ProgrammableMarketing/alkali.engineering (opens in a new tab)
PDF

Views

  • Plan
  • Roadmap
  • Research

On this page

  • The brief
  • Start here
  • Idea portfolio
  • Effort vs impact
  • This week · 7
  • This quarter · 11
  • Big bets · 3
  • Wildcards · 1
  • How we can help
Generated Nov 10, 2026
  1. Programmable
  2. Playbooks
  3. Alkali
Alkali logo

growth playbook · Nov 10, 2026

alkali.engineering (opens in a new tab)

Alkali growth playbook

Diagnosis

Alkali is a two-person, Houston-based YC company with strong usage for its size: 2,500+ steel projects across 48 states, project volume that more than doubled each month after launch, and a claimed share of more than 10% of U.S. structural steel. The product does a lot: takeoff, tonnage, revisions, 3D, nesting, collaboration, and now a detailing service.

The website has not kept up. The detailing page, the newest and highest-ticket offer, is missing from the sitemap and can only be bought through an email link. The sitemap lists empty app pages instead. There are no testimonials, case studies, pricing or accuracy figures. The homepage weighs 22.5 MB, while rivals like Steel Genie and Ferra show named proof and comparison pages.

This playbook first fixes the routes to revenue (detailing intake, proof, pricing, a drawing-set offer, page weight). It then builds the content and product loops that a data-rich, two-person team can run: an accuracy benchmark, comparison pages, Bluebeam switchers, share links, monthly shop reports and a takeoff-to-detailing upsell. Finally, it bets on the pre-bid estimate the founders already describe.

Growth Score

7/10

Biggest Opportunity

Turn usage nobody can see into public proof and paid services: the detailing service, the tonnage data and the speed demos are each underused on the site.

Biggest Risk

Steel Genie (backed by SDS2 and ALLPLAN), Ferra and Beam AI all court the same estimators with named testimonials, accuracy claims and free starts. A two-person team that can't show proof could lose on credibility, even with the better product.

01

Start here

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

  1. 1

    Put the detailing page in the sitemap and replace its email link with a bid-set upload form

    Add /detailing to the sitemap and submit it in Search Console today.

    High impact First result in 2 weeks
  2. 2

    Publish how pricing works, since the docs already say features require a subscription

    Write down the plan names, the free limits and the unit of pricing you already use in deals.

    High impact First result in 1 month
  3. 3

    Turn the 21-logo wall into three named fabricator stories with hours and tons

    Email the fabricator behind the "Tesla of estimating" quote and ask to publish it with their name.

    High impact First result in 1 month

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 · 3Big bets · 9Fill-ins · 8Reconsider · 2
Effort →

Showing 22 of 22 ideas

This week 7

Quick wins with a clear first action.

  • 01

    Alkali's detailing service (3D models, Tekla shop drawings, connection design, "days, not weeks") is the highest-ticket thing it sells. Its page, /detailing, is not in the sitemap. The sitemap does list /viewer, /status, /company and /nesting, which share one generic title and show search engines little or no text (/viewer shows only "Loading..."). The page also has no meta description, and both "Set Up Discovery Call" buttons are mailto links. A fabricator who clicks on a work computer without a mail app set up gets nothing, and Alkali never learns they tried.

    High impactLow effort First result in 2 weeks Founder/CEO

    First action

    Add /detailing to the sitemap and submit it in Search Console today.

    How to measure it

    Scoped detailing requests

    Now
    –
    Target
    8 a month
    By
    60 days

    Steps

    1. 1

      Fix the sitemap

      Add /detailing; remove or retitle /viewer, /status and /company.

      Founder/CTO 1 hour

    2. 2

      Write the meta

      Title and description for 3D models, shop drawings and connection design.

      Founder/CEO 1 hour

    3. 3

      Build the form

      Company, bid date, tonnage band, scope and an optional PDF upload.

      Founder/CTO 1 day

    4. 4

      Route it

      Send to the founders' inbox and Crisp with an auto-reply stating the quote window.

      Founder/CTO 2 hours

    5. 5

      Add proof

      Sample shop drawing and the minutes-hours-days timeline next to the form.

      Founder/CEO 2 hours

    6. 6

      Link it

      From the homepage nav, the footer and the in-app export screen.

      Founder/CTO 2 hours

    7. 7

      Request indexing

      Search Console for /detailing and the homepage.

      Founder/CTO 15 minutes

    8. 8

      Track

      Form starts, submissions and quotes sent each week.

      Founder/CEO weekly

    First result
    2 weeks
    Cost
    $0, about two days of founder time
    Owner
    Founder/CEO

    Smallest test

    Add /detailing to the sitemap and swap one button for a three-field form this week.

    Why it works

    Add /detailing to the sitemap with a meta description. Take the app shells (/viewer, /status, /company) out of it, or give them their own titles. Replace both mailto buttons with a short form: company, bid date, tonnage band, scope (model, shop drawings, connection design) and an optional PDF upload. Send each submission to the founders' inbox and Crisp, and reply with a quote window. Add the sample shop drawing and the bid-set-to-fab timeline next to the form, and link /detailing from the takeoff homepage and from inside the app after an export.

    Search engines rank what they are told about and can read. Buyers of services convert more often through a form that asks for the job details than through a blank email. The upload also gives the founders a priced, scoped lead instead of "tell me more".

    Ready to use

    ChecklistDetailing intake form fields

    [ ] Company and estimator name
    [ ] Work email and phone
    [ ] Project location and bid/award date
    [ ] Approximate tonnage (under 50 / 50-250 / 250-1,000 / 1,000+)
    [ ] Scope: 3D model / shop drawings / erection drawings / connection design
    [ ] Upload bid set or paste a link
    [ ] Already estimated in Alkali? (link the project)
    [ ] Auto-reply: "We'll send a quote and turnaround within one business day"

    Risks and how to handle them

    • Large bid-set uploads time out

      Accept a link to the bid set, or a few sheets, as an alternative.

    • More requests than detailers can handle

      Ask for the bid date and quote the turnaround honestly.

    Evidence

    • Your websiteSitemap (16 URLs, checked November 10, 2026): the homepage, /blog, /docs, /nesting, /viewer, /company, /status, /privacy, /terms and seven blog posts, but not /detailing. /nesting, /viewer, /company and /status all carry the same title, "Alkali | AI Steel Takeoff Software", and show a reader little or no text (/viewer shows only "Loading...").Source
    • Your websiteDetailing page: "The smarter way to detail steel." with 3D models, Tekla shop drawings and connection design, "Drawings from bid set to fab shop in days, not weeks" and "Every sheet is checked by a human who has done this before"; both "Set Up Discovery Call" buttons are mailto links (subject "Alkali Detailing Project"); no price, turnaround or file upload; no meta description.Source
    • Your websiteHomepage: H1 "The smarter way to estimate steel."; stat line "2,500+ steel projects | 48 states | >10% of U.S. steel"; a no-signup "Try Here" plus Google, Microsoft and email sign-in; 21 fabricator logos captioned "Anonymized subset of projects estimated with Alkali in 2026", but no testimonials, case studies or pricing; exports to Tekla, Bluebeam, Adobe and Excel; the only contact is a mailto link; the footer reads "© 2025 Alkali Engineering, Inc. · Built in Houston, TX".Source
  • 02

    Google's phone test scores the homepage 33. The page is 22.5 MB over 112 requests, the largest element is the "Final Estimodel cinematic" video, and 1.2 MB of JavaScript goes unused. Real visitors still pass loading (LCP 2.2s) and layout shift, so this is not an emergency. But responsiveness (INP 246 ms) is above the 200 ms "good" line, and estimators often open links on site Wi-Fi or a phone.

    Medium impactLow effort First result in 1 month Founder/CTO

    First action

    Replace the hero video with a poster image and a play button.

    How to measure it

    Field INP (p75)

    Now
    –
    Target
    Under 200 ms
    By
    60 days

    Steps

    1. 1

      Poster the hero

      Static frame and play button; video on click.

      Founder/CTO 2 hours

    2. 2

      Lazy-load media

      Gallery and demo videos load in view only.

      Founder/CTO 3 hours

    3. 3

      Compress images

      Modern formats and correct sizes; about 4.7 MB to save.

      Founder/CTO 3 hours

    4. 4

      Split the code

      Load the 3D viewer bundle only where it is used.

      Founder/CTO 1 day

    5. 5

      Fix accessibility

      Button names, form labels, link text and contrast.

      Founder/CTO 2 hours

    6. 6

      Re-test

      PageSpeed mobile and desktop.

      Founder/CTO 30 minutes

    7. 7

      Watch the field data

      INP in the Chrome UX Report over 28 days.

      Founder/CTO monthly

    First result
    1 month
    Cost
    $0, about two days of engineering
    Owner
    Founder/CTO

    Smallest test

    Swap the hero video for a poster and re-run PageSpeed.

    Why it works

    Show a compressed poster image with a play button in place of the autoplaying cinematic, and load the video only on click or once it scrolls into view. Lazy-load the 3D gallery and demo videos below the fold, convert images to modern formats, and split the 3D viewer code off the homepage. Fix the accessibility items while you are there: name the icon buttons and label the form fields.

    Lighter pages respond faster to the first tap, and a faster sign-in click is the conversion that matters. Accessible names also help assistive tech and automated readers understand the page.

    Risks and how to handle them

    • The cinematic was the hook

      Keep it one click away, with a strong poster frame.

    Evidence

    • PageSpeedHomepage, tested November 10, 2026: mobile score 33 (desktop 57); 22.5 MB over 112 requests, including 8.8 MB of images; the largest element is a video ("Final Estimodel cinematic"); 1.2 MB of unused JavaScript (about 6.1s) and 4.1s of blocking time in the lab. Real visitors (Chrome UX Report, p75) pass LCP (2.2s) and layout shift (0.09), but responsiveness (INP 246 ms) is above the 200 ms "good" line. Accessibility 82: buttons without accessible names and unlabeled form elements.Source
  • 03

    The homepage shows 21 fabricator logos but no testimonial, quote or case study, and calls the logos an "anonymized subset". Yet the founders' own milestones post quotes a fabricator calling Alkali the "Tesla of estimating" and shows 350+ tons taken off in about 15 minutes. Steel Genie publishes named estimator testimonials and a 67% throughput case study. Ferra shows three named testimonials and a customer story. Estimators trust other estimators.

    High impactMedium effort First result in 1 month Founder/CEO

    First action

    Email the fabricator behind the "Tesla of estimating" quote and ask to publish it with their name.

    How to measure it

    Homepage sign-in rate

    Now
    –
    Target
    +20%
    By
    60 days

    Steps

    1. 1

      Pick three shops

      Different sizes and regions, heavy users.

      Founder/CEO 1 day

    2. 2

      Pull their numbers

      Projects and tonnage from the Company Page metrics.

      Founder/CTO 2 hours

    3. 3

      Interview

      20-minute call: before, after, one quote.

      Founder/CEO 1 week

    4. 4

      Write and approve

      One page each, signed off by the shop.

      Founder/CEO 1 week

    5. 5

      Publish

      /customers page and one quote on the homepage.

      Founder/CTO 1 day

    6. 6

      Reuse

      In LinkedIn posts and outreach emails.

      Founder/CEO ongoing

    First result
    1 month
    Cost
    Staff time; optional gift card per story
    Owner
    Founder/CEO

    Smallest test

    Publish one approved quote under the logo wall.

    Why it works

    Ask three of the 350+ fabricators the team has spoken with for a short story: shop size, bid volume before and after, takeoff hours per project, tons estimated in Alkali and one quote with name and title. Publish them on a /customers page and pull one quote into the homepage under the logos. Get written approval for each name and number.

    Named peers with numbers answer the first question a chief estimator asks ("who else uses this, and did it work?") and compete directly with rivals' testimonials.

    Risks and how to handle them

    • Shops don't want competitors to know their tools

      Offer a role-and-region attribution if a full name is refused.

    Evidence

    • Your websiteHomepage: H1 "The smarter way to estimate steel."; stat line "2,500+ steel projects | 48 states | >10% of U.S. steel"; a no-signup "Try Here" plus Google, Microsoft and email sign-in; 21 fabricator logos captioned "Anonymized subset of projects estimated with Alkali in 2026", but no testimonials, case studies or pricing; exports to Tekla, Bluebeam, Adobe and Excel; the only contact is a mailto link; the footer reads "© 2025 Alkali Engineering, Inc. · Built in Houston, TX".Source
    • Your website"Alkali Early Milestones April 2026" (May 2, 2026): project volume "more than doubled" each month since launch; over 25,000 tons estimated in April, about 7% of average monthly U.S. structural steel volume; the same project is sometimes uploaded by several customers; 350+ fabricators spoken with in a year; a Baltimore video shows 350+ tons taken off in about 15 minutes; a fabricator called it the "Tesla of estimating"; coverage "still sparse in many regions"; the goal of an estimate ready before a fabricator starts bidding, tuned to each shop's labor rates.Source
    • Competitor researchSteel Genie from SDS2 by ALLPLAN: "Up to 10x" faster takeoffs, a case study claiming 67% higher estimating throughput, named estimator testimonials, "Book Demo" only, no published price or free trial.Source
    • Competitor researchFerra: bid intelligence, steel takeoff, revision control and 3D models; "Start for free"; a co-founder's "Send me a drawing set" offer; 10 customer logos and 3 named testimonials; a /compare page against SketchDeck, Steel Genie, Beam AI and SteelFlo that does not mention Alkali.Source
  • 04

    Alkali's docs say customization "requires a subscription", but the site has no pricing page and the homepage mentions no plan. Steel Genie is demo-only with no published price. Fabricators who like the free "Try Here" demo have no idea what the next step costs, or whether it is priced per seat, per project or per ton.

    High impactLow effort First result in 1 month Founder/CEO

    First action

    Write down the plan names, the free limits and the unit of pricing you already use in deals.

    How to measure it

    Free-to-paid conversion

    Now
    –
    Target
    +25%
    By
    90 days

    Steps

    1. 1

      List current deals

      What each paying shop pays and for what.

      Founder/CEO 1 day

    2. 2

      Choose the unit

      Per seat, per project or per ton; one clear rule.

      Founder/CEO 1 day

    3. 3

      Define free vs paid

      Free scans, paid features, team plan.

      Founder/CEO 1 day

    4. 4

      Build the page

      /pricing with an FAQ and a detailing section.

      Founder/CTO 1 day

    5. 5

      Add upgrade prompts

      In-app where a paid feature is hit.

      Founder/CTO 2 days

    6. 6

      Measure

      Pricing views to paid, in PostHog.

      Founder/CTO monthly

    First result
    1 month
    Cost
    $0
    Owner
    Founder/CEO

    Smallest test

    Publish a short "How pricing works" page with the unit and free limits.

    Why it works

    Publish a /pricing page with what is free (the no-signup viewer, a set number of scans), what a subscription adds (customization, team seats, exports, nesting, 3D) and how detailing is quoted (per project, by tonnage and scope). If exact prices aren't ready, publish a "starts at" figure or the pricing unit and a "talk to us for teams" path. Link it from the homepage nav and in-app upgrade prompts.

    Small and mid-size fabricators buy faster when they can check budget fit alone. Being the transparent option in a demo-only category is a positioning win.

    Risks and how to handle them

    • Locks in a price too early

      Mark it as current pricing and grandfather early customers.

    Evidence

    • Your websiteDocs: exports (CSV, PDF markup, IFC 3D Estimodel, nesting plans); a "Bluebeam Desktop" controls preset; customization "requires a subscription"; a Company Page with team usage metrics (projects uploaded, tonnage analyzed); 3D Estimodeling in Beta/Pilot; share links with Owner, Editor and Viewer roles.Source
    • ResearchSite stack: Next.js on Railway with Google Analytics, PostHog and Crisp chat; no pricing page, no newsletter and no YouTube channel; social presence is LinkedIn and GitHub; domain registered May 2025.Source
    • Competitor researchSteel Genie from SDS2 by ALLPLAN: "Up to 10x" faster takeoffs, a case study claiming 67% higher estimating throughput, named estimator testimonials, "Book Demo" only, no published price or free trial.Source
  • 05

    The sitemap lists /nesting, but to a reader it is an empty page with the generic site title. The "Intro to Nesting" guide explains why waste can cut "a few thousand dollars" from a small job's margin, yet its buttons point to the homepage, not the nesting tool. Search visitors who arrive with a nesting problem are never shown the nesting answer.

    Medium impactLow effort First result in 1 month Founder/CTO

    First action

    Change the guide's two buttons to point at /nesting.

    How to measure it

    Organic nesting runs

    Now
    –
    Target
    50 a month
    By
    90 days

    Steps

    1. 1

      Relink the guide

      Both buttons to /nesting.

      Founder/CTO 15 minutes

    2. 2

      Write the page

      300-500 words, example and FAQ.

      Founder/CEO 1 day

    3. 3

      Render it

      Text served in the HTML, not only after the app loads.

      Founder/CTO 1 day

    4. 4

      Title and meta

      Unique for nesting.

      Founder/CTO 15 minutes

    5. 5

      Add a sample

      Downloadable cut-list PDF.

      Founder/CTO 2 hours

    6. 6

      Track

      Nesting runs by source in PostHog.

      Founder/CTO monthly

    First result
    1 month
    Cost
    $0
    Owner
    Founder/CTO

    Smallest test

    Relink the guide today and watch clicks to /nesting.

    Why it works

    Add server-rendered text to /nesting: what it does (stock-aware patterns, kerf, end trim, nest across jobs), a before-and-after drop example, a sample cut-list PDF and a "nest your cut list free" button. Give it its own title and meta description. Point both buttons in the guide to /nesting, and link it from the homepage nesting section.

    A dedicated, readable page can rank for nesting searches, and a reader who just learned about kerf and drop is primed to try the tool.

    Risks and how to handle them

    • Page text and app share a route

      Show the landing text to signed-out visitors only.

    Evidence

    • Your website"Alkali Guide to Nesting" (March 2, 2026): stock lengths, kerf, end trim and nesting across jobs, noting waste can cut "a few thousand dollars" from a small job's margin; its buttons go to the homepage, not the /nesting tool.Source
    • Your websiteSitemap (16 URLs, checked November 10, 2026): the homepage, /blog, /docs, /nesting, /viewer, /company, /status, /privacy, /terms and seven blog posts, but not /detailing. /nesting, /viewer, /company and /status all carry the same title, "Alkali | AI Steel Takeoff Software", and show a reader little or no text (/viewer shows only "Loading...").Source
  • 06

    A project loads in about 30 seconds, and the Baltimore video shows 350+ tons taken off in about 15 minutes. Ferra's homepage has a co-founder's "Send me a drawing set" offer, and Beam AI sells done-for-you takeoffs in 24-72 hours. Alkali asks a busy estimator to learn a new tool on their own first.

    High impactLow effort First result in 1 week Founder/CEO

    First action

    Add the offer as a line under the homepage hero with a simple upload form.

    How to measure it

    Submission-to-active rate

    Now
    –
    Target
    40%
    By
    60 days

    Steps

    1. 1

      Write the offer

      One line, one promise, same day.

      Founder/CEO 1 hour

    2. 2

      Add the form

      Email, bid date, PDF or link.

      Founder/CTO 4 hours

    3. 3

      Run the takeoff

      Founders process and check it.

      Founder/CEO per request

    4. 4

      Share the result

      Project link, tonnage, short recording.

      Founder/CEO per request

    5. 5

      Follow up

      Call two days later about the bid.

      Founder/CEO per request

    6. 6

      Track

      Submissions, sign-ins, paid.

      Founder/CTO weekly

    First result
    1 week
    Cost
    Founder time per request
    Owner
    Founder/CEO

    Smallest test

    Make the offer in one LinkedIn post and see how many sets arrive.

    Why it works

    Add a "Send us a live bid" form beside "Try Here". The founders run the takeoff and return a shared project link, tonnage and a two-minute screen recording, the same business day. The estimator gets the result inside Alkali, already shared to their email, so the next step is to sign in and check it.

    Seeing their own job done removes the biggest doubt ("will it work on my drawings?") and gets the account created around real work.

    Risks and how to handle them

    • Too many requests for two founders

      Cap it per day and promise next-day once full.

    Evidence

    • Your website"Alkali Early Milestones April 2026" (May 2, 2026): project volume "more than doubled" each month since launch; over 25,000 tons estimated in April, about 7% of average monthly U.S. structural steel volume; the same project is sometimes uploaded by several customers; 350+ fabricators spoken with in a year; a Baltimore video shows 350+ tons taken off in about 15 minutes; a fabricator called it the "Tesla of estimating"; coverage "still sparse in many regions"; the goal of an estimate ready before a fabricator starts bidding, tuned to each shop's labor rates.Source
    • Competitor researchFerra: bid intelligence, steel takeoff, revision control and 3D models; "Start for free"; a co-founder's "Send me a drawing set" offer; 10 customer logos and 3 named testimonials; a /compare page against SketchDeck, Steel Genie, Beam AI and SteelFlo that does not mention Alkali.Source
    • Competitor researchBeam AI: multi-trade takeoff including structural steel, done-for-you takeoffs in 24-72 hours, "99%+ accuracy", "3X your bidding capacity" and customer hours-saved figures.Source
  • 07

    The homepage claims ">10% of U.S. steel". The YC job post says about 10% by volume. The April milestones post works it out as about 7% of monthly volume and calls that rough. The footer still reads "© 2025". Estimators check numbers for a living; one that changes between pages invites doubt about the takeoff numbers too.

    Medium impactLow effort First result in 1 week Founder/CEO

    First action

    Recalculate the share from the latest full month and set one figure.

    How to measure it

    Pages with the sourced figure

    Now
    –
    Target
    All
    By
    2 weeks

    Steps

    1. 1

      Recalculate

      Last full month's tonnage against the benchmark.

      Founder/CTO 2 hours

    2. 2

      Write the note

      Method and caveats in 150 words.

      Founder/CEO 1 hour

    3. 3

      Update pages

      Homepage stat, blog and job posts.

      Founder/CTO 1 hour

    4. 4

      Fix the footer

      Year set automatically.

      Founder/CTO 15 minutes

    5. 5

      Announce

      A LinkedIn milestone post.

      Founder/CEO 1 hour

    6. 6

      Refresh

      Each quarter.

      Founder/CEO quarterly

    First result
    1 week
    Cost
    $0
    Owner
    Founder/CEO

    Smallest test

    Fix the footer and add a source link under the stat today.

    Why it works

    Publish a short methodology note (tons estimated in the last full month and the AISC benchmark used), link it from the homepage stat, and use the same figure everywhere. Update the footer year automatically. Post the updated number as a milestone on LinkedIn.

    A sourced, consistent figure turns a claim that invites doubt into proof, and a current footer signals an active company.

    Risks and how to handle them

    • The careful number is lower

      Lead with tons and projects, which are larger and exact.

    Evidence

    • Your websiteHomepage: H1 "The smarter way to estimate steel."; stat line "2,500+ steel projects | 48 states | >10% of U.S. steel"; a no-signup "Try Here" plus Google, Microsoft and email sign-in; 21 fabricator logos captioned "Anonymized subset of projects estimated with Alkali in 2026", but no testimonials, case studies or pricing; exports to Tekla, Bluebeam, Adobe and Excel; the only contact is a mailto link; the footer reads "© 2025 Alkali Engineering, Inc. · Built in Houston, TX".Source
    • Your website"Alkali Early Milestones April 2026" (May 2, 2026): project volume "more than doubled" each month since launch; over 25,000 tons estimated in April, about 7% of average monthly U.S. structural steel volume; the same project is sometimes uploaded by several customers; 350+ fabricators spoken with in a year; a Baltimore video shows 350+ tons taken off in about 15 minutes; a fabricator called it the "Tesla of estimating"; coverage "still sparse in many regions"; the goal of an estimate ready before a fabricator starts bidding, tuned to each shop's labor rates.Source
    • HiringFounding Applied AI Engineer (New York; $150K–$220K, 0.75%–1.50% equity) on Y Combinator's job board: improve vision models that extract quantities and sizes from drawings, build agents that navigate large plan sets, and build datasets, evaluation pipelines and test harnesses. The company description says about 10% of U.S. structural steel by volume flows through Alkali. YC lists Alkali as a Spring 2025 company in Houston with a team of 2; the role was also announced on the Alkali blog on July 20, 2026.Source

This quarter 11

The core plays for the next three months.

  • 08

    The beam scanner post shows a predicted-versus-actual tonnage chart across 50 projects but gives no accuracy figure, and the column scanner says "Results should be verified by an estimator!" Beam AI advertises "99%+ accuracy". The open Founding Applied AI Engineer role is hired to build "evaluation pipelines and test harnesses", which is exactly what a public benchmark needs.

    High impactMedium effort First result in 6 weeks Founding Applied AI Engineer

    First action

    Turn the 50-project chart into a median tonnage-error figure.

    How to measure it

    Accuracy-page sign-ins

    Now
    –
    Target
    10% of visitors
    By
    1 quarter

    Steps

    1. 1

      Pick the test set

      Held-out projects with checked takeoffs.

      Founding Applied AI Engineer 1 week

    2. 2

      Define metrics

      Tonnage error, count recall and precision by element.

      Founding Applied AI Engineer 2 days

    3. 3

      Run it

      Automated in the release pipeline.

      Founding Applied AI Engineer 2 weeks

    4. 4

      Write the page

      Results, limits and where to review.

      Founder/CEO 3 days

    5. 5

      Publish the protocol

      How a shop can test its own past job.

      Founder/CTO 2 days

    6. 6

      Update each release

      Changelog of accuracy.

      Founding Applied AI Engineer per release

    First result
    6 weeks
    Cost
    Engineering time
    Owner
    Founding Applied AI Engineer

    Smallest test

    Publish the beam-scanner median error with the existing chart.

    Why it works

    Use the evaluation harness to publish a /accuracy page: tonnage error and member-count recall by element type (beams, columns, joists, bracing, base plates) on a held-out set of real projects, updated each release. Show where human review is still needed. Offer the test protocol so a shop can check it on its own past job.

    Estimators buy on trust in the numbers. An honest, element-by-element benchmark beats a single headline percentage and gives the sales conversation something to point to.

    Risks and how to handle them

    • Numbers look weaker than rivals' claims

      Explain the method; rivals' single figures rarely show one.

    Builds on

    Evidence

    • Your website"Steel Estimation with Beam Sweeper" (October 10, 2025): results in about 30 seconds, first data under 10 seconds; statistics across 50 projects with per-sheet tonnage of about 1 to 150 tons; a predicted-versus-ground-truth tonnage chart but no accuracy figure in the text. The Column Scanner (January 31, 2026) is a pilot: "Results should be verified by an estimator!"Source
    • Competitor researchBeam AI: multi-trade takeoff including structural steel, done-for-you takeoffs in 24-72 hours, "99%+ accuracy", "3X your bidding capacity" and customer hours-saved figures.Source
    • HiringFounding Applied AI Engineer (New York; $150K–$220K, 0.75%–1.50% equity) on Y Combinator's job board: improve vision models that extract quantities and sizes from drawings, build agents that navigate large plan sets, and build datasets, evaluation pipelines and test harnesses. The company description says about 10% of U.S. structural steel by volume flows through Alkali. YC lists Alkali as a Spring 2025 company in Houston with a team of 2; the role was also announced on the Alkali blog on July 20, 2026.Source
  • 09

    Ferra has a /compare page against SketchDeck, Steel Genie, Beam AI and SteelFlo that doesn't mention Alkali, so estimators researching options hear the field described without Alkali in it. Alkali's only comparison post is against a general chat assistant, not the steel tools it actually competes with.

    High impactMedium effort First result in 6 weeks Founder/CEO

    First action

    List the rivals prospects mention most on calls.

    How to measure it

    Comparison-page sign-ins

    Now
    –
    Target
    40 a month
    By
    90 days

    Steps

    1. 1

      Pick rivals

      The four most named on calls.

      Founder/CEO 1 day

    2. 2

      Check facts

      Each rival's own pages, dated.

      Founder/CEO 2 days

    3. 3

      Write

      Fair tables and where each fits best.

      Founder/CEO 1 week

    4. 4

      Review claims

      Trademarks and accuracy.

      Founder/CTO 2 days

    5. 5

      Publish

      /compare hub and four pages in the sitemap.

      Founder/CTO 1 day

    6. 6

      Refresh

      Each quarter.

      Founder/CEO quarterly

    First result
    6 weeks
    Cost
    Staff time
    Owner
    Founder/CEO

    Smallest test

    Publish "Alkali vs manual takeoff" first.

    Why it works

    Write four fair, dated pages: Alkali vs Steel Genie, vs Ferra, vs Beam AI, and vs manual takeoff in a PDF tool. Each covers the free start, pricing visibility, element coverage, revisions, 3D/IFC, nesting, detailing and collaboration, with a "try your own bid set" button. Cite each rival's own site.

    Comparison searches come from buyers close to a decision, and the pages also shape how search and AI assistants describe the category.

    Risks and how to handle them

    • Out-of-date claims about rivals

      Date every page and link the source.

    Evidence

    • Competitor researchFerra: bid intelligence, steel takeoff, revision control and 3D models; "Start for free"; a co-founder's "Send me a drawing set" offer; 10 customer logos and 3 named testimonials; a /compare page against SketchDeck, Steel Genie, Beam AI and SteelFlo that does not mention Alkali.Source
    • Competitor researchSteel Genie from SDS2 by ALLPLAN: "Up to 10x" faster takeoffs, a case study claiming 67% higher estimating throughput, named estimator testimonials, "Book Demo" only, no published price or free trial.Source
    • Competitor researchBeam AI: multi-trade takeoff including structural steel, done-for-you takeoffs in 24-72 hours, "99%+ accuracy", "3X your bidding capacity" and customer hours-saved figures.Source
    • Your websiteBlog index: seven posts between October 2025 and July 2026 (beam scanner, takeoff how-to, column detection, nesting guide, milestones, a comparison with a general-purpose chat assistant, a hiring post); nothing published since July 20, 2026; no newsletter signup.Source
  • 10

    The docs include a "Bluebeam Desktop" navigation preset, and the homepage says Alkali exports PDFs and material lists to Bluebeam. That is a ready-made switching story for the many estimators who do takeoff by hand in Bluebeam today. The site never says it.

    Medium impactLow effort First result in 1 month Founder/CEO

    First action

    Record a one-sheet takeoff in Bluebeam and in Alkali, side by side.

    How to measure it

    Bluebeam-preset activations

    Now
    –
    Target
    30% of new users
    By
    90 days

    Steps

    1. 1

      Record the comparison

      One framing sheet, both tools.

      Founder/CEO 1 day

    2. 2

      Write the page

      Preset, export, time saved.

      Founder/CEO 2 days

    3. 3

      Ask at sign-up

      "What do you use today?"

      Founder/CTO 1 day

    4. 4

      Default the preset

      For Bluebeam answers.

      Founder/CTO 1 day

    5. 5

      Promote

      LinkedIn and estimator forums.

      Founder/CEO 2 weeks

    6. 6

      Measure

      Activation by prior tool.

      Founder/CTO monthly

    First result
    1 month
    Cost
    $0
    Owner
    Founder/CEO

    Smallest test

    Post the side-by-side video on LinkedIn.

    Why it works

    Publish a page for Bluebeam users: "Keep your shortcuts, stop measuring by hand". Show the preset, the markup export back into Bluebeam, a side-by-side time comparison on one sheet, and a 3-minute setup video. Make the preset the default when a user says they come from Bluebeam at sign-up.

    Lowering the cost of switching (muscle memory and file habits) is often what gets a reluctant estimator to try a new tool.

    Risks and how to handle them

    • Trademark use

      Use the name descriptively and avoid implying partnership.

    Evidence

    • Your websiteDocs: exports (CSV, PDF markup, IFC 3D Estimodel, nesting plans); a "Bluebeam Desktop" controls preset; customization "requires a subscription"; a Company Page with team usage metrics (projects uploaded, tonnage analyzed); 3D Estimodeling in Beta/Pilot; share links with Owner, Editor and Viewer roles.Source
    • Your websiteHomepage: H1 "The smarter way to estimate steel."; stat line "2,500+ steel projects | 48 states | >10% of U.S. steel"; a no-signup "Try Here" plus Google, Microsoft and email sign-in; 21 fabricator logos captioned "Anonymized subset of projects estimated with Alkali in 2026", but no testimonials, case studies or pricing; exports to Tekla, Bluebeam, Adobe and Excel; the only contact is a mailto link; the footer reads "© 2025 Alkali Engineering, Inc. · Built in Houston, TX".Source
  • 11

    The no-signup "Try Here" opens a sample project, and every blog post sends readers to the viewer. The beam scanner returns results in about 30 seconds. The value only shows on the estimator's own drawings, though, and the site has PostHog installed to measure this step.

    High impactMedium effort First result in 1 month Founder/CTO

    First action

    Build a PostHog funnel: sample opened, own PDF uploaded, scan run, signed in.

    How to measure it

    Sample-to-own-scan rate

    Now
    –
    Target
    25%
    By
    60 days

    Steps

    1. 1

      Build the funnel

      Four events in PostHog.

      Founder/CTO 1 day

    2. 2

      Add the prompt

      Upload card in the sample project.

      Founder/CTO 2 days

    3. 3

      Auto-scan

      Beams preselected on upload.

      Founder/CTO 3 days

    4. 4

      Gate the save

      Sign in to keep results.

      Founder/CTO 1 day

    5. 5

      Rescue emails

      Upload but no return in 48 hours.

      Founder/CEO 2 days

    6. 6

      Review weekly

      Where users drop off.

      Founder/CTO weekly

    First result
    1 month
    Cost
    Engineering time
    Owner
    Founder/CTO

    Smallest test

    Add the upload card to the sample for one week.

    Why it works

    After two minutes in the sample, prompt: "Drop your current bid set; we'll scan the framing plans while you watch." Pre-select Beams, run the scan automatically, and ask for sign-in only to save the results. Track time from first visit to first own-project scan in PostHog, and email anyone who uploads but doesn't return.

    The moment their own tonnage appears is when a sample user becomes a prospect. Cutting the steps to it raises activation.

    Risks and how to handle them

    • Confidential drawings

      State how uploads are stored and who can see them.

    Evidence

    • Your websiteHomepage: H1 "The smarter way to estimate steel."; stat line "2,500+ steel projects | 48 states | >10% of U.S. steel"; a no-signup "Try Here" plus Google, Microsoft and email sign-in; 21 fabricator logos captioned "Anonymized subset of projects estimated with Alkali in 2026", but no testimonials, case studies or pricing; exports to Tekla, Bluebeam, Adobe and Excel; the only contact is a mailto link; the footer reads "© 2025 Alkali Engineering, Inc. · Built in Houston, TX".Source
    • Your website"Steel Estimation with Beam Sweeper" (October 10, 2025): results in about 30 seconds, first data under 10 seconds; statistics across 50 projects with per-sheet tonnage of about 1 to 150 tons; a predicted-versus-ground-truth tonnage chart but no accuracy figure in the text. The Column Scanner (January 31, 2026) is a pilot: "Results should be verified by an estimator!"Source
    • ResearchSite stack: Next.js on Railway with Google Analytics, PostHog and Crisp chat; no pricing page, no newsletter and no YouTube channel; social presence is LinkedIn and GitHub; domain registered May 2025.Source
  • 12

    The product already tracks team usage on a Company Page: projects uploaded, tonnage analyzed and a performance metric. Estimators log in to it, but the shop owner who pays the subscription may never see it.

    Medium impactLow effort First result in 1 month Founder/CTO

    First action

    Send ten hand-built reports to the most active shops and ask what they'd change.

    How to measure it

    Renewal rate

    Now
    –
    Target
    90%+
    By
    2 quarters

    Steps

    1. 1

      Pick the numbers

      Projects, tons, hours, revisions.

      Founder/CEO 1 day

    2. 2

      Hand-build ten

      Test with active shops.

      Founder/CEO 2 days

    3. 3

      Automate

      Monthly from the metrics data.

      Founder/CTO 1 week

    4. 4

      Add the invite

      One-click seat invite.

      Founder/CTO 1 day

    5. 5

      Add detailing

      A line for awarded jobs.

      Founder/CEO 1 hour

    6. 6

      Measure

      Opens, invites, renewals.

      Founder/CTO monthly

    First result
    1 month
    Cost
    $0-50 a month for email
    Owner
    Founder/CTO

    Smallest test

    Email five owners their tonnage this month.

    Why it works

    Send a monthly email to the account owner (and an opt-in for the shop's president): projects estimated, tons analyzed, estimated hours saved at the shop's own rate, and revisions caught. Include a one-click invite for another estimator and a line on detailing for jobs they won.

    Showing the buyer the value every month lowers churn, and the invite and detailing lines turn the report into expansion.

    Risks and how to handle them

    • Hours-saved estimate questioned

      Let each shop set its own baseline.

    Evidence

    • Your websiteDocs: exports (CSV, PDF markup, IFC 3D Estimodel, nesting plans); a "Bluebeam Desktop" controls preset; customization "requires a subscription"; a Company Page with team usage metrics (projects uploaded, tonnage analyzed); 3D Estimodeling in Beta/Pilot; share links with Owner, Editor and Viewer roles.Source
  • 13

    Alkali lets estimators share projects by email invite or public link, with Viewer roles. Those links reach project managers, erectors, detailers and other fabricators' estimators: people who do takeoffs or hire for them. Today a viewer has no prompt to try it on their own job.

    Medium impactLow effort First result in 1 month Founder/CTO

    First action

    Count how many public share links were opened last month by people outside the account.

    How to measure it

    Share-link sign-ups

    Now
    –
    Target
    15% of new accounts
    By
    1 quarter

    Steps

    1. 1

      Measure shares

      Links opened by outside viewers.

      Founder/CTO 1 day

    2. 2

      Add the banner

      Dismissible, on shared views.

      Founder/CTO 1 day

    3. 3

      Tag referrals

      Source the sharing account.

      Founder/CTO 1 day

    4. 4

      Set the reward

      Free month or seat per paid referral.

      Founder/CEO 1 day

    5. 5

      Let shops opt out

      For white-label sharing with clients.

      Founder/CTO 1 day

    6. 6

      Track

      Sign-ups and paid from shares.

      Founder/CTO monthly

    First result
    1 month
    Cost
    $0 plus reward credits
    Owner
    Founder/CTO

    Smallest test

    Show the banner on new share links for two weeks.

    Why it works

    Add a slim banner to shared views: "Estimated with Alkali: try it on your own bid set". Link it to the no-signup demo with a referral tag. Credit the sharing shop (an extra month or seat) when a viewer becomes a paying account. Track shares and conversions in PostHog.

    Every shared estimate becomes a product demo seen by a qualified peer, at no cost.

    Risks and how to handle them

    • Shops don't want branding on client links

      Make it switch-off-able on paid plans.

    Evidence

    • Your websiteHomepage product sections: AI takeoff assist for beams, columns, joists, braces and base plates; tonnage from the AISC manual; revision compare across addenda with Accept/Reject (example: a beam changed from W12x16 to W16x26 in Addendum 2); drawing Q&A including "What does a 20KSP joist mean?" for new estimators; real-time collaboration and public share links; "Alkali AI gets better with every project".Source
    • Your websiteDocs: exports (CSV, PDF markup, IFC 3D Estimodel, nesting plans); a "Bluebeam Desktop" controls preset; customization "requires a subscription"; a Company Page with team usage metrics (projects uploaded, tonnage analyzed); 3D Estimodeling in Beta/Pilot; share links with Owner, Editor and Viewer roles.Source
  • 14

    The founders write that coverage is "still sparse in many regions" and it is still rare for two shops to bid on the same project, while some projects are already uploaded by several customers. The homepage says 48 states. Usage data shows exactly where Alkali is strong and where it isn't.

    High impactMedium effort First result in 6 weeks Founder/CEO

    First action

    Export project count by state from the last six months.

    How to measure it

    Target-region new accounts

    Now
    –
    Target
    15
    By
    1 quarter

    Steps

    1. 1

      Map coverage

      Projects by state.

      Founder/CTO 1 day

    2. 2

      Pick regions

      Three with high bidding and low coverage.

      Founder/CEO 1 day

    3. 3

      Build lists

      Fabricators in each region from public directories.

      Founder/CEO 1 week

    4. 4

      Write outreach

      Regional proof and the bid-set offer.

      Founder/CEO 2 days

    5. 5

      Show up

      One regional association meeting.

      Founder/CEO 1 quarter

    6. 6

      Measure

      Accounts and projects by region.

      Founder/CTO monthly

    First result
    6 weeks
    Cost
    Travel and staff time
    Owner
    Founder/CEO

    Smallest test

    Email 30 fabricators in one under-covered state.

    Why it works

    Map uploaded projects by state against where steel is being bid. Pick three under-covered regions. Run founder-led outreach to fabricators there, with a regional proof point ("X projects in your state estimated in Alkali"), LinkedIn posts aimed at estimators in those areas, and a visit to a regional fabricator association meeting.

    Targeting gaps is cheaper than competing in saturated areas. Regional density also feeds the network effect the founders describe, where shops increasingly bid the same jobs.

    Risks and how to handle them

    • Founder time is scarce

      Start with one region and email-only outreach.

    Builds on

    Evidence

    • Your website"Alkali Early Milestones April 2026" (May 2, 2026): project volume "more than doubled" each month since launch; over 25,000 tons estimated in April, about 7% of average monthly U.S. structural steel volume; the same project is sometimes uploaded by several customers; 350+ fabricators spoken with in a year; a Baltimore video shows 350+ tons taken off in about 15 minutes; a fabricator called it the "Tesla of estimating"; coverage "still sparse in many regions"; the goal of an estimate ready before a fabricator starts bidding, tuned to each shop's labor rates.Source
    • Your websiteHomepage: H1 "The smarter way to estimate steel."; stat line "2,500+ steel projects | 48 states | >10% of U.S. steel"; a no-signup "Try Here" plus Google, Microsoft and email sign-in; 21 fabricator logos captioned "Anonymized subset of projects estimated with Alkali in 2026", but no testimonials, case studies or pricing; exports to Tekla, Bluebeam, Adobe and Excel; the only contact is a mailto link; the footer reads "© 2025 Alkali Engineering, Inc. · Built in Houston, TX".Source
  • 15

    The homepage shows the drawing chat answering "What does a 20KSP joist mean?" for new estimators, plus questions on embedment, footing depth and galvanizing. Those are exactly the questions junior estimators search for. The blog has seven posts and nothing new since July.

    Medium impactMedium effort First result in 2 months Founder/CEO

    First action

    List the 50 most common chat question topics from the last quarter.

    How to measure it

    Glossary visits

    Now
    –
    Target
    2,000 a month
    By
    2 quarters

    Steps

    1. 1

      Mine questions

      Top topics from chat, anonymised.

      Founding Applied AI Engineer 2 days

    2. 2

      Template

      Definition, drawing snippet, button.

      Founder/CTO 2 days

    3. 3

      Write 20

      Checked by an experienced estimator.

      Founder/CEO 2 weeks

    4. 4

      Publish

      Server-rendered, in the sitemap.

      Founder/CTO 1 day

    5. 5

      Add weekly

      Five new entries.

      Founder/CEO weekly

    6. 6

      Measure

      Visits and demo clicks per entry.

      Founder/CTO monthly

    First result
    2 months
    Cost
    Writing time; optional freelance estimator review
    Owner
    Founder/CEO

    Smallest test

    Publish ten entries and check impressions after a month.

    Why it works

    Publish a /learn glossary of 50 short, illustrated entries: joist designations, camber, studs, base plates, embeds, kerf, copes, AISC shapes. Each entry has a drawing snippet and a "find this on your drawings" button that opens the demo with the chat. Add five entries a week from the most common real chat questions (anonymised).

    Many small, precise answers rank for long-tail searches and catch estimators early in their careers, when tool habits form.

    Risks and how to handle them

    • Thin or wrong entries

      Every entry reviewed by someone who has estimated steel.

    Evidence

    • Your websiteHomepage product sections: AI takeoff assist for beams, columns, joists, braces and base plates; tonnage from the AISC manual; revision compare across addenda with Accept/Reject (example: a beam changed from W12x16 to W16x26 in Addendum 2); drawing Q&A including "What does a 20KSP joist mean?" for new estimators; real-time collaboration and public share links; "Alkali AI gets better with every project".Source
    • Your websiteBlog index: seven posts between October 2025 and July 2026 (beam scanner, takeoff how-to, column detection, nesting guide, milestones, a comparison with a general-purpose chat assistant, a hiring post); nothing published since July 20, 2026; no newsletter signup.Source
  • 16

    "How to Do a Steel Takeoff" walks through framing plans, columns, base plates and connection details ("the hardest part"), and warns that small missed pieces hurt bids. It offers no checklist or template. The site has no newsletter, so readers who aren't ready to try the product leave without a way back.

    Medium impactLow effort First result in 1 month Founder/CEO

    First action

    Draft the checklist from the guide's "Watch out for" and "Don't forget" sections.

    How to measure it

    Checklist sign-ups

    Now
    –
    Target
    300
    By
    90 days

    Steps

    1. 1

      Draft

      From the guide's warnings.

      Founder/CEO 1 day

    2. 2

      Design

      PDF and spreadsheet versions.

      Founder/CEO 1 day

    3. 3

      Gate

      Email field on the guide and /learn.

      Founder/CTO 4 hours

    4. 4

      Set up email

      A simple newsletter tool.

      Founder/CTO 2 hours

    5. 5

      Send monthly

      Tip, update, number.

      Founder/CEO monthly

    6. 6

      Measure

      Sign-ups and sign-ins from email.

      Founder/CTO monthly

    First result
    1 month
    Cost
    Under $50 a month
    Owner
    Founder/CEO

    Smallest test

    Add the checklist to the guide and count downloads in two weeks.

    Why it works

    Turn the guide into a one-page PDF and spreadsheet checklist: sheets to review, schedules to find, commonly missed items (embeds, hoist beams, base plates, misc steel). Gate it with an email field. Follow up with a short monthly email: one estimating tip, one product update, one customer number.

    A practical checklist is something estimators keep and share, and the list gives a two-person team a cheap channel back to readers who aren't ready yet.

    Ready to use

    ChecklistSteel takeoff checklist (draft)

    [ ] Read general notes and specs for steel grades, galvanizing, fireproofing
    [ ] Framing plans: every beam size, count and length, studs, camber
    [ ] Column schedule: sizes and heights (or derive from TOS / foundation levels)
    [ ] Base plates and anchor rods from foundation details
    [ ] Bracing and moment connections
    [ ] Embeds, hoist beams, lintels and misc steel
    [ ] Stairs and rails
    [ ] Addenda compared against the last set
    [ ] Tonnage check against a similar past job

    Risks and how to handle them

    • Gating cuts shares

      Keep the article open and gate only the file.

    Evidence

    • Your website"How to Do a Steel Takeoff" (January 26, 2026): about 1,300 words on framing plans, columns, base plates and connection details ("the hardest part"), with two "Try Alkali Here" buttons to the viewer; no downloadable checklist or template.Source
    • ResearchSite stack: Next.js on Railway with Google Analytics, PostHog and Crisp chat; no pricing page, no newsletter and no YouTube channel; social presence is LinkedIn and GitHub; domain registered May 2025.Source
  • 17

    Every Alkali takeoff builds a 3D estimodel that exports to IFC. The detailing service starts from exactly that model ("Alkali: Minutes, Tekla: Hours, Detailers: Days"). The software and the service are sold on separate pages through separate emails, and nothing in the product asks a shop what happened to the bid.

    High impactMedium effort First result in 2 months Founder/CEO

    First action

    Ask the five most active shops which of last month's bids they won and who detailed them.

    How to measure it

    Takeoff-to-detailing conversions

    Now
    –
    Target
    4 a month
    By
    2 quarters

    Steps

    1. 1

      Ask shops

      Who details their won jobs today.

      Founder/CEO 1 week

    2. 2

      Add bid status

      Pending, won, lost on each project.

      Founder/CTO 2 days

    3. 3

      Build the prompt

      On "won", offer a prefilled quote.

      Founder/CTO 3 days

    4. 4

      Set pricing

      A discount for jobs that start from an Alkali estimodel.

      Founder/CEO 2 days

    5. 5

      Quote fast

      One business day.

      Founder/CEO per request

    6. 6

      Track

      Won jobs, quotes, starts.

      Founder/CTO monthly

    First result
    2 months
    Cost
    Engineering time plus detailer capacity
    Owner
    Founder/CEO

    Smallest test

    Email shops who won a recent bid with a detailing quote.

    Why it works

    Add a "Bid status" field (pending, won, lost) to each project. When a shop marks a job won, offer: "Send this model to Alkali detailing; quote in one business day", with the tonnage and scope already filled in from the takeoff. Price detailing with a discount when it starts from an Alkali estimodel, since the model is already built.

    The shop has just won the work and needs shop drawings fast. Alkali already holds the model, so it can offer speed no outside detailer can match.

    Risks and how to handle them

    • Shops have loyal detailers

      Offer IFC handoff to their detailer as the free alternative.

    Builds on

    Evidence

    • Your websiteDetailing page: "The smarter way to detail steel." with 3D models, Tekla shop drawings and connection design, "Drawings from bid set to fab shop in days, not weeks" and "Every sheet is checked by a human who has done this before"; both "Set Up Discovery Call" buttons are mailto links (subject "Alkali Detailing Project"); no price, turnaround or file upload; no meta description.Source
    • Your websiteDocs: exports (CSV, PDF markup, IFC 3D Estimodel, nesting plans); a "Bluebeam Desktop" controls preset; customization "requires a subscription"; a Company Page with team usage metrics (projects uploaded, tonnage analyzed); 3D Estimodeling in Beta/Pilot; share links with Owner, Editor and Viewer roles.Source
  • 18

    Revision compare is one of Alkali's features: it finds changes across addenda and shows Removed/Changed/Added with Accept/Reject (the homepage example is a W12x16 becoming a W16x26 in Addendum 2). Every estimator gets addenda, often with little time before bid day, and often doesn't have a takeoff in any tool yet.

    Medium impactMedium effort First result in 6 weeks Founder/CTO

    First action

    Wrap the existing revisions feature in a no-login upload page for two PDFs.

    How to measure it

    Compare-to-sign-in rate

    Now
    –
    Target
    20%
    By
    1 quarter

    Steps

    1. 1

      Scope it

      Sheet limits and file size.

      Founder/CTO 1 day

    2. 2

      Build the page

      Two uploads, one result.

      Founder/CTO 1 week

    3. 3

      Write the landing

      Indexable explanation and example.

      Founder/CEO 1 day

    4. 4

      Add the upsell

      Carry changes into a takeoff.

      Founder/CTO 2 days

    5. 5

      Promote

      LinkedIn and estimator communities.

      Founder/CEO 2 weeks

    6. 6

      Measure

      Runs, sign-ins, paid.

      Founder/CTO monthly

    First result
    6 weeks
    Cost
    Engineering and compute
    Owner
    Founder/CTO

    Smallest test

    Offer manual addendum compares by email for two weeks.

    Why it works

    Launch a standalone page: upload the old and new structural sheets and get a highlighted list of steel changes, free with no sign-up for up to a set number of sheets. Offer to carry the changes into a full takeoff with sign-in. Give it its own indexable landing page.

    A narrow, urgent, free job brings in estimators who aren't shopping for takeoff software yet and shows the product's precision on their own drawings.

    Risks and how to handle them

    • Compute cost from free use

      Limit sheets per run and rate-limit.

    Evidence

    • Your websiteHomepage product sections: AI takeoff assist for beams, columns, joists, braces and base plates; tonnage from the AISC manual; revision compare across addenda with Accept/Reject (example: a beam changed from W12x16 to W16x26 in Addendum 2); drawing Q&A including "What does a 20KSP joist mean?" for new estimators; real-time collaboration and public share links; "Alkali AI gets better with every project".Source
    • Your websiteDocs: exports (CSV, PDF markup, IFC 3D Estimodel, nesting plans); a "Bluebeam Desktop" controls preset; customization "requires a subscription"; a Company Page with team usage metrics (projects uploaded, tonnage analyzed); 3D Estimodeling in Beta/Pilot; share links with Owner, Editor and Viewer roles.Source

Big bets 3

Larger plays that can change the trajectory.

  • 19

    The founders already describe it: some projects are now uploaded by several customers, and a central system could produce an estimate before a fabricator starts bidding, with labor rates tuned to each shop's past jobs. The open AI engineer role is about agents that navigate large plan sets. Rivals sell faster takeoff; none of them claims a takeoff that is already waiting.

    High impactHigh effort First result in 1 quarter Founder/CTO

    First action

    Pick ten shops and one region, and agree how they want to receive bid sets.

    How to measure it

    Pilot bid volume

    Now
    –
    Target
    +50%
    By
    2 quarters

    Steps

    1. 1

      Pick the pilot

      Ten shops, one region.

      Founder/CEO 2 weeks

    2. 2

      Agree sources

      How bid sets reach Alkali (shops forward invitations).

      Founder/CEO 1 week

    3. 3

      Profile shops

      Project types, sizes, labor rates.

      Founder/CEO 2 weeks

    4. 4

      Automate overnight runs

      Agents run takeoff and summary.

      Founding Applied AI Engineer 6 weeks

    5. 5

      Notify

      Morning email with tonnage and scope.

      Founder/CTO 1 week

    6. 6

      Keep it private

      Each shop's estimate and rates separate.

      Founder/CTO ongoing

    7. 7

      Price it

      Premium tier after the pilot.

      Founder/CEO 1 month

    8. 8

      Measure

      Bids per estimator and win rate.

      Founder/CTO monthly

    First result
    1 quarter
    Cost
    Engineering time and compute
    Owner
    Founder/CTO

    Smallest test

    For three shops, pre-run every bid invitation they forward for a month.

    Why it works

    Pilot with ten shops in one region: they set the project types and sizes they bid, and when a matching public bid set appears, Alkali runs the takeoff overnight and sends a notification with tonnage, scope summary and a labor estimate at the shop's own rates, ready for review. Charge it as a premium tier. Keep every estimate private to the shop.

    It changes the product from a faster tool to more bids with the same team. It also uses the cross-customer coverage that only a platform with Alkali's volume has.

    Risks and how to handle them

    • Drawing access and permissions

      Only process sets the shop forwards or is invited to.

    • Shops fear competitors get the same estimate

      Keep labor rates, edits and bid decisions private per shop.

    Builds on

    Evidence

    • Your website"Alkali Early Milestones April 2026" (May 2, 2026): project volume "more than doubled" each month since launch; over 25,000 tons estimated in April, about 7% of average monthly U.S. structural steel volume; the same project is sometimes uploaded by several customers; 350+ fabricators spoken with in a year; a Baltimore video shows 350+ tons taken off in about 15 minutes; a fabricator called it the "Tesla of estimating"; coverage "still sparse in many regions"; the goal of an estimate ready before a fabricator starts bidding, tuned to each shop's labor rates.Source
    • HiringFounding Applied AI Engineer (New York; $150K–$220K, 0.75%–1.50% equity) on Y Combinator's job board: improve vision models that extract quantities and sizes from drawings, build agents that navigate large plan sets, and build datasets, evaluation pipelines and test harnesses. The company description says about 10% of U.S. structural steel by volume flows through Alkali. YC lists Alkali as a Spring 2025 company in Houston with a team of 2; the role was also announced on the Alkali blog on July 20, 2026.Source
    • Competitor researchFerra: bid intelligence, steel takeoff, revision control and 3D models; "Start for free"; a co-founder's "Send me a drawing set" offer; 10 customer logos and 3 named testimonials; a /compare page against SketchDeck, Steel Genie, Beam AI and SteelFlo that does not mention Alkali.Source
  • 20

    Alkali's nesting turns a takeoff into stock-aware cut plans and cut lists. Its own guide lists "have a distributor nest" as one of the three options fabricators use today, and warns it may limit stock options. A nest built in Alkali is a ready-made material order for a service centre.

    High impactHigh effort First result in 1 quarter Founder/CEO

    First action

    Ask five active shops which service centres they buy from and whether they'd use one-click quotes.

    How to measure it

    Quote requests

    Now
    –
    Target
    100 a month
    By
    2 quarters

    Steps

    1. 1

      Interview shops

      Suppliers and ordering habits.

      Founder/CEO 2 weeks

    2. 2

      Approach suppliers

      Two or three regional service centres.

      Founder/CEO 1 month

    3. 3

      Define the format

      Stock list the supplier can price.

      Founder/CTO 1 week

    4. 4

      Build the button

      On nesting output.

      Founder/CTO 2 weeks

    5. 5

      Agree terms

      Referral fee or co-marketing.

      Founder/CEO 1 month

    6. 6

      Pilot

      One region for a quarter.

      Founder/CEO 1 quarter

    7. 7

      Measure

      Requests, orders, new shops referred.

      Founder/CTO monthly

    First result
    1 quarter
    Cost
    Staff time
    Owner
    Founder/CEO

    Smallest test

    Email one shop's nest to its supplier and time the quote.

    Why it works

    Pilot with two or three regional steel service centres: a "Request material quote" button on the nesting output sends the stock list (shapes, lengths, quantities) to the chosen supplier. Suppliers reply with price and lead time inside Alkali. Suppliers pay a referral fee or subscription for qualified quote requests, or promote Alkali to their fabricator customers.

    It puts Alkali in the purchase step after the estimate and gives distributors a reason to recommend it to their fabricator customers.

    Risks and how to handle them

    • Shops worry about pricing exposure

      The shop chooses whether and to whom to send.

    Builds on

    Evidence

    • Your website"Alkali Guide to Nesting" (March 2, 2026): stock lengths, kerf, end trim and nesting across jobs, noting waste can cut "a few thousand dollars" from a small job's margin; its buttons go to the homepage, not the /nesting tool.Source
    • Your websiteDocs: exports (CSV, PDF markup, IFC 3D Estimodel, nesting plans); a "Bluebeam Desktop" controls preset; customization "requires a subscription"; a Company Page with team usage metrics (projects uploaded, tonnage analyzed); 3D Estimodeling in Beta/Pilot; share links with Owner, Editor and Viewer roles.Source
  • 21

    Alkali has estimated 2,500+ projects across 48 states, over 25,000 tons in a single month, with project type, size and member mix in every takeoff. Nobody else publishes aggregate data on what steel is being bid, where and in what shapes, and fabricators, distributors and trade press would all use it.

    High impactMedium effort First result in 1 quarter Founder/CEO

    First action

    Check which fields can be aggregated safely and set a minimum group size for privacy.

    How to measure it

    Report downloads

    Now
    –
    Target
    1,000
    By
    90 days after launch

    Steps

    1. 1

      Set privacy rules

      Aggregate only, minimum group size, customer consent in terms.

      Founder/CEO 1 week

    2. 2

      Pull the data

      Tonnage, region, type, shapes, revisions.

      Founding Applied AI Engineer 2 weeks

    3. 3

      Analyse

      Ten clear findings.

      Founder/CEO 2 weeks

    4. 4

      Design

      Report PDF and charts.

      Founder/CEO 2 weeks

    5. 5

      Pitch press

      Steel and construction trade outlets.

      Founder/CEO 2 weeks

    6. 6

      Launch

      Webinar and LinkedIn series.

      Founder/CEO 1 week

    7. 7

      Benchmark shops

      Each customer's numbers against the market.

      Founder/CTO 1 month

    First result
    1 quarter
    Cost
    Design and staff time
    Owner
    Founder/CEO

    Smallest test

    Publish one chart (tons by building type) on LinkedIn.

    Why it works

    Publish an anonymised, aggregated report: tons bid by region and building type, most common shapes, average tons per project type, revision churn between addenda, and trends across the year. Release it with a short webinar and a LinkedIn series. Gate the full PDF with an email field and offer each shop its own benchmark against the market.

    Original data earns press, links and trust in a way product pages can't, and it shows the scale of usage to every reader.

    Risks and how to handle them

    • Customers object to data use

      Aggregate only, check terms, offer opt-out.

    Builds on

    Evidence

    • Your websiteHomepage: H1 "The smarter way to estimate steel."; stat line "2,500+ steel projects | 48 states | >10% of U.S. steel"; a no-signup "Try Here" plus Google, Microsoft and email sign-in; 21 fabricator logos captioned "Anonymized subset of projects estimated with Alkali in 2026", but no testimonials, case studies or pricing; exports to Tekla, Bluebeam, Adobe and Excel; the only contact is a mailto link; the footer reads "© 2025 Alkali Engineering, Inc. · Built in Houston, TX".Source
    • Your website"Alkali Early Milestones April 2026" (May 2, 2026): project volume "more than doubled" each month since launch; over 25,000 tons estimated in April, about 7% of average monthly U.S. structural steel volume; the same project is sometimes uploaded by several customers; 350+ fabricators spoken with in a year; a Baltimore video shows 350+ tons taken off in about 15 minutes; a fabricator called it the "Tesla of estimating"; coverage "still sparse in many regions"; the goal of an estimate ready before a fabricator starts bidding, tuned to each shop's labor rates.Source

Wildcards 1

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

  • 22

    The homepage already features "#2: Baltimore Speed Run", and the milestones post cites 350+ tons in about 15 minutes. That is the most persuasive thing Alkali shows, yet there is no YouTube channel and the only social presence is LinkedIn and GitHub. The Show HN post got 5 points; a speed run speaks the language of estimators, not developers.

    Medium impactLow effort First result in 1 month Founder/CEO

    First action

    Create a YouTube channel and upload the Baltimore run as episode 2.

    How to measure it

    Video-sourced sign-ins

    Now
    –
    Target
    50 a month
    By
    2 quarters

    Steps

    1. 1

      Set up the channel

      YouTube plus the existing Baltimore run.

      Founder/CEO 1 hour

    2. 2

      Pick sets

      Public bid sets of varied types.

      Founder/CEO weekly

    3. 3

      Record

      Clock on screen, no cuts.

      Founder/CEO weekly

    4. 4

      Cut a short

      60 seconds for LinkedIn.

      Founder/CEO weekly

    5. 5

      Link the project

      Shared view for viewers to explore.

      Founder/CTO weekly

    6. 6

      Take requests

      Estimators submit sets.

      Founder/CEO ongoing

    7. 7

      Measure

      Views, clicks, sign-ins.

      Founder/CTO monthly

    First result
    1 month
    Cost
    Founder time, about 2 hours a week
    Owner
    Founder/CEO

    Smallest test

    Post one LinkedIn speed run and compare sign-ins to a normal post.

    Why it works

    Publish a 10-15 minute speed run every week on a public bid set: load, scan, check, tonnage, with the clock on screen and the misses fixed live. Post the full video on YouTube and a 60-second cut on LinkedIn. Number the series, invite estimators to send in sets to be run, and end each one with "try this set yourself" linking to the shared project.

    A repeatable format builds an audience and lets buyers check the speed claims for themselves, showing the misses as well as the wins.

    Risks and how to handle them

    • Drawing ownership

      Use public bid sets or sets submitted with permission.

    Evidence

    • Your websiteHomepage: H1 "The smarter way to estimate steel."; stat line "2,500+ steel projects | 48 states | >10% of U.S. steel"; a no-signup "Try Here" plus Google, Microsoft and email sign-in; 21 fabricator logos captioned "Anonymized subset of projects estimated with Alkali in 2026", but no testimonials, case studies or pricing; exports to Tekla, Bluebeam, Adobe and Excel; the only contact is a mailto link; the footer reads "© 2025 Alkali Engineering, Inc. · Built in Houston, TX".Source
    • Your website"Alkali Early Milestones April 2026" (May 2, 2026): project volume "more than doubled" each month since launch; over 25,000 tons estimated in April, about 7% of average monthly U.S. structural steel volume; the same project is sometimes uploaded by several customers; 350+ fabricators spoken with in a year; a Baltimore video shows 350+ tons taken off in about 15 minutes; a fabricator called it the "Tesla of estimating"; coverage "still sparse in many regions"; the goal of an estimate ready before a fabricator starts bidding, tuned to each shop's labor rates.Source
    • ResearchSite stack: Next.js on Railway with Google Analytics, PostHog and Crisp chat; no pricing page, no newsletter and no YouTube channel; social presence is LinkedIn and GitHub; domain registered May 2025.Source
    • Hacker NewsShow HN (December 2025): "Agentic PDF Viewer for Schematics", 5 points and no comments; the post says the team uses it mainly for steel estimation.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.

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.

  • Gruve growth strategyAn AI services and infrastructure company: managed AI-powered security operations, Cisco-aligned security, data center and networking…21 growth ideas
  • HealthVerity growth strategyA privacy-protected real-world healthcare data ecosystem (HealthVerity Marketplace), identity and de-identification technology…21 growth ideas
  • Skyflow growth strategyA runtime AI data control platform built on a data privacy vault: it detects, tokenizes and governs sensitive data (PII, PHI, PCI) across…22 growth ideas
  • Headway growth strategyA mental health network and practice platform: it credentials independent therapists and psychiatrists with insurance plans, handles…21 growth ideas
  • SplitMetrics growth strategyMobile marketing software and services: SplitMetrics Acquire for Apple Ads management with the Iris AI agent, App Radar for ASO, Optimize…21 growth ideas
  • Nike growth strategyThe world's largest athletic footwear and apparel company, selling through Nike Direct (nike.com, the Nike app, SNKRS and Nike stores) and…21 growth ideas
  • Delinea growth strategyAn identity security platform for privileged access: secrets vaulting (Secret Server), server and cloud privilege control, remote access…21 growth ideas
  • Arrive growth strategyA global mobility platform bringing together EasyPark, Flowbird, RingGo, ParkMobile and Parkopedia: parking apps and payments, pay…20 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. Alkali did not commission, review or endorse it, and Programmable is not affiliated with Alkali. 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 Alkali runs its business.

Something wrong, or do you represent Alkali? 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 307 ideas

    Fix the routes to revenue: the detailing page in the sitemap with a drawing-set form, a lighter homepage, named proof, a pricing page, a drawing-set offer, a linked nesting page and consistent numbers.

  2. Days 31 to 608 ideas

    Content and product loops: accuracy benchmark, comparison pages, a Bluebeam switcher path, first-upload activation, monthly shop reports, share-link loop, an estimator glossary and a takeoff checklist.

  3. Days 61 to 907 ideas

    Regional coverage, the takeoff-to-detailing upsell, a free addendum compare, and commit to the big bets: the pre-bid estimate, a service-centre nesting partnership and the annual steel bidding report; start the speed-run series.

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 speed: hundreds of tons taken off in minutes, a project loaded in about 30 seconds. The most cited problem: manual takeoff from paper or PDF bid sets.

  • Ease of use (a fabricator called it the "Tesla of estimating")
  • Missing small pieces and connection details that hurt bid margins
  • In their words: “takeoff”

Market

Competitive landscape

4 competitors compared: Steel Genie (SDS2 by ALLPLAN), Ferra, Beam AI and Manual takeoff in PDF tools.

  • Gap to exploit: Published, measured takeoff accuracy for steel
  • Steel Genie (SDS2 by ALLPLAN): Demo-only, no free trial or published price
  • Ferra: Smaller project footprint claims

Hiring signals

Alkali is a team of two hiring its first engineer: a Founding Applied AI Engineer in New York ($150K–$220K, 0.75%–1.50% equity) to improve the vision models, build agents that navigate large plan sets, and build…

  • Founding Applied AI Engineer: Vision models, plan-set agents and evaluation harnesses. Ideas 8, 15, 19, 21.

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 first (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)
  • 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