ProgrammableMarketing/os3robotics.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 · 8
  • This quarter · 10
  • Big bets · 3
  • Wildcards · 2
  • How we can help
Generated Nov 10, 2026
  1. Programmable
  2. Playbooks
  3. OS3
OS3 logo

growth playbook · Nov 10, 2026

os3robotics.com (opens in a new tab)

OS3 growth playbook

Diagnosis

OS3 is a two-person Y Combinator (Summer 2026) team with a working robot, a narrow first job (folding and shelving hotel linens) and a published price that most rivals keep hidden. Its site is fast and clean, but it asks two very different buyers to take a lot on trust.

Hotels get "as low as $2.50/hour" and "reduce operating costs by half" with no math, no named property and only a booking link. Developers are asked to reserve a $14,995 unit with no docs, FAQ or ship date, and the launch post quoted different prices and a Fall 2026 ship date. The footage is unlabelled, and press has already noted that OS3 publishes no performance data.

This playbook first makes the offer concrete: a per-hotel quote on the Deploy page, labelled footage, one pricing note, depositor updates and a developer FAQ. Next it builds proof: a measured deployment, a weekly report for each hotel, a paid pilot and developer docs. The big bets are outcome pricing, a second hotel shift for the same robot, and selling through laundry equipment dealers.

Growth Score

6/10

Biggest Opportunity

OS3 is one of the few hotel robotics companies with a public hourly price. Turning that price into a quote each hotel can trust, backed by measured results, would make OS3 the easiest robot for a hotel to say yes to.

Biggest Risk

With a team of two and a robot still "in active development", unproven claims and silent waits for depositors could cost trust faster than proof arrives. Meanwhile better-funded rivals such as Dyna (towel-folding workflows, a pilot offer) and TechForce (linen-moving robots sold by subscription) court the same hotels.

01

Start here

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

  1. 1

    Turn "as low as $2.50/hour" into a quote for each hotel: a laundry-shift calculator and a "what's included" panel on the Deploy page

    Work out the formula from one live deployment: linen volume per occupied room, robot hours per day and the attendant hours it covers.

    High impact First result in 2 weeks
  2. 2

    Label every footage clip: autonomous or teleoperated, playback speed and uncut length

    Write the speed, mode and date for the four existing clips.

    High impact First result in 2 weeks
  3. 3

    Send every $99 depositor a monthly build update with their place in the queue

    Export the depositor list and draft the first update.

    High impact First result in 1 month

02

Your idea portfolio

23 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.

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

    1 of these are things Programmable can build and run for you.

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 · 4Big bets · 8Fill-ins · 8Reconsider · 3
Effort →

Showing 23 of 23 ideas

This week 8

Quick wins with a clear first action.

  • 01

    The Deploy page asks a hotel operations lead to book a call based on two numbers it never explains: "As low as $2.50/hour" and "Reduce operating costs by half." It doesn't say what an hour is (a robot hour or an hour of the laundry shift), what OS3 runs and what the hotel still does, or how many robot hours a property of their size would need. There is no form, FAQ or result, so every question waits for a call, and the founder joins that call knowing nothing about the property. Rivals set expectations up front: TechForce's linen robot is reported at roughly $150 a day, with setup, training and daily oversight included.

    High impactLow effort First result in 2 weeks Co founder

    First action

    Work out the formula from one live deployment: linen volume per occupied room, robot hours per day and the attendant hours it covers.

    How to measure it

    Qualified Deploy-page bookings

    Now
    –
    Target
    10 a month
    By
    60 days

    Steps

    1. 1

      Write the math

      From one live deployment: robot hours per occupied room and the attendant tasks covered.

      Founder 2 days

    2. 2

      Define the hour

      One sentence on what $2.50/hour measures and anything billed on top.

      Co-founder 1 day

    3. 3

      Write the panel

      What OS3 runs, what the hotel provides, and the safety basics from the tech page.

      Co-founder 1 day

    4. 4

      Build the calculator

      Five inputs and a range as the result; works on phones.

      Co-founder 3 days

    5. 5

      Pass the answers

      Prefill the booking link with rooms, occupancy and state.

      Co-founder 1 day

    6. 6

      Check against reality

      Compare estimates with actual usage at each deployed property.

      Founder monthly

    7. 7

      Track

      Calculator completions, bookings and how many hotels lack on-site laundry.

      Co-founder weekly

    First result
    2 weeks
    Cost
    Founder time only
    Owner
    Co founder

    Smallest test

    This week, add a static table under the Deploy hero for three hotel sizes (80, 150 and 300 rooms).

    Why it works

    Add three things above the "Schedule a call" button. (1) A calculator with five inputs: rooms, average occupancy, on-site laundry (yes/no), current laundry attendant hours per day and hourly wage. It returns estimated robot hours per day, the monthly HALE cost at the published rate, today's labour cost for the same six tasks, and the difference, shown as a range. (2) A "what $2.50/hour includes" panel. On OS3's side: it runs the robot, a remote operator steps in when a task needs correcting (as the training loop already does), and the robot charges from a standard outlet, lasts a full workday and fits standard doorways. On the hotel's side: floor space, its washers and dryers, and a staff member who starts each task. (3) Pass the calculator answers into the booking link, so the first call opens with the property's own numbers. Keep the existing "terms may apply" line, and say the final price is set after a site visit.

    Hotel buyers judge labour costs against an hourly wage they already know, and $2.50 an hour is a strong number. Showing the math for their property turns a claim into a budget line a GM can forward to the owner or management company. It also screens out hotels without on-site laundry before they take up one of a two-person team's call slots.

    Ready to use

    ChecklistDeploy page quote checklist

    [ ] One sentence: what "$2.50/hour" measures
    [ ] Calculator: rooms, occupancy, on-site laundry, attendant hours, wage
    [ ] Output as a range: robot hours/day, monthly cost, today's cost for the same six tasks
    [ ] What OS3 provides / what the hotel provides
    [ ] Safety basics: person starts each task, emergency stop, collision avoidance
    [ ] Booking link prefilled with the answers
    [ ] "Final price after a site visit" line

    Risks and how to handle them

    • Hotels read the estimate as a binding quote

      Show a range and say the final price is set after a site visit.

    • The real rate varies by property

      Use the published floor rate and list what moves the price.

    Evidence

    • Your websiteDeploy page: "The laundry shift, run by our robot." A six-step workflow (move linens, load the washer, washer to dryer, unload the dryer, fold, shelve); "Reduce operating costs by half."; "As low as $2.50/hour" with "Additional terms and conditions may apply"; "Deploying in three states": California, Michigan and Texas. The only call to action is "Schedule a call" (an external booking link); there is no form, FAQ, named hotel or measured result.Source
    • Your websiteTech page: arms 6+1 DoF each, 9 lb nominal and 15 lb peak per arm; telescoping torso from floor to top shelf; omnidirectional base that fits standard doorways; battery "Full workday", charged from a standard outlet; cloud planner plus on-robot policy, no onboard GPU; four demonstration rigs, including a 3D-printable rig and a handheld gripper; training loop in which a remote operator completes steps the robot cannot recover, followed by reinforcement learning on the failures; safety: motion starts only when a person initiates a task, camera and LiDAR collision avoidance, emergency stop, compliant arms. Reach, speed and battery hours are not listed.Source
    • Competitor researchTechForce's TIM-E (Hotel Dive, February 25, 2026): a robot that moves linens and waste at a Homewood Suites in Del Mar, California, offered by subscription at roughly $150 a day with setup, training and daily oversight included; the stated goal is to be "cost-neutral on the property"; housekeepers there clean 16 to 18 rooms a day; a comparable robot once cost upwards of $50,000 before mapping and training.Source
    • Your websiteHomepage: H1 "US-built semi-humanoid robots and video-trained intelligence"; HALE 1.0, a bimanual mobile robot, "Deployed in hospitality today"; 22 degrees of freedom on OS3-built actuators, 150,000 hours of video pretraining, 2,000+ real-world tasks, 2-4 hours of demonstrations to teach a new task; hospitality deployments from $2.50/hour; developer units $14,995 or $1,800/month with a $99 refundable deposit; calls to action "Deploy at your hotel" and "Partner with OS3"; startup partners mentioned but none named.Source
  • 02

    The footage page shows four clips (move linens, pack parts, robot overview, fold and align) with no label for speed, length or how the robot was controlled. A trade article has already pointed out that OS3's materials do not show which steps run autonomously in deployments. Buyers and developers who have seen sped-up robot demos assume the worst when nothing is labelled.

    High impactLow effort First result in 2 weeks Founder

    First action

    Write the speed, mode and date for the four existing clips.

    How to measure it

    Footage-to-CTA click rate

    Now
    –
    Target
    +30%
    By
    60 days

    Steps

    1. 1

      Gather facts

      Speed, length, mode and date for each clip.

      Founder 1 day

    2. 2

      Write captions

      One standard line under every clip.

      Co-founder 2 hours

    3. 3

      Record an uncut run

      Ten minutes of fold and align at 1x.

      Founder 2 days

    4. 4

      Explain the loop

      A short note linking to the training loop on the tech page.

      Co-founder 2 hours

    5. 5

      Add calls to action

      Deploy and Partner buttons under the clips.

      Co-founder 1 hour

    6. 6

      Measure

      Clicks and time on page before and after.

      Co-founder weekly

    First result
    2 weeks
    Cost
    $0
    Owner
    Founder

    Smallest test

    Caption the fold-and-align clip today.

    Why it works

    Add a caption line under each clip: playback speed (1x or the speed-up), clip length, control mode (autonomous, teleoperated or mixed), retries or remote corrections in the clip, and the date recorded. Add one uncut 10-minute fold-and-align run at 1x. Link a short note explaining that remote correction is part of the training loop by design, and that each correction becomes training data.

    Honest labels turn a doubt into a trust signal. Because OS3's training loop already includes remote correction, saying so is consistent with the tech page, not an admission.

    Risks and how to handle them

    • Some clips turn out to be teleoperated

      Label them as data collection and pair them with an autonomous clip of the same task.

    Evidence

    • Your websiteFootage page: "The robot, on the floor." Four clips (Move linens, Pack parts, Robot overview, Fold and align), with no label for playback speed, clip length or whether the robot was autonomous or teleoperated, and no metrics.Source
    • PressRuntimeWire startup spotlight (September 18, 2026): HALE's first job is towel folding in hotel laundry rooms; the $2.50/hour figure is OS3's own claim and was not independently verified; deployment scale, customer count and performance data are not provided, and OS3's materials do not establish that every step runs autonomously in deployments.Source
    • Your websiteTech page: arms 6+1 DoF each, 9 lb nominal and 15 lb peak per arm; telescoping torso from floor to top shelf; omnidirectional base that fits standard doorways; battery "Full workday", charged from a standard outlet; cloud planner plus on-robot policy, no onboard GPU; four demonstration rigs, including a 3D-printable rig and a handheld gripper; training loop in which a remote operator completes steps the robot cannot recover, followed by reinforcement learning on the failures; safety: motion starts only when a person initiates a task, camera and LiDAR collision avoidance, emergency stop, compliant arms. Reach, speed and battery hours are not listed.Source
  • 03

    OS3's pre-order terms say the deposit "holds a place in line", is refundable any time before shipment, and that ship windows are estimates. The launch post said shipments begin in Fall 2026, while the partner page gives no delivery date. Silent waits are when deposits get refunded and developers move to another platform.

    High impactLow effort First result in 1 month Co founder

    First action

    Export the depositor list and draft the first update.

    How to measure it

    Deposits refunded per quarter

    Now
    –
    Target
    Under 5%
    By
    90 days

    Steps

    1. 1

      Export the list

      Depositors, dates and option chosen (purchase or lease).

      Co-founder 1 hour

    2. 2

      Set the queue view

      Position and current estimated window per depositor.

      Co-founder 1 day

    3. 3

      Write update one

      Build progress, a clip and one question.

      Founder 1 day

    4. 4

      Publish a status page

      The same headline facts, dated.

      Co-founder 1 day

    5. 5

      Read the replies

      Tag planned use cases by vertical.

      Founder monthly

    6. 6

      Track

      Opens, replies and refunds.

      Co-founder monthly

    First result
    1 month
    Cost
    $0
    Owner
    Co founder

    Smallest test

    Send one update to all depositors this week and count replies.

    Build the monitor. A live page fed by the public registries and re-checked every day. Usually live within days.

    Book a working session

    Why it works

    Once a month, email every depositor: their place in the queue, the current ship window (as an estimate), what changed in hardware or the model, one new clip, and one question (what are you planning to build?). Keep a public build-status page with the same headline facts. Use their answers to plan the developer docs and partner verticals.

    Regular, honest updates keep reservations warm and turn a list of $99 deposits into a research panel on what developers will build.

    Risks and how to handle them

    • A later ship window prompts refunds

      Say it early, with the reason and what depositors get first.

    Evidence

    • Your websitePre-order terms (seller OS3 Inc.): the $99 deposit "holds a place in line; it is not a purchase or lease"; it is refundable any time before shipment; shipment windows are estimates, not guarantees; HALE 1.0 is in active development and specifications and timing may change; when a unit nears availability OS3 confirms the $14,995 purchase or the $1,800/month lease.Source
    • PressLaunch YC post (about three months old): OS3 is "already working with hospitality providers and biotech companies"; price targets under $10k for businesses (private beta) and under $15k for developers; a $99 refundable preorder deposit; "Shipments begin Fall 2026"; the technical report is "coming soon"; custom actuators for higher payloads are in development, and businesses with such use cases are invited to get in touch.Source
    • Your websitePartner page: "Own the application. Skip the robot R&D." OS3 provides HALE 1.0 hardware, control stack, base model, reinforcement-learning pipeline and support; the partner owns the vertical, integrations, customer and go-to-market. Developer unit $14,995 (not returnable once shipped) or Leased Developer $1,800/month (additional lease terms reviewed after reservation), each reserved with a $99 deposit via Stripe. "Partners currently deploying in" wet labs, hospitals and data centers, with no partner named; no SDK or documentation links, no FAQ and no delivery dates.Source
  • 04

    The Launch YC post, still a common way in, targets under $10k for businesses and under $15k for developers, with shipments from Fall 2026. The site now offers a $14,995 developer unit, a $1,800 monthly lease and an hourly hospitality service, gives no ship date, and has no business purchase option. A visitor arriving from the launch post meets different prices and no explanation.

    Medium impactLow effort First result in 1 week Co founder

    First action

    Decide whether the under-$10k business option still exists.

    How to measure it

    Calls where pricing confusion comes up

    Now
    –
    Target
    Halved
    By
    60 days

    Steps

    1. 1

      Decide the line-up

      Which options are on sale and which are planned.

      Founder 1 day

    2. 2

      Write the note

      Prices, ship window as an estimate, and the date.

      Co-founder 2 hours

    3. 3

      Publish

      On the partner page, linked from the homepage.

      Co-founder 1 hour

    4. 4

      Update the launch post

      Ask YC to add a pointer to the note.

      Founder 1 day

    5. 5

      Brief calls

      Same answer on every call.

      Founder 1 hour

    6. 6

      Review

      Update the date each time anything changes.

      Co-founder monthly

    First result
    1 week
    Cost
    $0
    Owner
    Co founder

    Smallest test

    Add a single dated line under the developer prices today.

    Why it works

    Add a short dated block to the partner page and link it from the homepage. It lists what is on sale today (the hospitality service per hour, a developer purchase, a developer lease), whether a business purchase is still planned, and the current ship window as an estimate. Ask YC to add an update to the launch post pointing to it.

    One source of truth keeps a price change from reading as a bait-and-switch, and it answers the first question on most calls.

    Risks and how to handle them

    • Early depositors expected the lower target

      Honour or explain it in the depositor update.

    Evidence

    • PressLaunch YC post (about three months old): OS3 is "already working with hospitality providers and biotech companies"; price targets under $10k for businesses (private beta) and under $15k for developers; a $99 refundable preorder deposit; "Shipments begin Fall 2026"; the technical report is "coming soon"; custom actuators for higher payloads are in development, and businesses with such use cases are invited to get in touch.Source
    • Your websitePartner page: "Own the application. Skip the robot R&D." OS3 provides HALE 1.0 hardware, control stack, base model, reinforcement-learning pipeline and support; the partner owns the vertical, integrations, customer and go-to-market. Developer unit $14,995 (not returnable once shipped) or Leased Developer $1,800/month (additional lease terms reviewed after reservation), each reserved with a $99 deposit via Stripe. "Partners currently deploying in" wet labs, hospitals and data centers, with no partner named; no SDK or documentation links, no FAQ and no delivery dates.Source
    • Your websitePre-order terms (seller OS3 Inc.): the $99 deposit "holds a place in line; it is not a purchase or lease"; it is refundable any time before shipment; shipment windows are estimates, not guarantees; HALE 1.0 is in active development and specifications and timing may change; when a unit nears availability OS3 confirms the $14,995 purchase or the $1,800/month lease.Source
  • 05

    OS3 is being shared through its YC profile, a Forbes feature on YC startups and trade coverage, but the homepage has no social share image. Links to os3robotics.com on LinkedIn and X show a bare card, when a robot folding towels is the most shareable image the company has.

    Medium impactLow effort First result in 1 week Co founder

    First action

    Export a 1200x630 still of HALE folding a towel.

    How to measure it

    Referral visits from LinkedIn and X

    Now
    –
    Target
    +30%
    By
    30 days

    Steps

    1. 1

      Pick stills

      One per page from existing footage.

      Founder 2 hours

    2. 2

      Add tags

      Open Graph and card tags in the page layout.

      Co-founder 2 hours

    3. 3

      Write descriptions

      One per page, written for that page's buyer.

      Co-founder 1 hour

    4. 4

      Check previews

      LinkedIn and X preview tools.

      Co-founder 1 hour

    5. 5

      Reshare

      Re-post the launch with the new card.

      Founder 1 hour

    6. 6

      Measure

      Referral clicks before and after.

      Co-founder monthly

    First result
    1 week
    Cost
    $0
    Owner
    Co founder

    Smallest test

    Add the homepage image today and share one post.

    Why it works

    Add an Open Graph and X card image to every page: HALE folding linens for Deploy, HALE with its developer specs for Partner, the hardware view for Tech, a clip still for Footage. Write a distinct description for each page around the buyer it serves. Check each one in the LinkedIn and X preview tools.

    Founder and press posts are OS3's main traffic source at this stage, and a picture in the link preview reliably earns more clicks than a bare link.

    Risks and how to handle them

    • Platforms cache old previews

      Refresh the cache in each platform's preview tool.

    Evidence

    • Your websiteSite check (November 10, 2026): page title "OS3 · US-built robots and video-trained intelligence"; the homepage has no social share image; no newsletter, live chat or contact page; the only published email address appears in the privacy policy and pre-order terms; the domain was registered in May 2026.Source
    • PressY Combinator profile: Summer 2026 batch, team size 2, San Francisco; one-liner "Affordable, intelligent humanoid robots built to deploy at scale"; lists a Forbes feature on YC startups dated September 3, 2026.Source
    • PressRuntimeWire startup spotlight (September 18, 2026): HALE's first job is towel folding in hotel laundry rooms; the $2.50/hour figure is OS3's own claim and was not independently verified; deployment scale, customer count and performance data are not provided, and OS3's materials do not establish that every step runs autonomously in deployments.Source
  • 06

    OS3's only open role, a contract Mechanical Design Engineer for robot parts and enclosures, is listed on Y Combinator but nowhere on os3robotics.com: the sitemap has no careers page and "Work with OS3" links only to Hospitality and Partners. The robotics engineers who read the control-stack post have no path to the job.

    Medium impactLow effort First result in 2 weeks Founder

    First action

    Copy the role into a /careers page.

    How to measure it

    Applicants citing the site

    Now
    –
    Target
    10
    By
    60 days

    Steps

    1. 1

      Create the page

      Role, location options and the design loop.

      Co-founder 2 hours

    2. 2

      Link it

      From the company page and footer.

      Co-founder 1 hour

    3. 3

      Add to the sitemap

      So search can find it.

      Co-founder 10 minutes

    4. 4

      Add to posts

      A hiring line under the control-stack article.

      Founder 10 minutes

    5. 5

      Share

      A clip of a part going from CAD to robot.

      Founder 1 hour

    6. 6

      Track

      Where applicants came from.

      Founder monthly

    First result
    2 weeks
    Cost
    $0
    Owner
    Founder

    Smallest test

    Add a Join link to the company page today.

    Why it works

    Publish a /careers page with the role and the design loop it offers: design remotely, parts built in San Francisco, feedback from real robots within days. Add "Join" to the "Work with OS3" section and a one-line hiring note at the end of each engineering post.

    Technical posts and footage attract exactly the people a two-person hardware team needs, and a visible role costs nothing.

    Risks and how to handle them

    • Too many unqualified applicants

      Ask for one portfolio link of a built part.

    Evidence

    • HiringMechanical Design Engineer - Robot Parts & Enclosures (Contract), San Francisco or remote (India), open to new grads and may convert to full-time: design parts, enclosures and covers for 3D printing, CNC and sheet metal; "Package cameras, sensors, compute, actuators, and wiring into serviceable assemblies"; OS3 builds the parts in San Francisco and returns feedback from real robots within days; nice to have: grippers and "industrial design sensibility for covers and finishes".Source
    • Your websiteSitemap (10 URLs, checked November 10, 2026), all last modified September 2, 2026: home, deploy, partner, tech, company, footage, blog, one blog post, privacy and pre-order terms. There is no careers, FAQ, case study, contact or pricing page, and the company page's "Work with OS3" section links only to Hospitality and Partners.Source
    • PressY Combinator profile: Summer 2026 batch, team size 2, San Francisco; one-liner "Affordable, intelligent humanoid robots built to deploy at scale"; lists a Forbes feature on YC startups dated September 3, 2026.Source
  • 17

    The partner page says purchased units cannot be returned once shipped and that lease terms are reviewed only after reservation. The pre-order terms say ship windows are estimates and specs may change, and neither page mentions a warranty or repairs. A startup weighing $14,995 has to book a call to learn the basics.

    Medium impactLow effort First result in 2 weeks Co founder

    First action

    List the ten questions developers asked most on calls.

    How to measure it

    Deposits per 100 partner-page visitors

    Now
    –
    Target
    +20%
    By
    60 days

    Steps

    1. 1

      Collect questions

      From calls and emails.

      Founder 1 day

    2. 2

      Write answers

      Plain words that match the terms.

      Co-founder 1 day

    3. 3

      Decide gaps

      Warranty, repair and data ownership policy.

      Founder 2 days

    4. 4

      Legal check

      Consistency with the pre-order terms.

      Founder 3 days

    5. 5

      Publish

      On the partner page with FAQ markup.

      Co-founder 2 hours

    6. 6

      Update

      Add new questions as they arrive.

      Co-founder monthly

    First result
    2 weeks
    Cost
    $0
    Owner
    Co founder

    Smallest test

    Publish the five most common answers first.

    Why it works

    Add a ten-question FAQ to the partner page: what is in the box, what platform support includes, warranty and repair, lease length and any buyout, what happens if specs change before shipping, the current ship window, who owns data and trained policies, and how to cancel a deposit. Have every answer checked against the pre-order terms.

    Clear answers on risk raise the share of visitors who reserve, and stop the same questions taking up founder calls.

    Risks and how to handle them

    • Committing to terms too early

      Say "current policy" and date the FAQ.

    Evidence

    • Your websitePartner page: "Own the application. Skip the robot R&D." OS3 provides HALE 1.0 hardware, control stack, base model, reinforcement-learning pipeline and support; the partner owns the vertical, integrations, customer and go-to-market. Developer unit $14,995 (not returnable once shipped) or Leased Developer $1,800/month (additional lease terms reviewed after reservation), each reserved with a $99 deposit via Stripe. "Partners currently deploying in" wet labs, hospitals and data centers, with no partner named; no SDK or documentation links, no FAQ and no delivery dates.Source
    • Your websitePre-order terms (seller OS3 Inc.): the $99 deposit "holds a place in line; it is not a purchase or lease"; it is refundable any time before shipment; shipment windows are estimates, not guarantees; HALE 1.0 is in active development and specifications and timing may change; when a unit nears availability OS3 confirms the $14,995 purchase or the $1,800/month lease.Source
  • 23

    Real visitors already pass Core Web Vitals (largest content in 0.78 s at the 75th percentile), so speed is not a problem today. But field load time is 34% slower than before, Google's phone test shows the HALE hero image at 2.9 s, and fonts weigh 159 KB. Labelled footage, an uncut run and a live stream will add heavy video.

    Low impactLow effort First result in 1 week Co founder

    First action

    Re-export the hero image for mobile.

    How to measure it

    Mobile lab LCP

    Now
    –
    Target
    Under 2.5 s
    By
    30 days

    Steps

    1. 1

      Compress the hero

      Modern format, phone sizes, priority.

      Co-founder 2 hours

    2. 2

      Subset fonts

      Only the weights and characters used.

      Co-founder 2 hours

    3. 3

      Lazy-load video

      Posters first, video on scroll.

      Co-founder 2 hours

    4. 4

      Re-test

      PageSpeed on home, deploy and footage.

      Co-founder 1 hour

    5. 5

      Watch field data

      Check the real-visitor trend.

      Co-founder monthly

    6. 6

      Set a budget

      A page-weight limit for new pages.

      Co-founder 1 hour

    First result
    1 week
    Cost
    $0
    Owner
    Co founder

    Smallest test

    Swap the hero image and re-run PageSpeed.

    Why it works

    Serve the hero image in a modern format at phone sizes with high fetch priority (about 84 KB of savings), subset the fonts to the characters used, lazy-load video below the fold with a light poster image, and re-run PageSpeed after each change.

    It keeps the site fast for hotel buyers on phones as the footage-heavy ideas ship.

    Risks and how to handle them

    • Image quality drops

      Compare side by side before shipping.

    Evidence

    • PageSpeedGoogle PageSpeed (November 10, 2026): mobile score 92, desktop 100; accessibility, best practices and SEO all 100. Real visitors pass Core Web Vitals (75th percentile LCP 0.78 s, INP 33 ms, CLS 0), although field LCP is 34% slower than in the previous period; in the lab phone test the hero image of HALE appears at 2.9 s; the homepage weighs 1,421 KB, including 159 KB of fonts, with about 84 KB of image savings available.Source
    • Your websiteFootage page: "The robot, on the floor." Four clips (Move linens, Pack parts, Robot overview, Fold and align), with no label for playback speed, clip length or whether the robot was autonomous or teleoperated, and no metrics.Source

This quarter 10

The core plays for the next three months.

  • 07

    The homepage says HALE is "Deployed in hospitality today", the Deploy page says OS3 is deploying in three states, and the launch post says it already works with hospitality providers. Yet the site names no hotel and publishes no result, and press noted the missing performance data. "Reduce operating costs by half" has nothing behind it.

    High impactMedium effort First result in 6 weeks Founder

    First action

    Ask the longest-running hotel for permission and agree what can be shown.

    How to measure it

    Deploy page to booked call

    Now
    –
    Target
    +40%
    By
    90 days

    Steps

    1. 1

      Get permission

      Named or anonymised, and what data can be shown.

      Founder 1 week

    2. 2

      Pull the data

      Hours, items, corrections and uptime from the robot logs.

      Founder 3 days

    3. 3

      Get the baseline

      Laundry hours and costs before HALE, from the hotel.

      Co-founder 1 week

    4. 4

      Interview

      Operations lead and one laundry attendant.

      Co-founder 2 days

    5. 5

      Write and film

      One page and a 60-second clip.

      Co-founder 1 week

    6. 6

      Approve

      Hotel signs off the final copy.

      Founder 1 week

    7. 7

      Distribute

      Deploy page, outreach and social.

      Co-founder ongoing

    First result
    6 weeks
    Cost
    Founder time and a small video budget
    Owner
    Founder

    Smallest test

    Publish three numbers from one property on the Deploy page.

    Why it works

    With one hotel's permission (anonymised if it prefers, such as "a 200-room hotel in Texas"), publish a one-page story: laundry-room hours before and after, items folded per hour, hours run unattended, remote corrections per shift and how they fell, what staff now do instead, and a quote from the operations lead. Show the cost comparison behind "half". Link it from Deploy, the calculator and outreach.

    Hotel buyers copy peers. One measured property does more than any claim, and it also answers the open questions in the press coverage.

    Risks and how to handle them

    • Results are below the headline claim

      Publish the real number and the improvement trend, and adjust the claim.

    • The hotel declines to be named

      Describe it by size, type and state.

    Evidence

    • Your websiteDeploy page: "The laundry shift, run by our robot." A six-step workflow (move linens, load the washer, washer to dryer, unload the dryer, fold, shelve); "Reduce operating costs by half."; "As low as $2.50/hour" with "Additional terms and conditions may apply"; "Deploying in three states": California, Michigan and Texas. The only call to action is "Schedule a call" (an external booking link); there is no form, FAQ, named hotel or measured result.Source
    • PressRuntimeWire startup spotlight (September 18, 2026): HALE's first job is towel folding in hotel laundry rooms; the $2.50/hour figure is OS3's own claim and was not independently verified; deployment scale, customer count and performance data are not provided, and OS3's materials do not establish that every step runs autonomously in deployments.Source
    • PressLaunch YC post (about three months old): OS3 is "already working with hospitality providers and biotech companies"; price targets under $10k for businesses (private beta) and under $15k for developers; a $99 refundable preorder deposit; "Shipments begin Fall 2026"; the technical report is "coming soon"; custom actuators for higher payloads are in development, and businesses with such use cases are invited to get in touch.Source
    • Your websiteHomepage: H1 "US-built semi-humanoid robots and video-trained intelligence"; HALE 1.0, a bimanual mobile robot, "Deployed in hospitality today"; 22 degrees of freedom on OS3-built actuators, 150,000 hours of video pretraining, 2,000+ real-world tasks, 2-4 hours of demonstrations to teach a new task; hospitality deployments from $2.50/hour; developer units $14,995 or $1,800/month with a $99 refundable deposit; calls to action "Deploy at your hotel" and "Partner with OS3"; startup partners mentioned but none named.Source
  • 08

    The partner page asks startups to reserve a $14,995 unit and "skip the robot R&D", but links no SDK or documentation. The tech page and the control-stack post already describe most of what a developer needs: the planner and on-robot executor, four data rigs, 2-4 hours of demonstrations per task, and the coming move to end-effector commands.

    High impactMedium effort First result in 1 month Co founder

    First action

    Outline the docs from the tech page and the control-stack post.

    How to measure it

    Deposits per 100 partner-page visitors

    Now
    –
    Target
    +50%
    By
    1 quarter

    Steps

    1. 1

      Outline

      Interface, data, training, support and limits.

      Founder 2 days

    2. 2

      Write the first task

      A worked example from demos to deployed policy.

      Founder 1 week

    3. 3

      Document the rigs

      What each rig records and when to use it.

      Founder 3 days

    4. 4

      Write the 30-day plan

      What a partner does from unboxing to first pilot.

      Co-founder 2 days

    5. 5

      Publish

      Link from the partner page and depositor updates.

      Co-founder 1 day

    6. 6

      Collect questions

      Add each call question to the docs.

      Co-founder ongoing

    First result
    1 month
    Cost
    Founder time only
    Owner
    Co founder

    Smallest test

    Publish the worked example as a blog post first.

    Why it works

    Publish a /developers page covering what a partner controls (task commands now, hand-pose commands when iteration 2 ships), how to collect demonstrations with each rig, who runs training and how long it takes, what platform support includes, and a worked example that teaches one new task from start to finish. Add a short "first 30 days with a developer unit" plan.

    Developers buy platforms they can picture building on. Docs before the deposit lift reservations and cut questions on calls.

    Risks and how to handle them

    • Interfaces change before ship

      Label docs as preview, versioned and dated.

    Evidence

    • Your websitePartner page: "Own the application. Skip the robot R&D." OS3 provides HALE 1.0 hardware, control stack, base model, reinforcement-learning pipeline and support; the partner owns the vertical, integrations, customer and go-to-market. Developer unit $14,995 (not returnable once shipped) or Leased Developer $1,800/month (additional lease terms reviewed after reservation), each reserved with a $99 deposit via Stripe. "Partners currently deploying in" wet labs, hospitals and data centers, with no partner named; no SDK or documentation links, no FAQ and no delivery dates.Source
    • Your websiteTech page: arms 6+1 DoF each, 9 lb nominal and 15 lb peak per arm; telescoping torso from floor to top shelf; omnidirectional base that fits standard doorways; battery "Full workday", charged from a standard outlet; cloud planner plus on-robot policy, no onboard GPU; four demonstration rigs, including a 3D-printable rig and a handheld gripper; training loop in which a remote operator completes steps the robot cannot recover, followed by reinforcement learning on the failures; safety: motion starts only when a person initiates a task, camera and LiDAR collision avoidance, emergency stop, compliant arms. Reach, speed and battery hours are not listed.Source
    • Your websiteBlog: one post, "From tokens to torque" (7-minute read, undated): RMS jerk down 75%, motion lag from 33.3 ms to 11.1 ms, command latency from 1.78 ms to 0.2 ms; it says a technical report on training is coming and invites questions by email. The blog has no newsletter signup or RSS feed.Source
  • 09

    The tech page says Rig 02 is 3D-printable and "reproducible anywhere", and the handheld gripper records demonstrations without a robot. The launch post and the blog both promise a training report "coming soon". The site has no newsletter or email capture, so the developers this content would attract leave no trace.

    High impactMedium effort First result in 6 weeks Founder

    First action

    Package the Rig 02 files and a one-page build guide.

    How to measure it

    Rig and report downloads

    Now
    –
    Target
    500
    By
    1 quarter

    Steps

    1. 1

      Package the files

      Print files, parts list and assembly photos.

      Mechanical Design Engineer (contract) 1 week

    2. 2

      Write the guide

      How to record demonstrations OS3 can train on.

      Founder 3 days

    3. 3

      Finish the report

      The promised training write-up.

      Founder 3 weeks

    4. 4

      Gate lightly

      An email field, nothing more.

      Co-founder 1 day

    5. 5

      Launch

      Post with a clip of the rig driving HALE.

      Co-founder 1 day

    6. 6

      Run a dataset call

      Train on the best submission and show the result.

      Founder quarterly

    First result
    6 weeks
    Cost
    Founder and contractor time
    Owner
    Founder

    Smallest test

    Post the rig parts list and count requests for the files.

    Why it works

    Publish the training report alongside the Rig 02 print files, a parts list and a recording guide, available for an email address. Developers who build the rig collect demonstrations in OS3's format before they own a robot. Run a quarterly call for datasets: OS3 trains a policy on the best one and shows it running on HALE.

    It turns curiosity into a list of qualified developers, and every dataset collected in OS3's format is a step toward a developer unit.

    Risks and how to handle them

    • Competitors copy the rig

      Release the rig, keep the training pipeline and model in-house.

    Evidence

    • Your websiteBlog: one post, "From tokens to torque" (7-minute read, undated): RMS jerk down 75%, motion lag from 33.3 ms to 11.1 ms, command latency from 1.78 ms to 0.2 ms; it says a technical report on training is coming and invites questions by email. The blog has no newsletter signup or RSS feed.Source
    • Your websiteTech page: arms 6+1 DoF each, 9 lb nominal and 15 lb peak per arm; telescoping torso from floor to top shelf; omnidirectional base that fits standard doorways; battery "Full workday", charged from a standard outlet; cloud planner plus on-robot policy, no onboard GPU; four demonstration rigs, including a 3D-printable rig and a handheld gripper; training loop in which a remote operator completes steps the robot cannot recover, followed by reinforcement learning on the failures; safety: motion starts only when a person initiates a task, camera and LiDAR collision avoidance, emergency stop, compliant arms. Reach, speed and battery hours are not listed.Source
    • PressLaunch YC post (about three months old): OS3 is "already working with hospitality providers and biotech companies"; price targets under $10k for businesses (private beta) and under $15k for developers; a $99 refundable preorder deposit; "Shipments begin Fall 2026"; the technical report is "coming soon"; custom actuators for higher payloads are in development, and businesses with such use cases are invited to get in touch.Source
  • 10

    OS3 runs the laundry service itself and is deploying in three states, while the robot is still in active development. Every new site means setup, visits and repairs, which cost far less when properties are close together. The only route in today is an inbound booking link.

    High impactMedium effort First result in 1 month Co founder

    First action

    List 50 hotels with on-site laundry near the busiest deployment.

    How to measure it

    Signed properties within an hour of a live site

    Now
    –
    Target
    3 per cluster
    By
    1 quarter

    Steps

    1. 1

      Map the clusters

      An hour's drive around each live site.

      Co-founder 3 days

    2. 2

      Build the list

      Hotels with on-property laundry and their management company.

      Co-founder 1 week

    3. 3

      Personalise

      A calculator result for each property's size.

      Co-founder ongoing

    4. 4

      Reach out

      Email and LinkedIn to operations leads.

      Founder weekly

    5. 5

      Visit

      Site visits batched by area.

      Founder monthly

    6. 6

      Review

      Reply, visit and sign rates by cluster.

      Co-founder monthly

    First result
    1 month
    Cost
    Founder time and travel
    Owner
    Co founder

    Smallest test

    Email 20 hotels near one site with their calculator result.

    Why it works

    For each live site, list the hotels with on-property laundry within an hour's drive, putting management companies that run several such properties first. Reach the operations lead with the calculator result for their size, the labelled footage and an offer of a site visit. Aim for clusters of three or more properties per area before adding a new state.

    Clustering lowers the cost per hour of supporting each robot, and peers in the same market are the best proof for one another.

    Risks and how to handle them

    • Founder time spread too thin

      Work one cluster at a time.

    Builds on

    Evidence

    • Your websiteDeploy page: "The laundry shift, run by our robot." A six-step workflow (move linens, load the washer, washer to dryer, unload the dryer, fold, shelve); "Reduce operating costs by half."; "As low as $2.50/hour" with "Additional terms and conditions may apply"; "Deploying in three states": California, Michigan and Texas. The only call to action is "Schedule a call" (an external booking link); there is no form, FAQ, named hotel or measured result.Source
    • Your websitePre-order terms (seller OS3 Inc.): the $99 deposit "holds a place in line; it is not a purchase or lease"; it is refundable any time before shipment; shipment windows are estimates, not guarantees; HALE 1.0 is in active development and specifications and timing may change; when a unit nears availability OS3 confirms the $14,995 purchase or the $1,800/month lease.Source
    • Competitor researchTechForce's TIM-E (Hotel Dive, February 25, 2026): a robot that moves linens and waste at a Homewood Suites in Del Mar, California, offered by subscription at roughly $150 a day with setup, training and daily oversight included; the stated goal is to be "cost-neutral on the property"; housekeepers there clean 16 to 18 rooms a day; a comparable robot once cost upwards of $50,000 before mapping and training.Source
  • 11

    Dyna's homepage leads with "Explore a pilot", and TechForce sells a subscription of about $150 a day with setup included. OS3 offers only a call. Hotel owners hesitate to commit to new labour technology without a short, bounded trial.

    High impactLow effort First result in 1 month Founder

    First action

    Set the pilot price and the three success measures.

    How to measure it

    Pilots that convert

    Now
    –
    Target
    60%
    By
    1 quarter

    Steps

    1. 1

      Set the terms

      Price, shifts, success bar and exit.

      Founder 2 days

    2. 2

      Legal check

      A one-page pilot agreement.

      Founder 1 week

    3. 3

      Publish

      A pilot block on the Deploy page.

      Co-founder 1 day

    4. 4

      Run check-ins

      Weekly numbers shared with the hotel.

      Founder weekly

    5. 5

      Convert

      Day-25 review with the measured results.

      Founder per pilot

    6. 6

      Learn

      Why pilots did or did not convert.

      Co-founder monthly

    First result
    1 month
    Cost
    Robot time at reduced margin
    Owner
    Founder

    Smallest test

    Offer the pilot to the next three hotels on calls.

    Why it works

    Publish a 30-day pilot on the Deploy page: a fixed price for a set number of shifts, success measures agreed up front (items folded per hour, hours run unattended, attendant hours moved to rooms), weekly check-ins, and the pilot fee credited to the first month if the hotel continues. Removal is free if the bar is missed.

    A bounded, paid trial is easier to approve than an open-ended contract, filters out tyre-kickers, and gives OS3 measured results for proof.

    Risks and how to handle them

    • Too many pilots for a small team

      Cap concurrent pilots and keep a waitlist.

    Evidence

    • Competitor researchDyna Robotics' homepage: its Dyna-2.1 model automates a long-horizon workflow of washing, drying and folding towels; it targets hotels and housekeeping, warehouses and food service; calls to action include "Explore a pilot" ("You'll believe it when you see it? Deal."), a contact form that asks company size, and a newsletter signup; no customers are named.Source
    • Competitor researchTechForce's TIM-E (Hotel Dive, February 25, 2026): a robot that moves linens and waste at a Homewood Suites in Del Mar, California, offered by subscription at roughly $150 a day with setup, training and daily oversight included; the stated goal is to be "cost-neutral on the property"; housekeepers there clean 16 to 18 rooms a day; a comparable robot once cost upwards of $50,000 before mapping and training.Source
    • Your websiteDeploy page: "The laundry shift, run by our robot." A six-step workflow (move linens, load the washer, washer to dryer, unload the dryer, fold, shelve); "Reduce operating costs by half."; "As low as $2.50/hour" with "Additional terms and conditions may apply"; "Deploying in three states": California, Michigan and Texas. The only call to action is "Schedule a call" (an external booking link); there is no form, FAQ, named hotel or measured result.Source
  • 12

    OS3's training loop records every step a remote operator has to finish and retrains on those failures. That improvement is the core of the product, but hotels never see it. A hotel paying for a service wants proof it is getting better at their laundry room.

    High impactMedium effort First result in 1 month Founder

    First action

    Pull last week's corrections per shift for one property.

    How to measure it

    Properties renewing after the first term

    Now
    –
    Target
    90%
    By
    2 quarters

    Steps

    1. 1

      Define the measures

      Hours, items, corrections and downtime.

      Founder 2 days

    2. 2

      Pull from logs

      Automate the weekly extract.

      Founder 1 week

    3. 3

      Design the report

      One screen, readable by a GM.

      Co-founder 3 days

    4. 4

      Send weekly

      Email to the operations lead.

      Co-founder weekly

    5. 5

      Add the fixes note

      Which failures were trained out.

      Founder monthly

    6. 6

      Aggregate

      A public trend across sites, with permission.

      Co-founder quarterly

    First result
    1 month
    Cost
    Founder time only
    Owner
    Founder

    Smallest test

    Send one hand-built report to one hotel this month.

    Why it works

    Send each property a weekly one-screen report: shifts and hours run, items folded and shelved, remote corrections per shift with a trend line, any downtime and its cause, and attendant hours freed. Add a monthly note on which failure types were fixed. Combined across sites, with permission, it becomes public proof.

    Visible improvement justifies renewals and price, and gives GMs something to show their owners. The data already exists in the training loop.

    Risks and how to handle them

    • A bad week looks alarming

      Show the trend and the cause together.

    Evidence

    • Your websiteTech page: arms 6+1 DoF each, 9 lb nominal and 15 lb peak per arm; telescoping torso from floor to top shelf; omnidirectional base that fits standard doorways; battery "Full workday", charged from a standard outlet; cloud planner plus on-robot policy, no onboard GPU; four demonstration rigs, including a 3D-printable rig and a handheld gripper; training loop in which a remote operator completes steps the robot cannot recover, followed by reinforcement learning on the failures; safety: motion starts only when a person initiates a task, camera and LiDAR collision avoidance, emergency stop, compliant arms. Reach, speed and battery hours are not listed.Source
    • Your websiteDeploy page: "The laundry shift, run by our robot." A six-step workflow (move linens, load the washer, washer to dryer, unload the dryer, fold, shelve); "Reduce operating costs by half."; "As low as $2.50/hour" with "Additional terms and conditions may apply"; "Deploying in three states": California, Michigan and Texas. The only call to action is "Schedule a call" (an external booking link); there is no form, FAQ, named hotel or measured result.Source
    • PressRuntimeWire startup spotlight (September 18, 2026): HALE's first job is towel folding in hotel laundry rooms; the $2.50/hour figure is OS3's own claim and was not independently verified; deployment scale, customer count and performance data are not provided, and OS3's materials do not establish that every step runs autonomously in deployments.Source
  • 13

    The partner page says partners are deploying in wet labs, hospitals and data centers, and the launch post mentions biotech companies, but no partner or workflow is shown. A startup deciding whether to build on HALE wants to know what is proven, what is taken and what is still open.

    Medium impactMedium effort First result in 1 month Co founder

    First action

    Ask each current partner what can be published.

    How to measure it

    Qualified partner calls a month

    Now
    –
    Target
    8
    By
    1 quarter

    Steps

    1. 1

      Ask partners

      Permission for name, logo and workflow.

      Founder 1 week

    2. 2

      Write vertical pages

      Workflow, fit and limits.

      Co-founder 1 week

    3. 3

      List open verticals

      With the payload and reach each would need.

      Founder 2 days

    4. 4

      Add a proposal form

      Vertical, workflow and team.

      Co-founder 1 day

    5. 5

      Publish

      Linked from the partner page.

      Co-founder 1 day

    6. 6

      Review proposals

      A monthly shortlist.

      Founder monthly

    First result
    1 month
    Cost
    Founder time only
    Owner
    Co founder

    Smallest test

    Add one sentence per vertical to the partner page.

    Why it works

    Write a short page for each active vertical: the workflow, why HALE fits (two arms at 9 lb nominal and 15 lb peak each, floor-to-shelf lift, standard doorways), and the partner's name and logo if they agree. Add an "open verticals" list, such as packing, which the footage already shows, and invite proposals for each.

    Concrete examples help a founder picture their own application, and an open-verticals list prompts proposals OS3 would not otherwise get.

    Risks and how to handle them

    • Partners want stealth

      Describe the vertical and workflow without naming them.

    Evidence

    • Your websitePartner page: "Own the application. Skip the robot R&D." OS3 provides HALE 1.0 hardware, control stack, base model, reinforcement-learning pipeline and support; the partner owns the vertical, integrations, customer and go-to-market. Developer unit $14,995 (not returnable once shipped) or Leased Developer $1,800/month (additional lease terms reviewed after reservation), each reserved with a $99 deposit via Stripe. "Partners currently deploying in" wet labs, hospitals and data centers, with no partner named; no SDK or documentation links, no FAQ and no delivery dates.Source
    • PressLaunch YC post (about three months old): OS3 is "already working with hospitality providers and biotech companies"; price targets under $10k for businesses (private beta) and under $15k for developers; a $99 refundable preorder deposit; "Shipments begin Fall 2026"; the technical report is "coming soon"; custom actuators for higher payloads are in development, and businesses with such use cases are invited to get in touch.Source
    • Your websiteFootage page: "The robot, on the floor." Four clips (Move linens, Pack parts, Robot overview, Fold and align), with no label for playback speed, clip length or whether the robot was autonomous or teleoperated, and no metrics.Source
    • Your websiteTech page: arms 6+1 DoF each, 9 lb nominal and 15 lb peak per arm; telescoping torso from floor to top shelf; omnidirectional base that fits standard doorways; battery "Full workday", charged from a standard outlet; cloud planner plus on-robot policy, no onboard GPU; four demonstration rigs, including a 3D-printable rig and a handheld gripper; training loop in which a remote operator completes steps the robot cannot recover, followed by reinforcement learning on the failures; safety: motion starts only when a person initiates a task, camera and LiDAR collision avoidance, emergency stop, compliant arms. Reach, speed and battery hours are not listed.Source
  • 14

    The tech page lists strong safety facts: a person starts every task, camera and LiDAR collision avoidance, emergency stop, compliant arms, and a stopped robot stays stopped. But they sit at the bottom of an engineering page. The Deploy page shows HALE reaching for elevator controls. A hotel's chief engineer, risk lead and owner usually have to sign off, and nothing is written for them.

    Medium impactLow effort First result in 1 month Co founder

    First action

    Copy the safety list from the tech page into a draft sheet.

    How to measure it

    Sales cycle length

    Now
    –
    Target
    -25%
    By
    1 quarter

    Steps

    1. 1

      Draft

      Safety, space, power and data in plain words.

      Co-founder 2 days

    2. 2

      Add the checklist

      What the property must have ready.

      Founder 1 day

    3. 3

      Review

      Check every statement against the robot.

      Founder 1 day

    4. 4

      Publish

      Page and PDF, linked from Deploy.

      Co-founder 1 day

    5. 5

      Send ahead

      Share before every site visit.

      Founder ongoing

    6. 6

      Update

      Add each new question from engineers.

      Co-founder monthly

    First result
    1 month
    Cost
    $0
    Owner
    Co founder

    Smallest test

    Send the draft sheet to the next hotel before its site visit.

    Why it works

    Publish a one-page safety and site-readiness sheet, as a web page and a PDF. It covers how the robot stops and who can start it, the space and doorway needs, charging from a standard outlet, how cameras in back-of-house areas are handled, what OS3 needs from the property (floor space, power, Wi-Fi, elevator access if used) and a site checklist. List any safety testing completed, without claiming certifications OS3 does not hold.

    It hands the operations champion the document their engineer and risk team will ask for, which shortens the approval loop.

    Risks and how to handle them

    • Overstating safety claims

      State only what the robot does, and list testing actually completed.

    Evidence

    • Your websiteTech page: arms 6+1 DoF each, 9 lb nominal and 15 lb peak per arm; telescoping torso from floor to top shelf; omnidirectional base that fits standard doorways; battery "Full workday", charged from a standard outlet; cloud planner plus on-robot policy, no onboard GPU; four demonstration rigs, including a 3D-printable rig and a handheld gripper; training loop in which a remote operator completes steps the robot cannot recover, followed by reinforcement learning on the failures; safety: motion starts only when a person initiates a task, camera and LiDAR collision avoidance, emergency stop, compliant arms. Reach, speed and battery hours are not listed.Source
    • Your websiteDeploy page: "The laundry shift, run by our robot." A six-step workflow (move linens, load the washer, washer to dryer, unload the dryer, fold, shelve); "Reduce operating costs by half."; "As low as $2.50/hour" with "Additional terms and conditions may apply"; "Deploying in three states": California, Michigan and Texas. The only call to action is "Schedule a call" (an external booking link); there is no form, FAQ, named hotel or measured result.Source
  • 15

    Hotel operations leads talk to peers at sister properties, in management companies and in local GM networks. OS3 is deploying in three states, but its only route in is a booking link, with no way for a happy property to bring in another.

    Medium impactLow effort First result in 1 quarter Co founder

    First action

    Set the credit and ask the happiest hotel for one introduction.

    How to measure it

    Signed referrals

    Now
    –
    Target
    3
    By
    2 quarters

    Steps

    1. 1

      Set the credit

      Hours per signed referral.

      Founder 1 day

    2. 2

      Build the share kit

      Report, clip and booking link.

      Co-founder 3 days

    3. 3

      Ask

      At each weekly report review.

      Founder ongoing

    4. 4

      Track

      Flag referrals in bookings.

      Co-founder 1 day

    5. 5

      Pay out

      Credit hours on signing.

      Co-founder per referral

    6. 6

      Thank publicly

      With permission, in the newsletter.

      Co-founder per referral

    First result
    1 quarter
    Cost
    Robot hours credited
    Owner
    Co founder

    Smallest test

    Ask one live hotel for one introduction this month.

    Why it works

    Offer a fixed credit of robot hours to a hotel for each referred property that completes a pilot and signs. Give the referring GM a share kit: their own weekly report (anonymised if they like), a short clip from their laundry room and a direct booking link that flags the referral.

    A peer's real numbers persuade more than vendor claims, and paying in robot hours costs OS3 less than cash.

    Risks and how to handle them

    • Referrals outside service clusters

      Pay the credit only once the new site goes live.

    Builds on

    Evidence

    • Your websiteDeploy page: "The laundry shift, run by our robot." A six-step workflow (move linens, load the washer, washer to dryer, unload the dryer, fold, shelve); "Reduce operating costs by half."; "As low as $2.50/hour" with "Additional terms and conditions may apply"; "Deploying in three states": California, Michigan and Texas. The only call to action is "Schedule a call" (an external booking link); there is no form, FAQ, named hotel or measured result.Source
    • Your websiteHomepage: H1 "US-built semi-humanoid robots and video-trained intelligence"; HALE 1.0, a bimanual mobile robot, "Deployed in hospitality today"; 22 degrees of freedom on OS3-built actuators, 150,000 hours of video pretraining, 2,000+ real-world tasks, 2-4 hours of demonstrations to teach a new task; hospitality deployments from $2.50/hour; developer units $14,995 or $1,800/month with a $99 refundable deposit; calls to action "Deploy at your hotel" and "Partner with OS3"; startup partners mentioned but none named.Source
  • 16

    OS3's one blog post is specific and credible, with measured cuts in jerk, lag and latency, and it previews an "iteration 2" move to hand-pose commands. But it is undated, and the blog has no newsletter or RSS feed, so readers arriving from YC and press attention cannot follow along.

    Medium impactLow effort First result in 1 month Founder

    First action

    Add a date and a signup box to the control-stack post.

    How to measure it

    Email subscribers

    Now
    –
    Target
    1,000
    By
    2 quarters

    Steps

    1. 1

      Date and feed

      Dates on posts and an RSS feed.

      Co-founder 2 hours

    2. 2

      Add signup

      On the blog and footage pages.

      Co-founder 2 hours

    3. 3

      Plan topics

      Six months of posts.

      Founder 1 day

    4. 4

      Write monthly

      One measured post a month.

      Founder monthly

    5. 5

      Add the three paths

      Deploy, partner and careers at the end of each post.

      Co-founder 10 minutes

    6. 6

      Share

      Founder posts on LinkedIn and X.

      Founder monthly

    First result
    1 month
    Cost
    $0
    Owner
    Founder

    Smallest test

    Add the signup today and count sign-ups for two weeks.

    Why it works

    Date the existing post, add RSS and an email signup, and publish one engineering post a month: iteration 2, the data rigs, field lessons from the laundry room, and how corrections fall over time. End each post with three paths: deploy, partner and careers.

    Regular, measured engineering writing builds the credibility a young hardware company needs with developers, investors and recruits.

    Risks and how to handle them

    • Writing slips under build pressure

      Keep posts short and draw on existing logs.

    Evidence

    • Your websiteBlog: one post, "From tokens to torque" (7-minute read, undated): RMS jerk down 75%, motion lag from 33.3 ms to 11.1 ms, command latency from 1.78 ms to 0.2 ms; it says a technical report on training is coming and invites questions by email. The blog has no newsletter signup or RSS feed.Source
    • Your websiteSite check (November 10, 2026): page title "OS3 · US-built robots and video-trained intelligence"; the homepage has no social share image; no newsletter, live chat or contact page; the only published email address appears in the privacy policy and pre-order terms; the domain was registered in May 2026.Source
    • PressY Combinator profile: Summer 2026 batch, team size 2, San Francisco; one-liner "Affordable, intelligent humanoid robots built to deploy at scale"; lists a Forbes feature on YC startups dated September 3, 2026.Source

Big bets 3

Larger plays that can change the trajectory.

  • 18

    Hotels budget laundry per occupied room. Pricing per robot hour leaves the speed risk with the hotel: a slower robot costs them more. TechForce prices per day, aiming to be "cost-neutral on the property". OS3's training loop should make HALE faster over time, and under per-hour pricing those gains go to the hotel as a discount rather than to OS3 as margin.

    High impactHigh effort First result in 1 quarter Founder

    First action

    Model the per-room price from one property's hours and volume.

    How to measure it

    Monthly margin per deployed robot

    Now
    –
    Target
    +30%
    By
    2 quarters

    Steps

    1. 1

      Model it

      Volume, hours and cost per occupied room.

      Founder 1 week

    2. 2

      Set the cap

      A monthly ceiling the hotel can budget.

      Founder 2 days

    3. 3

      Offer the choice

      To the next three hotels.

      Founder 1 month

    4. 4

      Meter volume

      Count items or weight in the workflow.

      Founder 1 month

    5. 5

      Report

      Show volume and price in the weekly report.

      Co-founder weekly

    6. 6

      Compare

      Margin and win rate under each model.

      Co-founder quarterly

    First result
    1 quarter
    Cost
    Founder time; some margin risk early on
    Owner
    Founder

    Smallest test

    Offer one renewing hotel a per-room price.

    Why it works

    Offer new hotels a choice: the hourly rate, or a price per occupied room-night (or per pound processed) with a monthly cap. Use the weekly reports to set the rate so it beats the hotel's current labour cost for the same tasks. As corrections fall and speed rises, OS3's margin grows while the hotel's price stays flat.

    It matches how hotels budget, removes the fear of paying for a slow robot, and lets OS3 keep the value of its own learning curve.

    Risks and how to handle them

    • Early speed is too low to profit

      Start outcome pricing only where corrections are already low.

    • Volume disputes

      Count from robot logs and share them.

    Builds on

    Evidence

    • Your websiteDeploy page: "The laundry shift, run by our robot." A six-step workflow (move linens, load the washer, washer to dryer, unload the dryer, fold, shelve); "Reduce operating costs by half."; "As low as $2.50/hour" with "Additional terms and conditions may apply"; "Deploying in three states": California, Michigan and Texas. The only call to action is "Schedule a call" (an external booking link); there is no form, FAQ, named hotel or measured result.Source
    • Competitor researchTechForce's TIM-E (Hotel Dive, February 25, 2026): a robot that moves linens and waste at a Homewood Suites in Del Mar, California, offered by subscription at roughly $150 a day with setup, training and daily oversight included; the stated goal is to be "cost-neutral on the property"; housekeepers there clean 16 to 18 rooms a day; a comparable robot once cost upwards of $50,000 before mapping and training.Source
    • Your websiteTech page: arms 6+1 DoF each, 9 lb nominal and 15 lb peak per arm; telescoping torso from floor to top shelf; omnidirectional base that fits standard doorways; battery "Full workday", charged from a standard outlet; cloud planner plus on-robot policy, no onboard GPU; four demonstration rigs, including a 3D-printable rig and a handheld gripper; training loop in which a remote operator completes steps the robot cannot recover, followed by reinforcement learning on the failures; safety: motion starts only when a person initiates a task, camera and LiDAR collision avoidance, emergency stop, compliant arms. Reach, speed and battery hours are not listed.Source
  • 19

    HALE already moves linens, fits standard doorways, lifts from floor to top shelf, runs a full workday, and is shown reaching for elevator controls. New tasks take 2-4 hours of demonstrations. Rivals are working on the same building: Dyna targets room turnover, and TechForce moves linens between floors. Every extra task HALE does per day lowers OS3's cost per hour and deepens the account.

    High impactHigh effort First result in 1 quarter Founder

    First action

    Time how attendants spend the hours between laundry cycles at one property.

    How to measure it

    Daily utilisation per robot

    Now
    –
    Target
    +40%
    By
    2 quarters

    Steps

    1. 1

      Observe

      Shadow attendants between cycles.

      Founder 1 week

    2. 2

      Pick the task

      Highest hours, lowest risk.

      Founder 2 days

    3. 3

      Collect demos

      2-4 hours with the data rigs.

      Founder 1 week

    4. 4

      Train and test

      In OS3's workspace first.

      Founder 1 month

    5. 5

      Pilot

      At one hotel, off-peak hours.

      Founder 1 month

    6. 6

      Price and report

      Add it to the report and the offer.

      Co-founder 2 weeks

    First result
    1 quarter
    Cost
    Engineering time over two quarters
    Owner
    Founder

    Smallest test

    Teach one closet-stocking task in OS3's workspace.

    Why it works

    Pilot a second workflow at one hotel: loading clean linen onto housekeeping carts and stocking floor closets between laundry cycles. Teach it from demonstrations, add it to the weekly report, and price it as extra hours or within outcome pricing. Pick the next task by what attendants spend the most time on.

    More useful hours per robot improve unit economics, and a robot that covers two jobs is much harder for a hotel to remove.

    Risks and how to handle them

    • Guest areas raise safety and privacy questions

      Start with back-of-house and service elevators only.

    Evidence

    • Your websiteDeploy page: "The laundry shift, run by our robot." A six-step workflow (move linens, load the washer, washer to dryer, unload the dryer, fold, shelve); "Reduce operating costs by half."; "As low as $2.50/hour" with "Additional terms and conditions may apply"; "Deploying in three states": California, Michigan and Texas. The only call to action is "Schedule a call" (an external booking link); there is no form, FAQ, named hotel or measured result.Source
    • Your websiteTech page: arms 6+1 DoF each, 9 lb nominal and 15 lb peak per arm; telescoping torso from floor to top shelf; omnidirectional base that fits standard doorways; battery "Full workday", charged from a standard outlet; cloud planner plus on-robot policy, no onboard GPU; four demonstration rigs, including a 3D-printable rig and a handheld gripper; training loop in which a remote operator completes steps the robot cannot recover, followed by reinforcement learning on the failures; safety: motion starts only when a person initiates a task, camera and LiDAR collision avoidance, emergency stop, compliant arms. Reach, speed and battery hours are not listed.Source
    • Competitor researchDyna Robotics' homepage: its Dyna-2.1 model automates a long-horizon workflow of washing, drying and folding towels; it targets hotels and housekeeping, warehouses and food service; calls to action include "Explore a pilot" ("You'll believe it when you see it? Deal."), a contact form that asks company size, and a newsletter signup; no customers are named.Source
    • Your websiteHomepage: H1 "US-built semi-humanoid robots and video-trained intelligence"; HALE 1.0, a bimanual mobile robot, "Deployed in hospitality today"; 22 degrees of freedom on OS3-built actuators, 150,000 hours of video pretraining, 2,000+ real-world tasks, 2-4 hours of demonstrations to teach a new task; hospitality deployments from $2.50/hour; developer units $14,995 or $1,800/month with a $99 refundable deposit; calls to action "Deploy at your hotel" and "Partner with OS3"; startup partners mentioned but none named.Source
  • 20

    Three of HALE's six workflow steps happen at the washer and dryer. The dealers who sell and service that equipment already know which hotels run on-site laundry, when machines are being replaced, and the person who signs off. OS3's partner model already lets others own the customer relationship, so a two-person team does not need to build a field sales force.

    High impactHigh effort First result in 1 quarter Founder

    First action

    List the commercial laundry dealers serving one live cluster.

    How to measure it

    Hotels signed through dealers

    Now
    –
    Target
    5
    By
    2 quarters

    Steps

    1. 1

      List dealers

      Those serving hotels near a live site.

      Co-founder 1 week

    2. 2

      Pitch one

      Revenue share and service role.

      Founder 2 weeks

    3. 3

      Train

      Site checks and first-line service.

      Founder 2 weeks

    4. 4

      Co-sell

      Joint visits on machine replacements.

      Founder 1 quarter

    5. 5

      Measure

      Leads, pilots and service response.

      Co-founder monthly

    6. 6

      Expand

      A second dealer in another state.

      Founder next quarter

    First result
    1 quarter
    Cost
    Revenue share and training time
    Owner
    Founder

    Smallest test

    Bring one dealer to a live site demonstration.

    Why it works

    Pilot with one regional commercial laundry dealer near a live cluster. The dealer introduces HALE when quoting new washers and dryers and handles site checks and first-line service. OS3 runs the robot and the remote help. The dealer earns a share of the service revenue for as long as the hotel stays.

    Dealers bring trust, timing and service visits OS3 cannot afford to build, and a robot that loads their machines makes their own offer stronger.

    Risks and how to handle them

    • Dealers see robots as a threat to service income

      Give them the robot service work and a revenue share.

    Builds on

    Evidence

    • Your websiteDeploy page: "The laundry shift, run by our robot." A six-step workflow (move linens, load the washer, washer to dryer, unload the dryer, fold, shelve); "Reduce operating costs by half."; "As low as $2.50/hour" with "Additional terms and conditions may apply"; "Deploying in three states": California, Michigan and Texas. The only call to action is "Schedule a call" (an external booking link); there is no form, FAQ, named hotel or measured result.Source
    • Your websitePartner page: "Own the application. Skip the robot R&D." OS3 provides HALE 1.0 hardware, control stack, base model, reinforcement-learning pipeline and support; the partner owns the vertical, integrations, customer and go-to-market. Developer unit $14,995 (not returnable once shipped) or Leased Developer $1,800/month (additional lease terms reviewed after reservation), each reserved with a $99 deposit via Stripe. "Partners currently deploying in" wet labs, hospitals and data centers, with no partner named; no SDK or documentation links, no FAQ and no delivery dates.Source
    • PressY Combinator profile: Summer 2026 batch, team size 2, San Francisco; one-liner "Affordable, intelligent humanoid robots built to deploy at scale"; lists a Forbes feature on YC startups dated September 3, 2026.Source

Wildcards 2

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

  • 21

    OS3's footage is four short, unlabelled clips, and press has noted that it publishes no performance data. The question every hotel and developer has is not whether HALE can fold a towel, but whether it keeps going for a whole shift.

    Medium impactMedium effort First result in 1 month Co founder

    First action

    Run a private eight-hour test with the counter and correction marker.

    How to measure it

    Bookings from the stream

    Now
    –
    Target
    15
    By
    1 month

    Steps

    1. 1

      Rehearse

      A private full-shift run.

      Founder 1 week

    2. 2

      Build the overlay

      Counter, clock and correction marker.

      Co-founder 3 days

    3. 3

      Announce

      Date and time, a week ahead.

      Co-founder 1 day

    4. 4

      Stream

      Uncut, on three platforms.

      Co-founder 1 day

    5. 5

      Publish totals

      Items, hours and corrections.

      Founder 1 day

    6. 6

      Cut highlights

      For footage and social.

      Co-founder 3 days

    First result
    1 month
    Cost
    Streaming gear and a day of robot time
    Owner
    Co founder

    Smallest test

    Post a one-hour uncut run first.

    Why it works

    Broadcast one full shift live from OS3's workspace (or, with consent, a hotel laundry room with no guests or staff on camera) on YouTube, LinkedIn and X. Show an on-screen counter of items folded and shelved, a clock, and a marker each time a remote operator steps in. Afterwards, publish the totals and cut highlights for the footage page.

    A long, plain, honest video is unusual in robotics. It earns attention and answers the reliability question directly.

    Risks and how to handle them

    • A failure happens live

      Show the correction; that honesty is the point.

    • Privacy on site

      Run it in OS3's workspace or with no people on camera.

    Builds on

    Evidence

    • Your websiteFootage page: "The robot, on the floor." Four clips (Move linens, Pack parts, Robot overview, Fold and align), with no label for playback speed, clip length or whether the robot was autonomous or teleoperated, and no metrics.Source
    • PressRuntimeWire startup spotlight (September 18, 2026): HALE's first job is towel folding in hotel laundry rooms; the $2.50/hour figure is OS3's own claim and was not independently verified; deployment scale, customer count and performance data are not provided, and OS3's materials do not establish that every step runs autonomously in deployments.Source
    • Your websiteHomepage: H1 "US-built semi-humanoid robots and video-trained intelligence"; HALE 1.0, a bimanual mobile robot, "Deployed in hospitality today"; 22 degrees of freedom on OS3-built actuators, 150,000 hours of video pretraining, 2,000+ real-world tasks, 2-4 hours of demonstrations to teach a new task; hospitality deployments from $2.50/hour; developer units $14,995 or $1,800/month with a $99 refundable deposit; calls to action "Deploy at your hotel" and "Partner with OS3"; startup partners mentioned but none named.Source
  • 22

    OS3 is hiring a contract engineer to design parts, enclosures and covers for 3D printing, CNC and sheet metal, with an eye for "covers and finishes". Partners own their vertical, and wet labs, hospitals and data centers will each want different grippers and cleanable covers. Today a partner has no drawing to design against.

    Medium impactMedium effort First result in 1 quarter Mechanical Design Engineer

    First action

    Ask the contractor to draw the wrist mount as a shareable file.

    How to measure it

    Partner end-effectors or covers in the library

    Now
    –
    Target
    5
    By
    2 quarters

    Steps

    1. 1

      Draw the interfaces

      Wrist mount, cover points and keep-out zones.

      Mechanical Design Engineer (contract) 2 weeks

    2. 2

      Set the limits

      Payload, cabling and heat.

      Founder 3 days

    3. 3

      Make references

      One gripper and one cover.

      Mechanical Design Engineer (contract) 3 weeks

    4. 4

      Publish

      With the developer docs.

      Co-founder 1 day

    5. 5

      Review submissions

      Approve partner designs.

      Mechanical Design Engineer (contract) monthly

    6. 6

      Feature

      One partner part per update.

      Co-founder monthly

    First result
    1 quarter
    Cost
    Contractor time
    Owner
    Mechanical Design Engineer (contract)

    Smallest test

    Share the wrist-mount drawing with current depositors.

    Why it works

    Have the new engineer publish a partner interface pack: wrist tool-mount drawings, payload limits (9 lb nominal and 15 lb peak per arm), cable and connector pass-through, cover attachment points and keep-out zones, plus two reference designs. Host a small library where partners can share approved end-effectors and covers.

    It lets partners adapt HALE without waiting on OS3, makes the platform stickier, and gives the new hire's work a public payoff.

    Risks and how to handle them

    • Unsafe partner parts

      Approve designs against the published limits before use.

    Builds on

    Evidence

    • HiringMechanical Design Engineer - Robot Parts & Enclosures (Contract), San Francisco or remote (India), open to new grads and may convert to full-time: design parts, enclosures and covers for 3D printing, CNC and sheet metal; "Package cameras, sensors, compute, actuators, and wiring into serviceable assemblies"; OS3 builds the parts in San Francisco and returns feedback from real robots within days; nice to have: grippers and "industrial design sensibility for covers and finishes".Source
    • Your websitePartner page: "Own the application. Skip the robot R&D." OS3 provides HALE 1.0 hardware, control stack, base model, reinforcement-learning pipeline and support; the partner owns the vertical, integrations, customer and go-to-market. Developer unit $14,995 (not returnable once shipped) or Leased Developer $1,800/month (additional lease terms reviewed after reservation), each reserved with a $99 deposit via Stripe. "Partners currently deploying in" wet labs, hospitals and data centers, with no partner named; no SDK or documentation links, no FAQ and no delivery dates.Source
    • Your websiteTech page: arms 6+1 DoF each, 9 lb nominal and 15 lb peak per arm; telescoping torso from floor to top shelf; omnidirectional base that fits standard doorways; battery "Full workday", charged from a standard outlet; cloud planner plus on-robot policy, no onboard GPU; four demonstration rigs, including a 3D-printable rig and a handheld gripper; training loop in which a remote operator completes steps the robot cannot recover, followed by reinforcement learning on the failures; safety: motion starts only when a person initiates a task, camera and LiDAR collision avoidance, emergency stop, compliant arms. Reach, speed and battery hours are not listed.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.

Have it built

ProgrammableAgency

1 of the 23 ideas are things Programmable builds and runs for its customers: pages, outreach, prospect lists, booking flows and monitors. Pick one, or send the whole list.

Monitors1A live page fed by the public registries and re-checked every day.

A 30-minute session with the engineer who would build it, not a sales call. You leave with a plan whether or not you hire us.

Book a working session

Usually live within days.

Tools for the rest

ProgrammableDigital

Show it, don't just describe it

These ideas need something to show: a demo, a video, an animation, graphics or a deck. Digital turns screenshots or the site into all of them, in the company's own brand.

Make a demo (opens programmable.digital in a new tab)
More example playbooks and reviews

More growth playbooks

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

  • Hygge Box growth strategyA monthly hygge subscription box (the Deluxe box, 5-6 handpicked seasonal items such as candles, tea, treats and home goods) plus an…20 growth ideas
  • Flowspark growth strategyA Dutch no-code automation consultancy that builds workflow automations, AI-driven workflows and AI agents on Zapier, Make, Airtable and…22 growth ideas
  • WITHIN growth strategyAn integrated media and creative agency ("performance branding") running paid media, SEO and AI search, affiliate, influencer, email and…20 growth ideas
  • Sourcebot growth strategyA self-hosted code understanding tool: regex and symbol-aware code search, code navigation, Ask Sourcebot (plain-English questions…22 growth ideas
  • item growth strategyAn AI-native CRM. It brings a company's customer data together from 100+ tools, and an Assistant and document-defined AI agents do the…24 growth ideas
  • ROG Orthodontics growth strategyA doctor-owned orthodontic practice with three offices in Wyomissing, Douglassville and Pottstown, PA, offering braces and clear aligners…21 growth ideas
  • Yazio growth strategyA calorie and nutrition tracking app with AI photo tracking, barcode scanning, intermittent fasting timers, recipes and fitness tracker…21 growth ideas
  • AirCaps AI growth strategyA developer API for speech-to-text with speaker labels (A5S Async), diarization, speaker ID and deployable streaming (A5S Small V2)…23 growth ideas

See every playbook and website review

Independent research. Programmable wrote this playbook from public sources, as they stood on November 10, 2026, and cites them throughout. OS3 did not commission, review or endorse it, and Programmable is not affiliated with OS3. 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 OS3 runs its business.

Something wrong, or do you represent OS3? 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 308 ideas

    Make the offer concrete: the per-hotel quote, labelled footage, one pricing note, depositor updates, the developer FAQ, share images, a careers page and a light speed pass.

    Programmable can ship 1 of these in days.

  2. Days 31 to 609 ideas

    Build proof and pipeline: a measured deployment, developer docs, the open data rig and training report, clustered hotel outreach, the paid pilot, partner verticals, the safety page, referrals and the engineering series.

  3. Days 61 to 906 ideas

    Weekly hotel reports, then commit to the bets: outcome pricing, a second hotel shift, a dealer channel, a live uncut shift and a published tool-mount interface.

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 A laundry shift covered without hiring. The most cited problem: hard, repetitive laundry-room labour and short-staffed shifts.

  • A published hourly price with OS3 running the robot
  • Not knowing what a robot will really cost or do at their property
  • In their words: “laundry shift”

Market

Competitive landscape

3 competitors compared: Dyna Robotics, TechForce Robotics (TIM-E) and Well-funded humanoid makers (Figure, Apptronik, Agility).

  • Gap to exploit: Transparent, per-property pricing for hotel robots
  • Dyna Robotics: No named customers or published price
  • TechForce Robotics (TIM-E): Transport only, no folding or manipulation

Hiring signals

OS3 is a two-person YC team with one open role: a contract Mechanical Design Engineer for robot parts, enclosures and covers, working remotely with parts built in San Francisco and feedback from real robots within days.

  • Mechanical Design Engineer - Robot Parts & Enclosures (Contract): Serviceable assemblies, covers and finishes; parts built in San Francisco with…

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