and "Deploy and feed the loop" invite teams to "Tell Us What You Need".Source
    • HiringAI Automation Engineer, Real-World Test Lab (posted September 29, 2026): "build the system that produces our evidence. Today our evaluations are manual, inconsistent, and slow"; "Instrument scorecards so a result can be compared across product versions, devices, and capture conditions. A number without its lineage is not evidence." A Solutions Architect (Enterprise), still open, focuses on "Energy and Industrial" customers using "Intelligent Field Operations, Reconstruction, Semantics, and Robotics" and will "Maintain a set of repeatable demos for commercial prospects".Source
    • HiringSenior Product Marketing Manager, Embodied AI (posted September 22, 2026): "Define the Market Narrative" for embodied AI across "our website, campaigns, sales materials, and partner communications"; "Shape pricing, packaging, and initial product-led growth motions, turning technical capabilities into clear offerings and activation paths"; "Build the target-account list for robotics and physical AI"; "Run competitive intelligence"; build "co-marketing and case-study partnerships with robotics OEMs, simulation ecosystems, and strategic customers"; and "Ensure every message clearly distinguishes what ships today, what is being validated, and what is on the roadmap".Source
    • Your websiteSitemap (checked October 22, 2026): three customer stories (Sphere, Forgers Rittai, The Running Man), none about robots; 39 /campaigns landing pages (including Gartner reports, "physical-ai-readiness-checklist", "simulation-vs-digital-twin" and "agentic-ai-vs-physical-ai"); no comparison pages against other approaches or vendors; a /forms/usdz-samples page where "7 more scenes unlock with the form"; a public map-free relocalization benchmark with a leaderboard under /research/map-free; blog posts on Coco Robotics and the Flexion humanoid real-to-sim work.Source
  • 08

    The site's three customer stories (Sphere, Forgers Rittai, The Running Man) are about collaboration and AR; none involves robots. Coco Robotics, the first robotics partner, and the Flexion humanoid study are told only as blog posts.

    High impactMedium effort First result in 1 quarter Senior Product Marketing Mana…

    First action

    Ask Coco and Flexion for permission to publish metrics.

    How to measure it

    Inquiries after case-study views

    Now
    –
    Target
    +25%
    By
    2 quarters

    Steps

    1. 1

      Get approvals

      Coco and Flexion.

      GTM, Embodied AI 2 weeks

    2. 2

      Interview

      Engineering leads.

      Senior Product Marketing Manager, Embodied AI 2 weeks

    3. 3

      Write

      Problem, approach, results.

      Senior Product Marketing Manager, Embodied AI 2 weeks

    4. 4

      Design

      Clips and charts.

      Web team 1 week

    5. 5

      Publish

      Embodied AI page and decks.

      Senior Product Marketing Manager, Embodied AI 1 week

    6. 6

      Add a third

      From an Evaluate customer.

      GTM, Embodied AI 1 quarter

    First result
    1 quarter
    Cost
    Writing time
    Owner
    Senior Product Marketing Manager, Embodied AI

    Smallest test

    Turn the Flexion post into a one-page case study.

    Why it works

    Write two case studies with numbers: Coco (localization where GPS fails) and Flexion (policy transfer). Add one warehouse or AMR story from an early Evaluate customer. Put them on the Embodied AI page and in sales decks.

    Robotics teams buy on peer proof, and the PMM is hired to build "case-study partnerships with robotics OEMs".

    Risks and how to handle them

    • Partners limit what can be said

      Agree the numbers before writing.

    Evidence

    • Your websiteSitemap (checked October 22, 2026): three customer stories (Sphere, Forgers Rittai, The Running Man), none about robots; 39 /campaigns landing pages (including Gartner reports, "physical-ai-readiness-checklist", "simulation-vs-digital-twin" and "agentic-ai-vs-physical-ai"); no comparison pages against other approaches or vendors; a /forms/usdz-samples page where "7 more scenes unlock with the form"; a public map-free relocalization benchmark with a leaderboard under /research/map-free; blog posts on Coco Robotics and the Flexion humanoid real-to-sim work.Source
    • Your websiteCoco Robotics partnership post: Coco, which runs a robot delivery fleet, is Niantic Spatial's "first robotics partner"; Niantic Spatial deploys its Visual Positioning System for "maintaining accurate geolocation where GPS fails" in urban canyons.Source
    • Your websiteEmbodied AI page: "Make real places simulatable"; "A deployment environment starts at a few hundred dollars vs. tens of thousands to professionally survey or hand-build synthetic twins"; "hours, not weeks"; delivered as "one USDZ per place" (Gaussian splat plus aligned collision mesh) that "Runs in: Isaac Sim · Isaac Lab · any OpenUSD-compatible simulator"; upload is "Drag and drop into a secure GCS bucket. Enterprise APIs coming soon." In a comparison with Flexion and NVIDIA, an RGB policy trained on the reconstruction cut the fall rate from 27.1% (synthetic office) to 2.1%; the livestream sits behind a form. Roadmap items "Vary" and "Deploy and feed the loop" invite teams to "Tell Us What You Need".Source
    • HiringSenior Product Marketing Manager, Embodied AI (posted September 22, 2026): "Define the Market Narrative" for embodied AI across "our website, campaigns, sales materials, and partner communications"; "Shape pricing, packaging, and initial product-led growth motions, turning technical capabilities into clear offerings and activation paths"; "Build the target-account list for robotics and physical AI"; "Run competitive intelligence"; build "co-marketing and case-study partnerships with robotics OEMs, simulation ecosystems, and strategic customers"; and "Ensure every message clearly distinguishes what ships today, what is being validated, and what is on the roadmap".Source
  • 09

    Buyers weigh synthetic data (Parallel Domain), generated 3D worlds (World Labs) and reality capture tools (Polycam, Matterport). The site has campaign pages such as "simulation vs digital twin" but none that compares these approaches for training robots.

    High impactMedium effort First result in 1 quarter Senior Product Marketing Mana…

    First action

    Draft the synthetic-vs-real-to-sim page from the sales team's most common objection.

    How to measure it

    Comparison-page inquiries

    Now
    –
    Target
    20 a month
    By
    2 quarters

    Steps

    1. 1

      List objections

      From GTM and SDR notes.

      Sales Development Representative, GTM 1 week

    2. 2

      Check rivals

      Current offers from their own pages.

      Senior Product Marketing Manager, Embodied AI 1 week

    3. 3

      Write pages

      Fair tables with sources.

      Senior Product Marketing Manager, Embodied AI 3 weeks

    4. 4

      Legal review

      Claims and trademarks.

      Head of Legal 1 week

    5. 5

      Publish

      Linked from Embodied AI.

      Web team 1 day

    6. 6

      Refresh

      Quarterly.

      Senior Product Marketing Manager, Embodied AI quarterly

    First result
    1 quarter
    Cost
    Writing time
    Owner
    Senior Product Marketing Manager, Embodied AI

    Smallest test

    Publish one comparison page.

    Why it works

    Publish three pages: real-to-sim vs synthetic environments, vs generated world models, vs manual scanning. Compare cost per place, metric accuracy, colliders, training rights and time to first scene, with sources. Keep claims within what ships today.

    Comparison searches come from teams close to buying, and the PMM is asked to run competitive intelligence and keep the story defensible.

    Risks and how to handle them

    • Rivals change quickly

      Date every page and review quarterly.

    Evidence

    • Competitor researchRivals' own homepages: Parallel Domain "generates photorealistic synthetic data and simulation to train and validate perception systems for autonomous vehicles, robotics, and AI"; World Labs builds "frontier models that can perceive, generate, and interact with the 3D world"; Polycam offers "AI 3D Scanning, Floor Plans & Reality Capture for Teams"; Matterport sells 3D cameras and a platform to "digitize your building".Source
    • Your websiteSitemap (checked October 22, 2026): three customer stories (Sphere, Forgers Rittai, The Running Man), none about robots; 39 /campaigns landing pages (including Gartner reports, "physical-ai-readiness-checklist", "simulation-vs-digital-twin" and "agentic-ai-vs-physical-ai"); no comparison pages against other approaches or vendors; a /forms/usdz-samples page where "7 more scenes unlock with the form"; a public map-free relocalization benchmark with a leaderboard under /research/map-free; blog posts on Coco Robotics and the Flexion humanoid real-to-sim work.Source
    • HiringSenior Product Marketing Manager, Embodied AI (posted September 22, 2026): "Define the Market Narrative" for embodied AI across "our website, campaigns, sales materials, and partner communications"; "Shape pricing, packaging, and initial product-led growth motions, turning technical capabilities into clear offerings and activation paths"; "Build the target-account list for robotics and physical AI"; "Run competitive intelligence"; build "co-marketing and case-study partnerships with robotics OEMs, simulation ecosystems, and strategic customers"; and "Ensure every message clearly distinguishes what ships today, what is being validated, and what is on the roadmap".Source
    • Your websiteEmbodied AI page: "Make real places simulatable"; "A deployment environment starts at a few hundred dollars vs. tens of thousands to professionally survey or hand-build synthetic twins"; "hours, not weeks"; delivered as "one USDZ per place" (Gaussian splat plus aligned collision mesh) that "Runs in: Isaac Sim · Isaac Lab · any OpenUSD-compatible simulator"; upload is "Drag and drop into a secure GCS bucket. Enterprise APIs coming soon." In a comparison with Flexion and NVIDIA, an RGB policy trained on the reconstruction cut the fall rate from 27.1% (synthetic office) to 2.1%; the livestream sits behind a form. Roadmap items "Vary" and "Deploy and feed the loop" invite teams to "Tell Us What You Need".Source
  • 10

    The GTM lead is asked to "Track funding, hiring, launches, and deployments" and the PMM to build the target-account list for robotics and physical AI. Robotics companies that hire simulation engineers or announce new warehouse deployments are signalling a sim-to-real need.

    High impactMedium effort First result in 1 quarter GTM

    First action

    List 50 robotics companies hiring simulation engineers now.

    How to measure it

    Meetings booked

    Now
    –
    Target
    12 a month
    By
    1 quarter

    Steps

    1. 1

      Pick signals

      Funding, hiring, deployments.

      Senior Product Marketing Manager, Embodied AI 1 week

    2. 2

      Build the list

      In HubSpot with scores.

      Sales Development Representative, GTM 2 weeks

    3. 3

      Match places

      Library scenes per account.

      Senior Product Marketing Manager, Embodied AI weekly

    4. 4

      Write briefs

      Two lines per account.

      Sales Development Representative, GTM weekly

    5. 5

      Reach out

      Email, LinkedIn and events.

      Sales Development Representative, GTM weekly

    6. 6

      Review

      Which signals convert.

      GTM, Embodied AI monthly

    First result
    1 quarter
    Cost
    Data tools
    Owner
    GTM, Embodied AI

    Smallest test

    Send ten signal-based emails this week.

    Why it works

    Build a weekly list of robotics and physical AI companies with new funding, open simulation or robot-learning roles, or new deployment announcements. Score them by robot type and environment, match each to library places, and send the SDR a short brief per account.

    Timing beats volume in a small market, and matching each account to specific places makes the first message useful.

    Risks and how to handle them

    • Noisy signals

      Review scores monthly against meetings.

    Evidence

    • HiringGTM, Embodied AI (posted September 22, 2026): an entrepreneurial lead working "directly alongside the group's General Manager" to "map the market, open doors, and close business"; "Track funding, hiring, launches, and deployments"; "Get into the communities where our customers spend time"; "Be useful to accounts we're not ready to serve yet"; "Agree what an evaluation must prove, who will decide, and what follows success"; and "Help design commercial models for a maturing product". The role "will not initially carry a sales quota" but is accountable for "qualified pipeline".Source
    • HiringSenior Product Marketing Manager, Embodied AI (posted September 22, 2026): "Define the Market Narrative" for embodied AI across "our website, campaigns, sales materials, and partner communications"; "Shape pricing, packaging, and initial product-led growth motions, turning technical capabilities into clear offerings and activation paths"; "Build the target-account list for robotics and physical AI"; "Run competitive intelligence"; build "co-marketing and case-study partnerships with robotics OEMs, simulation ecosystems, and strategic customers"; and "Ensure every message clearly distinguishes what ships today, what is being validated, and what is on the roadmap".Source
    • HiringSales Development Representative, GTM (posted September 14, 2026): the "founding Sales Development Representative", "Reporting into Marketing", will "Respond quickly to inbound interest", "Develop outbound campaigns with Marketing and Business Development, testing audiences, messages, channels, and calls to action", research accounts "across industries such as oil and gas, defense and the public sector, and embodied AI", and "Capture patterns from prospect conversations: recurring pain points, objections, terminology, competitive context". The CRM is HubSpot.Source
    • Your websitePlaces Library page: "a collection of 100 off-the-shelf warehouse, residential, and outdoor environments", but the page lists 16 places, with filters only for "Last Mile" and "Warehouse"; two are "Free Sample" downloads, the rest say "Select Plan to Access", and "+84 more places in the library; reach out for access!". A "Request A Place" vote (Warehouse, Factory, Refinery, Data center, Construction, Natural, Urban, Residential) shows 0 on every option. The launch post says the catalog spans "industrial environments, last-mile and logistics settings, commercial spaces, and residential environments".Source
  • 11

    The library's vote on what to capture next shows 0 on every option and collects no contact details. Teams that ask for a refinery or data center are telling you their deployment environment.

    Medium impactLow effort First result in 1 month Senior Product Marketing Mana…

    First action

    Add an email field to the vote.

    How to measure it

    Place-request leads

    Now
    –
    Target
    30 a month
    By
    60 days

    Steps

    1. 1

      Add fields

      Email and robot type.

      Web team 2 days

    2. 2

      Show counts

      Live tallies.

      Web team 2 days

    3. 3

      Set the reward

      Early access to new categories.

      Senior Product Marketing Manager, Embodied AI 1 day

    4. 4

      Route

      HubSpot tasks for the SDR.

      Sales Development Representative, GTM 1 day

    5. 5

      Capture

      Top-voted category first.

      Capture team 1 quarter

    6. 6

      Announce

      Email voters when it ships.

      Senior Product Marketing Manager, Embodied AI ongoing

    First result
    1 month
    Cost
    $0
    Owner
    Senior Product Marketing Manager, Embodied AI

    Smallest test

    Add the email field this week.

    Why it works

    Ask voters for a work email and their robot type, show live counts, and promise the first ten teams who voted for a new category early access to it. Route votes to the SDR with the category chosen.

    It turns a passive widget into a qualified intent signal and a reason to come back.

    Risks and how to handle them

    • Low votes look bad

      Seed counts with real requests from sales notes.

    Evidence

    • Your websitePlaces Library page: "a collection of 100 off-the-shelf warehouse, residential, and outdoor environments", but the page lists 16 places, with filters only for "Last Mile" and "Warehouse"; two are "Free Sample" downloads, the rest say "Select Plan to Access", and "+84 more places in the library; reach out for access!". A "Request A Place" vote (Warehouse, Factory, Refinery, Data center, Construction, Natural, Urban, Residential) shows 0 on every option. The launch post says the catalog spans "industrial environments, last-mile and logistics settings, commercial spaces, and residential environments".Source
    • HiringSales Development Representative, GTM (posted September 14, 2026): the "founding Sales Development Representative", "Reporting into Marketing", will "Respond quickly to inbound interest", "Develop outbound campaigns with Marketing and Business Development, testing audiences, messages, channels, and calls to action", research accounts "across industries such as oil and gas, defense and the public sector, and embodied AI", and "Capture patterns from prospect conversations: recurring pain points, objections, terminology, competitive context". The CRM is HubSpot.Source
    • Your websiteEmbodied AI page: "Make real places simulatable"; "A deployment environment starts at a few hundred dollars vs. tens of thousands to professionally survey or hand-build synthetic twins"; "hours, not weeks"; delivered as "one USDZ per place" (Gaussian splat plus aligned collision mesh) that "Runs in: Isaac Sim · Isaac Lab · any OpenUSD-compatible simulator"; upload is "Drag and drop into a secure GCS bucket. Enterprise APIs coming soon." In a comparison with Flexion and NVIDIA, an RGB policy trained on the reconstruction cut the fall rate from 27.1% (synthetic office) to 2.1%; the livestream sits behind a form. Roadmap items "Vary" and "Deploy and feed the loop" invite teams to "Tell Us What You Need".Source
  • 12

    Evaluate is a paid three-month proof, and Train is where "most teams start". The GTM lead is asked to "Agree what an evaluation must prove, who will decide, and what follows success". Without a written plan, trials drift.

    High impactMedium effort First result in 1 quarter GTM

    First action

    Write the evaluation plan template.

    How to measure it

    Conversion to Train

    Now
    –
    Target
    50%
    By
    2 quarters

    Steps

    1. 1

      Write template

      Metric, scenes, decider, offer.

      GTM, Embodied AI 1 week

    2. 2

      Track usage

      Downloads and minutes.

      Senior/Staff Product Engineer, Embodied AI 2 weeks

    3. 3

      Set check-ins

      Weeks 2, 6, 10.

      GTM, Embodied AI 1 week

    4. 4

      Bring the architect

      For technical blockers.

      Solutions Architect (Enterprise) ongoing

    5. 5

      Offer Train

      With credit for Evaluate.

      GTM, Embodied AI ongoing

    6. 6

      Review

      Why trials did not convert.

      Senior Product Marketing Manager, Embodied AI quarterly

    First result
    1 quarter
    Cost
    Staff time
    Owner
    GTM, Embodied AI

    Smallest test

    Use the template with the next Evaluate customer.

    Why it works

    Give every Evaluate customer a one-page evaluation plan: success metric, scenes, who decides and the Train offer if it succeeds. Check usage of downloads and capture minutes at weeks 2, 6 and 10 and step in when usage is low.

    Clear success criteria lift conversion of paid pilots and shorten the step to a 12-month plan.

    Risks and how to handle them

    • Customers lack time to test

      Offer a joint working session in week 2.

    Evidence

    • Your websiteEmbodied AI Plans page (checked October 22, 2026): the meta description says "from self-serve Evaluate to Enterprise", but no plan shows a price. Evaluate ("3-month minimum", "15 Places Library downloads across the term", "15 capture minutes a month (~1 place)"), Train ("Most teams start here", 12-month minimum), Deploy and Enterprise all end in the same "Select plan" form: "Connect with our team to discuss the shape of your plan". The FAQ calls Evaluate "a paid three-month plan".Source
    • HiringGTM, Embodied AI (posted September 22, 2026): an entrepreneurial lead working "directly alongside the group's General Manager" to "map the market, open doors, and close business"; "Track funding, hiring, launches, and deployments"; "Get into the communities where our customers spend time"; "Be useful to accounts we're not ready to serve yet"; "Agree what an evaluation must prove, who will decide, and what follows success"; and "Help design commercial models for a maturing product". The role "will not initially carry a sales quota" but is accountable for "qualified pipeline".Source
    • HiringAI Automation Engineer, Real-World Test Lab (posted September 29, 2026): "build the system that produces our evidence. Today our evaluations are manual, inconsistent, and slow"; "Instrument scorecards so a result can be compared across product versions, devices, and capture conditions. A number without its lineage is not evidence." A Solutions Architect (Enterprise), still open, focuses on "Energy and Industrial" customers using "Intelligent Field Operations, Reconstruction, Semantics, and Robotics" and will "Maintain a set of repeatable demos for commercial prospects".Source
  • 13

    The PMM is hired to build co-marketing with "robotics OEMs, simulation ecosystems, and strategic customers". Robot makers want their customers to deploy faster; a ready simulation of the customer's site helps them sell.

    High impactMedium effort First result in 2 quarters Senior Product Marketing Mana…

    First action

    List ten robot makers whose models already run in Isaac Sim.

    How to measure it

    Partner Evaluate starts

    Now
    –
    Target
    10 a quarter
    By
    2 quarters

    Steps

    1. 1

      List partners

      AMR and humanoid makers.

      Senior Product Marketing Manager, Embodied AI 2 weeks

    2. 2

      Pitch

      Joint value and offer.

      GTM, Embodied AI 1 month

    3. 3

      Build demos

      Their robot in your scenes.

      Staff Engineer - Robot Learning, Embodied AI 1 month

    4. 4

      Launch

      Webinar and pages.

      Senior Product Marketing Manager, Embodied AI 1 month

    5. 5

      Share leads

      Agreed rules.

      GTM, Embodied AI ongoing

    6. 6

      Measure

      Evaluate starts by partner.

      Senior Product Marketing Manager, Embodied AI quarterly

    First result
    2 quarters
    Cost
    Demo engineering time
    Owner
    Senior Product Marketing Manager, Embodied AI

    Smallest test

    Run one joint demo with one robot maker.

    Why it works

    Sign two robot makers (an AMR and a humanoid company). Publish their robot models running in Places Library scenes, joint webinars, and a bundled offer: buy the robot, get Evaluate. Share leads.

    Partners bring buyers at the moment they choose a deployment, and joint content reaches their audiences.

    Risks and how to handle them

    • Partner conflicts

      Start with non-competing robot types.

    Evidence

    • HiringSenior Product Marketing Manager, Embodied AI (posted September 22, 2026): "Define the Market Narrative" for embodied AI across "our website, campaigns, sales materials, and partner communications"; "Shape pricing, packaging, and initial product-led growth motions, turning technical capabilities into clear offerings and activation paths"; "Build the target-account list for robotics and physical AI"; "Run competitive intelligence"; build "co-marketing and case-study partnerships with robotics OEMs, simulation ecosystems, and strategic customers"; and "Ensure every message clearly distinguishes what ships today, what is being validated, and what is on the roadmap".Source
    • Your websiteEmbodied AI page: "Make real places simulatable"; "A deployment environment starts at a few hundred dollars vs. tens of thousands to professionally survey or hand-build synthetic twins"; "hours, not weeks"; delivered as "one USDZ per place" (Gaussian splat plus aligned collision mesh) that "Runs in: Isaac Sim · Isaac Lab · any OpenUSD-compatible simulator"; upload is "Drag and drop into a secure GCS bucket. Enterprise APIs coming soon." In a comparison with Flexion and NVIDIA, an RGB policy trained on the reconstruction cut the fall rate from 27.1% (synthetic office) to 2.1%; the livestream sits behind a form. Roadmap items "Vary" and "Deploy and feed the loop" invite teams to "Tell Us What You Need".Source
    • Your websiteCoco Robotics partnership post: Coco, which runs a robot delivery fleet, is Niantic Spatial's "first robotics partner"; Niantic Spatial deploys its Visual Positioning System for "maintaining accurate geolocation where GPS fails" in urban canyons.Source
  • 14

    The Embodied AI page says POCs run "six to twelve months" because tuning waits for hardware on site, while with a twin "Robots arrive with most of their environment-specific training already complete". Warehouse automation integrators run exactly these deployments for many customers.

    High impactMedium effort First result in 2 quarters Solutions Architect

    First action

    Interview three warehouse integrators about their commissioning time.

    How to measure it

    Partner-captured places

    Now
    –
    Target
    40 a year
    By
    3 quarters

    Steps

    1. 1

      Interview

      Three integrators.

      Solutions Architect (Enterprise) 1 month

    2. 2

      Design the plan

      Partner pricing and support.

      GTM, Embodied AI 1 month

    3. 3

      Train partners

      Capture and upload.

      Solutions Architect (Enterprise) 1 month

    4. 4

      Run a pilot

      One customer site.

      Solutions Architect (Enterprise) 1 quarter

    5. 5

      Publish

      Joint case study.

      Senior Product Marketing Manager, Embodied AI 1 month

    6. 6

      Scale

      Five partners.

      GTM, Embodied AI 2 quarters

    First result
    2 quarters
    Cost
    Partner support time
    Owner
    Solutions Architect (Enterprise)

    Smallest test

    Run one integrator site end to end.

    Why it works

    Offer integrators a partner plan: they capture each customer site with a 360 camera, upload it, and train before installation. Give them partner pricing on capture minutes and co-branded case studies.

    Integrators repeat deployments across many sites, so one partner can bring dozens of places a year.

    Risks and how to handle them

    • Capture quality varies

      QA on every partner batch.

    Evidence

    • Your websiteEmbodied AI page: "Make real places simulatable"; "A deployment environment starts at a few hundred dollars vs. tens of thousands to professionally survey or hand-build synthetic twins"; "hours, not weeks"; delivered as "one USDZ per place" (Gaussian splat plus aligned collision mesh) that "Runs in: Isaac Sim · Isaac Lab · any OpenUSD-compatible simulator"; upload is "Drag and drop into a secure GCS bucket. Enterprise APIs coming soon." In a comparison with Flexion and NVIDIA, an RGB policy trained on the reconstruction cut the fall rate from 27.1% (synthetic office) to 2.1%; the livestream sits behind a form. Roadmap items "Vary" and "Deploy and feed the loop" invite teams to "Tell Us What You Need".Source
    • HiringAI Automation Engineer, Real-World Test Lab (posted September 29, 2026): "build the system that produces our evidence. Today our evaluations are manual, inconsistent, and slow"; "Instrument scorecards so a result can be compared across product versions, devices, and capture conditions. A number without its lineage is not evidence." A Solutions Architect (Enterprise), still open, focuses on "Energy and Industrial" customers using "Intelligent Field Operations, Reconstruction, Semantics, and Robotics" and will "Maintain a set of repeatable demos for commercial prospects".Source
    • Your websiteEmbodied AI Plans page (checked October 22, 2026): the meta description says "from self-serve Evaluate to Enterprise", but no plan shows a price. Evaluate ("3-month minimum", "15 Places Library downloads across the term", "15 capture minutes a month (~1 place)"), Train ("Most teams start here", 12-month minimum), Deploy and Enterprise all end in the same "Select plan" form: "Connect with our team to discuss the shape of your plan". The FAQ calls Evaluate "a paid three-month plan".Source
  • 15

    The Embodied AI page names the details that break policies: uneven floors, clutter, changing layouts, glass and white walls, and mentions MVSAnywhere for low-texture surfaces. Engineers search for these problems by name.

    Medium impactMedium effort First result in 1 quarter Senior Product Marketing Mana…

    First action

    Pick the first topic: glass and white walls.

    How to measure it

    Organic visits

    Now
    –
    Target
    +50%
    By
    2 quarters

    Steps

    1. 1

      Pick topics

      Six failure modes.

      Senior Product Marketing Manager, Embodied AI 1 week

    2. 2

      Write

      One per month.

      Research Scientist, Embodied AI monthly

    3. 3

      Add clips

      Short splat videos.

      Capture team monthly

    4. 4

      Promote

      Communities and email.

      Senior Product Marketing Manager, Embodied AI monthly

    5. 5

      Link

      To samples and plans.

      Web team monthly

    6. 6

      Measure

      Visits and downloads.

      Senior Product Marketing Manager, Embodied AI monthly

    First result
    1 quarter
    Cost
    Research time
    Owner
    Senior Product Marketing Manager, Embodied AI

    Smallest test

    Publish the first post this month.

    Why it works

    Publish a monthly post by a researcher on one failure mode: what goes wrong, how the reconstruction handles it, a short splat clip and a sample scene. Promote it in robotics communities and to the signal list.

    Specific, honest engineering content earns attention in a technical community and feeds search.

    Risks and how to handle them

    • Research time is scarce

      Interview researchers and let marketing write.

    Evidence

    • Your websiteEmbodied AI page: "Make real places simulatable"; "A deployment environment starts at a few hundred dollars vs. tens of thousands to professionally survey or hand-build synthetic twins"; "hours, not weeks"; delivered as "one USDZ per place" (Gaussian splat plus aligned collision mesh) that "Runs in: Isaac Sim · Isaac Lab · any OpenUSD-compatible simulator"; upload is "Drag and drop into a secure GCS bucket. Enterprise APIs coming soon." In a comparison with Flexion and NVIDIA, an RGB policy trained on the reconstruction cut the fall rate from 27.1% (synthetic office) to 2.1%; the livestream sits behind a form. Roadmap items "Vary" and "Deploy and feed the loop" invite teams to "Tell Us What You Need".Source
    • Your websiteSitemap (checked October 22, 2026): three customer stories (Sphere, Forgers Rittai, The Running Man), none about robots; 39 /campaigns landing pages (including Gartner reports, "physical-ai-readiness-checklist", "simulation-vs-digital-twin" and "agentic-ai-vs-physical-ai"); no comparison pages against other approaches or vendors; a /forms/usdz-samples page where "7 more scenes unlock with the form"; a public map-free relocalization benchmark with a leaderboard under /research/map-free; blog posts on Coco Robotics and the Flexion humanoid real-to-sim work.Source
    • HiringSenior Product Marketing Manager, Embodied AI (posted September 22, 2026): "Define the Market Narrative" for embodied AI across "our website, campaigns, sales materials, and partner communications"; "Shape pricing, packaging, and initial product-led growth motions, turning technical capabilities into clear offerings and activation paths"; "Build the target-account list for robotics and physical AI"; "Run competitive intelligence"; build "co-marketing and case-study partnerships with robotics OEMs, simulation ecosystems, and strategic customers"; and "Ensure every message clearly distinguishes what ships today, what is being validated, and what is on the roadmap".Source
  • 16

    Scaniverse is a well-rated app (4.8 stars from 11,838 ratings) with self-serve Plus and Pro plans and USDZ export. Some of those accounts belong to robotics teams already scanning their sites.

    Medium impactLow effort First result in 1 month Sales Development Representat…

    First action

    Query Scaniverse accounts by company domain.

    How to measure it

    Meetings from Scaniverse users

    Now
    –
    Target
    8 a month
    By
    1 quarter

    Steps

    1. 1

      Query accounts

      Company domains and USDZ export.

      Senior/Staff Product Engineer, Embodied AI 1 week

    2. 2

      Match to the list

      Robotics companies.

      Sales Development Representative, GTM 1 week

    3. 3

      Write the note

      From capture to simulation.

      Sales Development Representative, GTM 1 day

    4. 4

      Send

      Personal outreach.

      Sales Development Representative, GTM weekly

    5. 5

      Add an in-app prompt

      For USDZ exporters.

      Senior/Staff Product Engineer, Embodied AI 1 month

    6. 6

      Measure

      Meetings and Evaluate starts.

      Sales Development Representative, GTM monthly

    First result
    1 month
    Cost
    $0
    Owner
    Sales Development Representative, GTM

    Smallest test

    Contact ten matching accounts.

    Why it works

    Flag Scaniverse accounts on company domains from robotics, logistics and autonomy companies, especially those exporting USDZ. Send them a note about the Embodied AI pipeline and an Evaluate offer.

    These teams already capture with your tools; they need only the step to simulation.

    Risks and how to handle them

    • Privacy expectations

      Use only data covered by the business terms and offer opt-out.

    Evidence

    • App StoreScaniverse - 3D Scanner on the US App Store: 4.79 stars from 11,838 ratings, updated October 19, 2026.Source
    • Your websiteThe /pricing page covers Scaniverse only: Free ($0, 20,000 monthly credits), Plus ($20 a month, "360° camera Gaussian splats and USDZ export"), Pro ($50 a month, commercial rights) and Enterprise, with "Get Started for Free", monthly/annual billing and credit top-ups; it has no link to the Embodied AI plans. A separate /plan-comparison-calculator recommends a Scaniverse plan only.Source
    • HiringSales Development Representative, GTM (posted September 14, 2026): the "founding Sales Development Representative", "Reporting into Marketing", will "Respond quickly to inbound interest", "Develop outbound campaigns with Marketing and Business Development, testing audiences, messages, channels, and calls to action", research accounts "across industries such as oil and gas, defense and the public sector, and embodied AI", and "Capture patterns from prospect conversations: recurring pain points, objections, terminology, competitive context". The CRM is HubSpot.Source
    • Your websiteEmbodied AI page: "Make real places simulatable"; "A deployment environment starts at a few hundred dollars vs. tens of thousands to professionally survey or hand-build synthetic twins"; "hours, not weeks"; delivered as "one USDZ per place" (Gaussian splat plus aligned collision mesh) that "Runs in: Isaac Sim · Isaac Lab · any OpenUSD-compatible simulator"; upload is "Drag and drop into a secure GCS bucket. Enterprise APIs coming soon." In a comparison with Flexion and NVIDIA, an RGB policy trained on the reconstruction cut the fall rate from 27.1% (synthetic office) to 2.1%; the livestream sits behind a form. Roadmap items "Vary" and "Deploy and feed the loop" invite teams to "Tell Us What You Need".Source

Big bets 3

Larger plays that can change the trajectory.

  • 17

    Upload is drag and drop into a GCS bucket and "Enterprise APIs coming soon". Teams training at scale want captures to flow into their pipelines without manual steps.

    High impactHigh effort First result in 3 quarters Senior/Staff Product Engineer

    First action

    Ask five customers which steps they would automate first.

    How to measure it

    API-processed places

    Now
    –
    Target
    30% of volume
    By
    3 quarters

    Steps

    1. 1

      Interview

      Five customers.

      Senior/Staff Product Engineer, Embodied AI 2 weeks

    2. 2

      Design

      Upload, jobs, delivery, webhooks.

      Senior/Staff Product Engineer, Embodied AI 1 month

    3. 3

      Build

      With usage billing.

      Senior/Staff Product Engineer, Embodied AI 1 quarter

    4. 4

      Document

      Reference and examples.

      Senior/Staff Product Engineer, Embodied AI 1 month

    5. 5

      Beta

      Three customers.

      GTM, Embodied AI 1 quarter

    6. 6

      Launch

      With pricing.

      Senior Product Marketing Manager, Embodied AI 1 month

    First result
    3 quarters
    Cost
    Engineering team
    Owner
    Senior/Staff Product Engineer, Embodied AI

    Smallest test

    Release a job-status endpoint to one customer.

    Why it works

    Build an API to upload captures, track processing jobs and pull finished USDZ files, with webhooks and usage billing. Offer it first on Deploy and Enterprise, then to Train.

    Automation drives volume, and volume is how the plans lower the cost per place.

    Risks and how to handle them

    • Support load

      Start with a closed beta.

    Evidence

    • Your websiteEmbodied AI page: "Make real places simulatable"; "A deployment environment starts at a few hundred dollars vs. tens of thousands to professionally survey or hand-build synthetic twins"; "hours, not weeks"; delivered as "one USDZ per place" (Gaussian splat plus aligned collision mesh) that "Runs in: Isaac Sim · Isaac Lab · any OpenUSD-compatible simulator"; upload is "Drag and drop into a secure GCS bucket. Enterprise APIs coming soon." In a comparison with Flexion and NVIDIA, an RGB policy trained on the reconstruction cut the fall rate from 27.1% (synthetic office) to 2.1%; the livestream sits behind a form. Roadmap items "Vary" and "Deploy and feed the loop" invite teams to "Tell Us What You Need".Source
    • HiringSenior/Staff Product Engineer, Embodied AI (posted September 30, 2026): will "Own how autonomy teams experience our work: the APIs, UIs, packages, and documentation" and "Turn scans and reconstructions into reliable, interoperable assets that work in customers' simulation tools"; "a capability isn't a product until someone outside the building has used it without help". Nice to have: Isaac Sim, Isaac Lab, MuJoCo or OpenUSD. A Staff Engineer, Robot Learning (posted October 2) will "lead engineering for our real-to-sim product".Source
    • Your websiteEmbodied AI Plans page (checked October 22, 2026): the meta description says "from self-serve Evaluate to Enterprise", but no plan shows a price. Evaluate ("3-month minimum", "15 Places Library downloads across the term", "15 capture minutes a month (~1 place)"), Train ("Most teams start here", 12-month minimum), Deploy and Enterprise all end in the same "Select plan" form: "Connect with our team to discuss the shape of your plan". The FAQ calls Evaluate "a paid three-month plan".Source
  • 18

    Niantic Spatial already runs a public map-free relocalization benchmark with a leaderboard. Embodied AI lacks a shared test of how well policies trained in simulation work in real places.

    High impactHigh effort First result in 4 quarters Research Scientist

    First action

    Pick five places and one navigation task.

    How to measure it

    Submitting teams

    Now
    –
    Target
    25
    By
    4 quarters

    Steps

    1. 1

      Define the task

      Navigation in five places.

      Staff Engineer - Robot Learning, Embodied AI 1 month

    2. 2

      Build the harness

      Automated evaluation.

      AI Automation Engineer, Real-World Test Lab 1 quarter

    3. 3

      Publish the data

      With terms.

      Research Scientist, Embodied AI 1 month

    4. 4

      Launch

      At a robotics conference.

      Senior Product Marketing Manager, Embodied AI 1 month

    5. 5

      Run the leaderboard

      Monthly updates.

      AI Automation Engineer, Real-World Test Lab ongoing

    6. 6

      Publish results

      Yearly report.

      Research Scientist, Embodied AI yearly

    First result
    4 quarters
    Cost
    Research and lab time
    Owner
    Research Scientist, Embodied AI

    Smallest test

    Share the protocol with three labs for feedback.

    Why it works

    Release a set of real places with a fixed evaluation protocol and a leaderboard for navigation policies, run with the Test Lab. Invite universities and robotics teams to submit.

    Owning the benchmark makes the company the reference point for real-to-sim, and every entrant uses its places.

    Risks and how to handle them

    • Benchmarks get gamed

      Hold out test places.

    Evidence

    • Your websiteSitemap (checked October 22, 2026): three customer stories (Sphere, Forgers Rittai, The Running Man), none about robots; 39 /campaigns landing pages (including Gartner reports, "physical-ai-readiness-checklist", "simulation-vs-digital-twin" and "agentic-ai-vs-physical-ai"); no comparison pages against other approaches or vendors; a /forms/usdz-samples page where "7 more scenes unlock with the form"; a public map-free relocalization benchmark with a leaderboard under /research/map-free; blog posts on Coco Robotics and the Flexion humanoid real-to-sim work.Source
    • HiringAI Automation Engineer, Real-World Test Lab (posted September 29, 2026): "build the system that produces our evidence. Today our evaluations are manual, inconsistent, and slow"; "Instrument scorecards so a result can be compared across product versions, devices, and capture conditions. A number without its lineage is not evidence." A Solutions Architect (Enterprise), still open, focuses on "Energy and Industrial" customers using "Intelligent Field Operations, Reconstruction, Semantics, and Robotics" and will "Maintain a set of repeatable demos for commercial prospects".Source
    • Your websiteHomepage: "Real-world foundation models for physical AI"; "30B+ Posed images from real-world locations", "50M+ Neural networks trained", "140+ Patents filed or pending"; "We created SPZ, the open-source file format for Gaussian splats now used across the industry"; three industries in the menu (Embodied AI, Defense and Intelligence, Oil and Gas).Source
  • 19

    The library will "keep growing, with new environments added in response to what the community asks for", and Scaniverse has a large, engaged user base with 360 camera support. Capturing every requested factory or refinery with staff does not scale.

    High impactHigh effort First result in 4 quarters General Manager

    First action

    Draft the capture brief and rights checklist for one category.

    How to measure it

    Community places

    Now
    –
    Target
    200
    By
    4 quarters

    Steps

    1. 1

      Set rules

      Permissions, privacy, quality.

      Head of Legal 1 month

    2. 2

      Pick categories

      Top requests.

      Senior Product Marketing Manager, Embodied AI 1 week

    3. 3

      Recruit

      Vetted Scaniverse users.

      General Manager, Embodied AI 1 month

    4. 4

      Review

      QA and rights checks.

      Capture team ongoing

    5. 5

      Pay

      Per accepted place.

      Finance ongoing

    6. 6

      Scale

      New regions.

      General Manager, Embodied AI 2 quarters

    First result
    4 quarters
    Cost
    Capture payments
    Owner
    General Manager, Embodied AI

    Smallest test

    Pay ten capturers for one category.

    Why it works

    Invite vetted Scaniverse users to capture requested place types with the property owner's written permission, paying per accepted place. Review quality and rights before adding to the library.

    A community supply of places grows the catalog faster and cheaper, and the most-requested categories fill first.

    Risks and how to handle them

    • Rights and privacy issues

      Require owner permission and blur people.

    Evidence

    • Your websitePlaces Library page: "a collection of 100 off-the-shelf warehouse, residential, and outdoor environments", but the page lists 16 places, with filters only for "Last Mile" and "Warehouse"; two are "Free Sample" downloads, the rest say "Select Plan to Access", and "+84 more places in the library; reach out for access!". A "Request A Place" vote (Warehouse, Factory, Refinery, Data center, Construction, Natural, Urban, Residential) shows 0 on every option. The launch post says the catalog spans "industrial environments, last-mile and logistics settings, commercial spaces, and residential environments".Source
    • App StoreScaniverse - 3D Scanner on the US App Store: 4.79 stars from 11,838 ratings, updated October 19, 2026.Source
    • Your websiteEmbodied AI page: "Make real places simulatable"; "A deployment environment starts at a few hundred dollars vs. tens of thousands to professionally survey or hand-build synthetic twins"; "hours, not weeks"; delivered as "one USDZ per place" (Gaussian splat plus aligned collision mesh) that "Runs in: Isaac Sim · Isaac Lab · any OpenUSD-compatible simulator"; upload is "Drag and drop into a secure GCS bucket. Enterprise APIs coming soon." In a comparison with Flexion and NVIDIA, an RGB policy trained on the reconstruction cut the fall rate from 27.1% (synthetic office) to 2.1%; the livestream sits behind a form. Roadmap items "Vary" and "Deploy and feed the loop" invite teams to "Tell Us What You Need".Source

Wildcards 3

Unexpected experiments, each with a cheap way to test it.

  • 20

    The GTM lead is asked to get "into the communities where our customers spend time". At robotics conferences everyone shows videos; few show their product working on the room people are standing in.

    Medium impactMedium effort First result in 1 quarter GTM

    First action

    Pick one upcoming robotics event and check venue capture rules.

    How to measure it

    Contacts from the demo

    Now
    –
    Target
    150 per event
    By
    1 quarter

    Steps

    1. 1

      Pick the event

      One robotics show.

      GTM, Embodied AI 1 week

    2. 2

      Get permission

      Venue capture rights.

      GTM, Embodied AI 2 weeks

    3. 3

      Capture

      Morning of day one.

      Capture team 1 day

    4. 4

      Process

      Same-day USDZ.

      Senior/Staff Product Engineer, Embodied AI 1 day

    5. 5

      Send

      To visitors who ask.

      Sales Development Representative, GTM 1 day

    6. 6

      Follow up

      Evaluate offer.

      Sales Development Representative, GTM 2 weeks

    First result
    1 quarter
    Cost
    Event costs
    Owner
    GTM, Embodied AI

    Smallest test

    Do it at one event.

    Why it works

    At the next robotics event, capture the hall or booth area in the morning and offer visitors the USDZ by email that afternoon, with a sample policy running in it on screen.

    A live demo on a familiar place is memorable and produces qualified emails from engineers who want the file.

    Risks and how to handle them

    • Processing delays

      Test the same-day pipeline before the event.

    Evidence

    • HiringGTM, Embodied AI (posted September 22, 2026): an entrepreneurial lead working "directly alongside the group's General Manager" to "map the market, open doors, and close business"; "Track funding, hiring, launches, and deployments"; "Get into the communities where our customers spend time"; "Be useful to accounts we're not ready to serve yet"; "Agree what an evaluation must prove, who will decide, and what follows success"; and "Help design commercial models for a maturing product". The role "will not initially carry a sales quota" but is accountable for "qualified pipeline".Source
    • Your websiteEmbodied AI page: "Make real places simulatable"; "A deployment environment starts at a few hundred dollars vs. tens of thousands to professionally survey or hand-build synthetic twins"; "hours, not weeks"; delivered as "one USDZ per place" (Gaussian splat plus aligned collision mesh) that "Runs in: Isaac Sim · Isaac Lab · any OpenUSD-compatible simulator"; upload is "Drag and drop into a secure GCS bucket. Enterprise APIs coming soon." In a comparison with Flexion and NVIDIA, an RGB policy trained on the reconstruction cut the fall rate from 27.1% (synthetic office) to 2.1%; the livestream sits behind a form. Roadmap items "Vary" and "Deploy and feed the loop" invite teams to "Tell Us What You Need".Source
    • HiringSales Development Representative, GTM (posted September 14, 2026): the "founding Sales Development Representative", "Reporting into Marketing", will "Respond quickly to inbound interest", "Develop outbound campaigns with Marketing and Business Development, testing audiences, messages, channels, and calls to action", research accounts "across industries such as oil and gas, defense and the public sector, and embodied AI", and "Capture patterns from prospect conversations: recurring pain points, objections, terminology, competitive context". The CRM is HubSpot.Source
  • 21

    The Solutions Architect focuses on Energy and Industrial customers using Intelligent Field Operations and Robotics, and the site has an Oil and Gas industry page. Energy companies increasingly use inspection robots and drones on the same sites.

    Medium impactMedium effort First result in 2 quarters Solutions Architect

    First action

    Ask two energy customers about their robot or drone plans.

    How to measure it

    Cross-sold energy accounts

    Now
    –
    Target
    3
    By
    2 quarters

    Steps

    1. 1

      Ask customers

      Robot and drone plans.

      Solutions Architect (Enterprise) 2 weeks

    2. 2

      Design the add-on

      Twin for training.

      GTM, Embodied AI 1 month

    3. 3

      Pilot

      One facility.

      Solutions Architect (Enterprise) 1 quarter

    4. 4

      Publish

      Case study with approval.

      Senior Product Marketing Manager, Embodied AI 1 month

    5. 5

      Offer

      To other energy accounts.

      Solutions Architect (Enterprise) ongoing

    6. 6

      Measure

      Revenue per account.

      General Manager, Embodied AI quarterly

    First result
    2 quarters
    Cost
    Pilot time
    Owner
    Solutions Architect (Enterprise)

    Smallest test

    Discuss with one energy customer.

    Why it works

    Offer energy customers a twin of one facility for training their inspection robots or drones, priced as an add-on to their field operations work.

    It expands accounts already won, using data captured for another purpose.

    Risks and how to handle them

    • Sites restrict data

      Keep captures private and excluded from training.

    Evidence

    • HiringAI Automation Engineer, Real-World Test Lab (posted September 29, 2026): "build the system that produces our evidence. Today our evaluations are manual, inconsistent, and slow"; "Instrument scorecards so a result can be compared across product versions, devices, and capture conditions. A number without its lineage is not evidence." A Solutions Architect (Enterprise), still open, focuses on "Energy and Industrial" customers using "Intelligent Field Operations, Reconstruction, Semantics, and Robotics" and will "Maintain a set of repeatable demos for commercial prospects".Source
    • Your websiteHomepage: "Real-world foundation models for physical AI"; "30B+ Posed images from real-world locations", "50M+ Neural networks trained", "140+ Patents filed or pending"; "We created SPZ, the open-source file format for Gaussian splats now used across the industry"; three industries in the menu (Embodied AI, Defense and Intelligence, Oil and Gas).Source
    • Your websiteEmbodied AI page: "Make real places simulatable"; "A deployment environment starts at a few hundred dollars vs. tens of thousands to professionally survey or hand-build synthetic twins"; "hours, not weeks"; delivered as "one USDZ per place" (Gaussian splat plus aligned collision mesh) that "Runs in: Isaac Sim · Isaac Lab · any OpenUSD-compatible simulator"; upload is "Drag and drop into a secure GCS bucket. Enterprise APIs coming soon." In a comparison with Flexion and NVIDIA, an RGB policy trained on the reconstruction cut the fall rate from 27.1% (synthetic office) to 2.1%; the livestream sits behind a form. Roadmap items "Vary" and "Deploy and feed the loop" invite teams to "Tell Us What You Need".Source
  • 22

    The YouTube channel has 33 million views but about 6,100 subscribers and 40 videos, and the best embodied AI content, the Flexion livestream, is behind a form.

    Medium impactLow effort First result in 1 quarter Senior Product Marketing Mana…

    First action

    Record the first walkthrough on a free sample.

    How to measure it

    YouTube-sourced downloads

    Now
    –
    Target
    100 a month
    By
    2 quarters

    Steps

    1. 1

      Plan episodes

      Six places.

      Senior Product Marketing Manager, Embodied AI 1 week

    2. 2

      Record

      One a month.

      Senior/Staff Product Engineer, Embodied AI monthly

    3. 3

      Edit

      Short, step by step.

      Senior Product Marketing Manager, Embodied AI monthly

    4. 4

      Link

      Samples and quickstart.

      Web team monthly

    5. 5

      Cut clips

      For LinkedIn and X.

      Senior Product Marketing Manager, Embodied AI monthly

    6. 6

      Measure

      Downloads and inquiries.

      Senior Product Marketing Manager, Embodied AI monthly

    First result
    1 quarter
    Cost
    Production time
    Owner
    Senior Product Marketing Manager, Embodied AI

    Smallest test

    Publish one walkthrough.

    Why it works

    Publish a short series: one place from capture to a trained policy in under ten minutes each, with the steps on screen. Link to the quickstart and samples.

    Engineers learn on YouTube, and showing the whole workflow is the shortest explanation of the product.

    Risks and how to handle them

    • Engineer time

      Record in one session per month.

    Evidence

    • YouTubeThe Niantic Spatial YouTube channel has about 6,100 subscribers, 40 videos and 33 million total views.Source
    • Your websiteEmbodied AI page: "Make real places simulatable"; "A deployment environment starts at a few hundred dollars vs. tens of thousands to professionally survey or hand-build synthetic twins"; "hours, not weeks"; delivered as "one USDZ per place" (Gaussian splat plus aligned collision mesh) that "Runs in: Isaac Sim · Isaac Lab · any OpenUSD-compatible simulator"; upload is "Drag and drop into a secure GCS bucket. Enterprise APIs coming soon." In a comparison with Flexion and NVIDIA, an RGB policy trained on the reconstruction cut the fall rate from 27.1% (synthetic office) to 2.1%; the livestream sits behind a form. Roadmap items "Vary" and "Deploy and feed the loop" invite teams to "Tell Us What You Need".Source
    • Your websiteDeveloper docs (/docs, 2,547 pages in the docs sitemap): the NSDK for Unity, Swift and Kotlin and a few Scaniverse guides; no page covers importing a Places Library USDZ into Isaac Sim or Isaac Lab, or the embodied AI pipeline.Source

03

How Programmable can help

Where an idea in this playbook is something Programmable builds or runs, it's listed here with what we'd do. Everything else is yours to run.

ProgrammableInfo

Turn the research into a list you can act on

These ideas start with finding the right companies. Info finds them and the right person at each, researches every one and drafts the outreach, then asks before it sends.

  • Build this list in Info for idea #10 (opens programmable.info in a new tab)

New accounts: 100,000 free credits with code BETA100K

Find the companies (opens programmable.info in a new tab)
ProgrammableTalk

Fill the events in this plan

These ideas are events. Talk gets registrants for them and tells you who's coming and what their company does before the day.

Plan an event (opens programmable.talk in a new tab)
More example playbooks and reviews
ProgrammableDigital

Show Niantic Spatial’s product working

Turn screenshots or the website into an animated product demo in its own brand, for the homepage, launch posts and outreach in this playbook.

Make a demo (opens programmable.digital in a new tab)

More growth playbooks

Other companies we have researched, each read from what they are hiring for.

  • Motus growth strategyA platform for managing employee driving: vehicle reimbursement programs (FAVR, cents-per-mile, car allowance), driver risk protection…22 growth ideas
  • Moritz growth strategyAn AI-native law firm: AI workflows draft about 80% of each matter and a named attorney reviews, negotiates and signs off, with same-day…22 growth ideas
  • Cyber Defense Advisors growth strategyA Tampa-based cybersecurity and IT advisory firm, founded in 2001, selling 11 services (virtual CISO/CIO, compliance, risk assessment…23 growth ideas
  • Kobie growth strategyAn enterprise loyalty company: the Kobie Alchemy Loyalty Cloud platform with the Bonnie AI assistant, loyalty strategy and consulting…21 growth ideas
  • HYDAWAY growth strategyDesigns collapsible water bottles, drink tumblers, pints, insulated bowls, travel cases and bundles that pack flat, plus partner packable…21 growth ideas
  • voize growth strategyAn AI companion for nurses and caregivers: staff speak at the point of care and voize structures the documentation and writes it into the…22 growth ideas
  • Coframe growth strategyAn AI platform with managed service that generates, tests and personalizes website variants and landing pages, with every change approved…20 growth ideas
  • Sezzle growth strategyA public US fintech offering pay-later plans (Pay in 4, Pay in 5, Pay Monthly), subscriptions (Premium, Anywhere), a virtual card, credit…22 growth ideas

See every playbook and website review

Independent research. Programmable wrote this playbook from public sources, as they stood on October 23, 2026, and cites them throughout. Niantic Spatial did not commission, review or endorse it, and Programmable is not affiliated with Niantic Spatial. Company and product names are trademarks of their owners, used only to identify them. The ideas are suggestions and opinion, not professional, legal or financial advice, and not statements about how Niantic Spatial runs its business.

Something wrong, or do you represent Niantic Spatial? Email [email protected] and we will correct it or take it down.

30/60/90 roadmap

What to focus on in each month, and the ideas that carry it.

  1. Days 1 to 306 ideas

    Priced Evaluate, the Isaac Sim quickstart, sample follow-up, the pricing switch, the full library and the homepage speed fixes.

  2. Days 31 to 6012 ideas

    Evidence scorecard, robotics stories, comparison pages, signal list, place-request leads, the Evaluate playbook, partners, integrators, technical posts and Scaniverse leads; the booth demo and YouTube series.

  3. Days 61 to 904 ideas

    Commit to the Enterprise API, an open sim-to-real benchmark and a capture community; start the energy cross-sell.

The research behind the plan

3 research areas, each summarised in a few lines. Open any of them for the full detail.

Customers

What customers say

Customers value metric-scale, collider-ready scenes from a commodity camera. The most cited problem: the sim-to-real gap.

  • Training rights on every place
  • Months spent hand-building synthetic twins
  • In their words: “sim-to-real”

Market

Competitive landscape

4 competitors compared: Parallel Domain, World Labs, Polycam and Matterport.

  • Gap to exploit: Priced, self-serve real-to-sim trials
  • Parallel Domain: Synthetic, not the deployment site
  • World Labs: Generated geometry is not measured reality

Hiring signals

Niantic Spatial is building a go-to-market team for embodied AI from scratch.

  • Senior Product Marketing Manager, Embodied AI: Pricing, packaging and product-led growth. Ideas 1, 4, 5, 9.
  • GTM, Embodied AI: Map the market and close early deals. Ideas 10, 12, 20.
  • Sales Development Representative, GTM: Founding SDR for inbound and outbound. Ideas 3, 11, 16.

From the team that wrote this playbook

All products
  • ProgrammableAgencyWorking AI systems in days, not quarters (opens programmable.agency in a new tab)
  • ProgrammableInfoAn agent that finds, researches and asks firstNew accounts get 100,000 free enrichment credits with code BETA100K (opens programmable.info in a new tab)
  • ProgrammableDigitalScreenshots in, an animated product demo out (opens programmable.digital in a new tab)
  • ProgrammableWatchKnow the day a supplier falls out of compliance (opens programmable.watch in a new tab)
  • ProgrammableTalkGet registrants for your events, and know who’s coming (opens programmable.talk in a new tab)
  • ProgrammableMediaAn AI note taker, interpreter and webinar host for your calls (opens programmable.media in a new tab)
  • ProgrammablePageText what you need, your customer gets a page that works (opens programmable.page in a new tab)
  • ProgrammableCommunityForums with labelled AI members, never empty (opens programmable.community in a new tab)
ProgrammableMarketing
ExamplesLeaderboardsProductsTermsPrivacyCookiesDPALegalContact

© 2026 AI-Powered Growth Intelligence