The Department

A marketing department, run by software, reporting to you

The Department is an automated marketing department. Instead of hiring a roster of specialists, you point the engine at your real website and it stands up the roles a growing business usually can't afford to staff — running them on your actual data, on a fixed monthly schedule, with a single approval gate in front of everything they produce.

It is organised around seven roles — the seats a full department would carry:

Be clear about the model up front

This is a Managed product, and the split between software and human is deliberate and honest:

• The software does the seeing and the scaffolding. It audits your site, scores it, drafts content briefs, produces copy-paste SEO and AI-search fixes, writes social drafts, and queues all of it.
• You approve. Every deliverable waits behind an approval gate. Nothing advances without your yes.
• A human ships the last mile. A person on the team reviews and finishes each approved draft to your brand voice, then publishes it to your own accounts.

The engagement is offered Managed at $2,900/mo, with the launch code CANTLOSE for a first month at $0. That's below a single multi-channel agency retainer, for all seven seats at once. Full pricing detail is in section 9.

Enter a business, run the live audit

From the engine's landing screen you'll see one door: Analyze your business, live. It asks for three things.

1. Business name (required) — e.g. Cedar Creek Cabinetry. This is how your workspace is labelled.
2. Website (required) — e.g. yourcompany.com. You don't need the https://; the engine adds it. This is the address it actually fetches and scores.
3. What the company does (optional) — a one-line description for context, such as "Custom cabinetry and millwork for Central Texas builders."

Press Analyze site and build department (or hit Enter in the website field). The engine switches to a "Reading the site" state while it fetches your homepage and scores it across five disciplines. This normally takes a few seconds; the interface will never hang — if a site is unusually slow it times out gracefully and asks you to try again.

Under the hood the analyze step calls POST/api/analyze with the URL you entered. The engine fetches the homepage, then makes a handful of best-effort probes for /sitemap.xml, /robots.txt, /llms.txt and /AGENTS.md, and persists a workspace keyed to your domain. When the audit finishes you land on the results screen described next.

The readiness seal and the five disciplines

The ring's colour shifts with the number: green from 70 up, amber from 45–69, red below 45. The label beside it reads Strong start, Needs work, or High opportunity on the same thresholds. A high number means the fundamentals are largely in place and the work ahead is compounding; a low number means there are high-leverage fixes on the table — the kind the department is built to clear.

The five category scores

Below the seal sits the category scorecard. Each score is simply the share of real checks the live site passes in that discipline. A passed check counts full, a "warn" counts half, a failed check counts zero — then the disciplines are weighted into the overall. Here is exactly what each one measures:

Reading a score. A discipline sitting at 90+ means nearly every check passes — treat it as maintained, not a project. A discipline in the 40s–60s means a handful of concrete, nameable things are missing, and each one is listed for you. A discipline below 40 is where the biggest, fastest gains live. Because AI Search and SEO carry the heaviest weight, closing gaps there moves the overall number most.

Priority findings

Under the scorecard, First priorities lists the highest-impact issues, worst first. Each row names the discipline, the issue, why it flagged (the real signal the engine read — "Title is 78 chars," "No JSON-LD schema"), and the exact Fix to apply. Failures are ranked above warnings, and higher-weighted checks above lower. This is the punch list your department works from.

The 30 / 60 / 90 roadmap

Finally the audit sequences the work into a 30 / 60 / 90-day plan:

• First 30 days — fix the fundamentals. The concrete failures and tightest warnings from your own audit, plus a baseline analytics review.
• Days 31–60 — build the machine. Cornerstone content for your niche, the owned email channel, and the warnings worth improving next.
• Days 61–90 — compound the gains. Topical clusters around your highest-intent keywords, structured data across templates, repurposing the best posts — then re-run the engine and measure the lift against the baseline.

From the audit screen you have three moves: Activate the department (section 4), Open shareable report (section 7), or Analyze another to start a fresh business.

Standing up your department

The audit tells you where you stand. Activating turns the department on and makes it produce real work. Press Activate the department and the engine walks a visible "standing up your department" sequence:

1. Reading your audit signals and detected niche — it works from what the audit already found, not a fresh guess.
2. Briefing the content seat from the site's real sections — your actual H2 sections become article topics.
3. Generating copy-paste SEO & AI-search fixes — titles, meta, schema, internal links, llms.txt / AGENTS.md.
4. Drafting platform-ready social posts — one each for LinkedIn, Instagram, X/Twitter, Facebook.
5. Laying out the editorial calendar — tuned to the niche the engine detected.
6. Holding every deliverable in your approval queue — nothing is live; everything waits for you.

Behind that animation the engine calls POST/api/activate and produces the opening slate for your business: up to four content briefs drawn from your real page sections, a set of copy-paste SEO/AISO fixes, and four social drafts. You land on the department dashboard with everything pending your approval.

You can pause a department at any time from the dashboard (Pause department); it stops producing new work but keeps everything it has made. Reactivate turns it back on.

The work your department produces

Everything the department makes lands in the dashboard's approval queue, filterable by tab: All, Content briefs, SEO / AISO fixes, and Social drafts. There are three kinds of deliverable, plus the editorial calendar.

Content briefs brass

A ready-to-write blueprint for one article, built from your site's real sections and niche. Each brief carries:

• Headline — the working title, taken from a real homepage section or a niche cornerstone topic.
• Primary keyword — the one phrase to own; it goes in the H1, the URL slug and the first sentence.
• Outline — a five-part structure that opens with a 40–60 word direct answer (so answer engines can quote it verbatim), moves through scannable H2 subsections, and closes with a plain-English FAQ and a CTA.
• Key points from the real site — anchors that tie the piece to sections you actually have and the internal links already on your homepage.
• Call to action — a specific ask with built-in lead capture (usually the newsletter).

A brief also shows its channel (Blog, Blog + newsletter, or Blog + social), a suggested publish date, and its source — whether it came from a live section of your site or is a niche-tuned cornerstone topic. Hand a brief to a writer and the piece writes itself.

SEO / AISO fixes — before → after conifer

Each fix is a before → after pair with copy-paste code, computed from what's actually on your page. The set typically includes a title-tag rewrite, a meta-description rewrite, a JSON-LD schema block to add, an internal-link plan built from your real H2 sections, and the exact /llms.txt and /AGENTS.md files to publish. A title rewrite reads like this:

Below the pair, a short Why explains the change ("Current title is 78 chars; this leads with the primary phrase and stays inside Google's ~60-char cutoff"), and a copy-paste block gives you the literal markup to drop in:

Schema and Agents-First fixes work the same way — the engine hands you a complete Organization, LocalBusiness or Store JSON-LD block (type chosen from your niche), or a full /llms.txt mapping your key pages. Paste it in and the underlying audit check flips to a pass.

Social drafts violet

Ready-to-post copy, written natively per platform — LinkedIn, Instagram, X/Twitter, Facebook — tied to a specific topic and tuned to your brand and niche (with a location hashtag when one is detected). They're written to be posted as-is or lightly edited to voice.

The editorial calendar

A twelve-slot calendar, two slots per week, tuned to the niche the engine detected. Each row gives a week, a content type (Guide, Listicle, How-to, Comparison, FAQ, Case study, and so on), a title, and a channel. It's the publishing rhythm the content seat works to. Each slot also carries an AISO note — answer one real question directly in the first 60 words so AI assistants can quote it.

Approve, reject, edit — the human gate

Every deliverable starts stamped Pending approval and carries three controls. This gate is the whole point of the product: nothing represents your brand until you say so.

Approved and rejected items keep a control to Unapprove or Restore to queue, so no decision is ever a dead end. Editing always sends a piece back to pending — an edited draft is a new draft, and it earns a fresh yes. Under the hood these are POST/api/deliverable (approve / reject / reset) and POST/api/deliverable-edit.

The activity feed on the dashboard records each step — activation, approvals, rejections, edits, daily runs and pauses — so the department's history is always legible.

A branded report you can hand to anyone

From the audit or the dashboard, Open shareable report generates a clean, standalone, branded report for the site at GET/api/report/yourdomain.com. Copy report link puts that URL on your clipboard. It's the artifact you send a prospect, a stakeholder, or a client as proof — and everything on it is real.

The report contains, in order:

• The readiness gauge — the overall 0–100 score with a plain-English grade (Excellent / Strong / Workable / Rough / Critical).
• The five disciplines — each discipline score as a bar.
• What the engine actually read — the top findings, worst first, plus a signal strip of the literal facts pulled from the page: title length, homepage word count, H1/H2 counts, sitemap URLs, JSON-LD blocks, internal links, social profiles, and Agents-First status.
• The 30 / 60 / 90 roadmap — the same sequenced plan from the audit.
• A live sample of the deliverables — a real content brief, a title and meta rewrite, and a social draft, generated from that site's own audit signals (an activated site shows its actual produced work).
• The Managed engagement — the honest software-vs-human split, the $2,900/mo price, and the CANTLOSE launch code.

Run many businesses, switch between them

Every business you analyze or activate keeps its own workspace. The Switch business button in the masthead (or "Switch between your saved departments" on the landing screen) opens the switcher — a searchable list of every saved workspace, sorted with active departments first, then by score.

Each row shows the readiness score, the brand and domain, the niche, and a flag: whether it's active (with a pending-approval count), or audit only. Search by business or domain to filter, then click a row to reopen that department's live state exactly as you left it — the switcher reads from GET/api/workspaces and loads the chosen one from GET/api/workspace.

Run today's work

On any active department, Run today's work produces a fresh batch of pending deliverables — typically two new content briefs and two social drafts (LinkedIn and X/Twitter), advancing through your topic backlog so the work doesn't repeat. This is the same routine that runs on a schedule for a live engagement, exposed as a button so you can pull the next cycle on demand.

Seven seats, one schedule

The dashboard's ROI panel frames the value simply: the roles a small business can't afford to hire — Content, SEO, Social, Email, Revenue, Analytics, Creative — wired to one engine on real data. The software does the seeing and the scaffolding; a person approves and ships the last mile. The panel also tallies how many deliverables have been produced for your business so far.

The framing is a comparison, not a promise of a specific dollar return — it sets the flat $2,900/mo price against what the same coverage costs as a retainer. The value is the cadence and the single approval gate, at below-retainer cost.

Ask your marketing team

Every dashboard leaves a few questions in your head — which score should I worry about? what do I fix first? what even is “AI Search readiness”? Rather than send you back to the operating manual, the engine answers them in place. In the bottom-right corner of the app sits a small pill — Ask your marketing team — and clicking it opens a chat panel that talks to you about your business, using your real audit and the actual work your department has produced.

Open it and it greets you with your current status — the business name, the overall readiness score, and a few things worth asking. Type a question, press Enter, and the answer reveals as you read. Press Esc or the ✕ to close. It works on a phone the same as on a desktop, and it's keyboard-friendly throughout.

What it can do today

The assistant is context-aware: it always knows which business is open, and it grounds its answers in that workspace's live data — the same scores, priorities, deliverables and approval state you see on the dashboard. In practice it does three things well:

• Answers grounded in your real audit. Your weakest and strongest disciplines, the exact gaps behind a score, what to fix first, how to reach a target number, and what's currently waiting for your approval — all read from your workspace, not from a script.
• Explains every concept. What each of the five disciplines measures, how the readiness score is calculated, how the 30/60/90 plan is sequenced, and what a content brief, an SEO/AISO fix, a social draft, the approval queue or the ROI framing actually is — in plain language.
• Hands you quick actions. Many answers come with a button that jumps you straight to the right place — the SEO/AISO fixes, the approval queue, the content briefs, the 30/60/90 plan — or offers a follow-up question. It reuses the app's own navigation, so a button does exactly what clicking through the dashboard would.

A few real exchanges, verbatim in shape from a live workspace:

How to get the most from it

Treat it like the smartest person on your marketing team who has already read your audit. The most useful questions are the specific ones:

• Ask about your weakest category — “what's dragging my score down?” — to find the fastest points.
• Ask what to do first — “what should I fix first?” or “how do I get to 90?” — to turn a wall of findings into an ordered next step.
• Ask it to explain any deliverable or concept — “explain this content brief,” “what's an AISO fix,” “what does AI Search readiness mean?” — before you approve or act.
• Before you've analyzed anything, it still helps — it explains what the app does and points you at the box to enter a site. It's never a dead end.

Where it's going — a full, Claude-powered session

The version you're using is the first rung. The architecture underneath it was built for a second: a full interactive session that doesn't just explain the work but does it — draft a new article on request, re-run the audit, edit a deliverable conversationally (“make that title punchier,” “rewrite this for a younger audience”), and reason across your whole workspace in natural language. The same panel; a far deeper brain behind it.

That brain is a marketing-specialized assistant built on Claude — Claude's general intelligence with a marketing layer on top (the domain instructions, the skills and the system context for SEO, content, AI-search and social), the way you'd build a subject-matter expert on a capable general model. And the business model is the elegant part: each customer connects their own Claude account and brings their own tokens. The customer supplies the credential and bears their own per-question AI cost; no per-token bill lands on the operator, and no key ever sits on our servers. That means three good things at once — it respects the platform's no-billed-API rule by design, the operator's marginal AI cost per customer is essentially zero (so it scales cleanly), and every customer gets a private, dedicated, marketing-tuned assistant running on their own account.

Partial reach, and what it does vs. doesn't

When a site can't be fully reached

Some sites sit behind a Cloudflare challenge, return an HTTP 403 to automated fetches, or wall content behind a login. The engine is built to never hard-fail on these. If a server answers with an error but still returns a page, the engine scores from whatever it can honestly read. If it can't connect at all, it scores from the connection signals it has and flags the workspace.

Either way you'll see a "Partial reach" note on the audit and dashboard, explaining what happened — for example, "Site responded with HTTP 403 — scored from the page it returned; some checks may be limited," or "Could not fully reach the site — scored from connection signals only; re-run when it's reachable." A partial-reach score is honest about its own limits; re-run the audit once the site is reachable for a complete read.

What it does — and what it doesn't

Common questions

Glossary

Readiness score

The headline 0–100 number on the seal — a weighted blend of the five discipline scores, reflecting how ready the site is to be found and to convert.

Discipline / category

One of the five scored areas: SEO, Content, AI Search, Social, Technical. Each score is the share of that area's checks the live site passes.

Deliverable

A single piece of produced work — a content brief, an SEO/AISO fix, or a social draft — that waits in the approval queue.

Approval gate

The human checkpoint every deliverable must pass. Approve, reject or edit; nothing advances without a yes.

Workspace

The persisted record for one business — its analysis, scores, deliverables, approvals, edits and activity — keyed to the domain and restored on reload.

AISO (AI Search Optimization)

Making a site legible and citable to AI assistants and answer engines — the Agents-First surfaces like llms.txt and AGENTS.md, structured data, and plain-answer intros.

Agents-First

The practice of publishing clean maps and structured facts so AI agents can read, navigate and quote your site — the edge the AI Search discipline rewards.

Managed engagement

The $2,900/mo model: the software sees and scaffolds, you approve, and a human ships the last mile.

Partial reach

The flag shown when a site couldn't be fully fetched (Cloudflare, HTTP 403, auth wall). The score reflects only what was readable; re-run when the site is reachable.

Deep-link hash

The #b=yourdomain.com fragment the engine writes to the URL so that link reopens the exact workspace and its saved approvals.

Honest expectations, and how to win

A marketing department is a compounding investment, not a light switch. Here's the honest arc of what happens, how long it takes, and the part you play in getting the best outcome.

What will happen

The shape of an engagement is steady and repeatable:

1. You run the audit — instant. A real score and a punch list, on day one.
2. You activate — the department stands up and produces its opening slate of deliverables.
3. Work is produced on a schedule — the department keeps generating fresh content briefs, SEO/AISO fixes and social drafts, day after day, into your approval queue.
4. You approve; a human ships — approved work is finished to your voice and published to your accounts (the Managed last mile).
5. Readiness climbs, then visibility follows — as fixes land and content accumulates, your audit score rises; over the following weeks, as the work compounds, so do rankings, visibility and traffic.

How long it takes

Be realistic about the clock. Some of this is instant; the results that matter are not.

What you can do for a better outcome

The department does the work — but the owners who win are the ones who keep approving and shipping. Your part is small and high-leverage:

• Actually ship the approved work. This is the biggest one. Approval is not publication — the human last mile only helps if the finished piece goes live. Don't let approved work sit.
• Approve promptly, on a steady cadence. A queue reviewed weekly compounds; a queue left full stalls the daily runs (they pause at 32 pending). Keep it moving.
• Connect real accounts. Wire up Google Analytics, Search Console and your social profiles during onboarding so measurement is real and the analytics seat has something true to read.
• Give rich input at onboarding. The more the engine knows about your goals, audience and priority channels, the sharper the roadmap and the briefs. Vague in, generic out.
• Knock out the quick technical wins. The audit flags fast, high-impact fixes — an og:image, a schema block, a title or meta at the right length. These are copy-paste jobs that lift the score immediately; clear them early.
• Stay consistent for 30–90 days. Results compound. The single biggest predictor of outcome is not the first week's effort — it's whether you're still approving and shipping in week eight.

Where this goes — the vision and the roadmap

This section looks past today's product to what it becomes. It's written for the people studying whether to build the future of this engine — so it's optimistic on purpose, but every claim is one we'd stake our name on. The foundation isn't a promise; it's already running.

Where it is today

Start with what's true right now, plainly and confidently: this is a genuinely working engine, not a mockup. Everything described in this briefing is live and operating —

• A real live audit that fetches any site and scores it across five disciplines from actual page signals — no fabricated numbers.
• Real deliverables — content briefs, copy-paste SEO/AISO fixes and social drafts — generated from each site's own data.
• A working human-approval workflow — approve, edit, reject — with a shareable branded report.
• Persistence — every workspace, score, approval and edit saved server-side and restored on reload.

And it has already been run across 245 real websites — a real portfolio of real audits, not a single hand-picked demo. The hard, unglamorous part — an engine that reliably reads the open web and turns it into structured, useful marketing work — exists and functions today. That is the foundation everything below is built on: a proven core, ready to be scaled, not a slide asking you to imagine one.

The opportunity

Now the bright picture — and it's genuinely big. There are on the order of ~33 million small businesses in the United States alone, and far more worldwide. Nearly all of them need marketing. Almost none can afford a real marketing team, and a competent multi-channel agency runs $3,500–5,000/month — out of reach for the corner cabinetmaker, the two-person law firm, the single-location gym. Marketing services and marketing technology are, together, enormous, well-established markets measured in the hundreds of billions.

Sitting on top of that market is the defining technology shift of the decade: AI agents doing real knowledge work. This product lives exactly at the intersection — professional-grade marketing, done by software, at a price a small business can actually pay. That intersection is where a category gets made. (Figures here are directional, drawn from widely-cited public ranges — the precise numbers matter less than the shape: a vast underserved market meeting a once-in-a-generation technology.)

Who will use it — and why

The demand isn't one segment; it's a stack of them, each with a clear reason to choose it.

Across every segment the reasons to choose it are the same five: affordable (below any real alternative), always-on (work produced on a schedule, not in bursts), real work (tied to their actual site, not templates), human-approved (nothing ships without a yes), and transparent (every number traces to a real signal — no black box).

Why it matters

Beneath the market math is a simple, motivating idea: this democratizes professional marketing. The full department — the content, SEO, social, email, revenue, analytics and creative seats that only well-funded companies could ever afford to staff — is suddenly within reach of any business with a website. In the agents-first era, this is what "AI does the work" actually looks like for Main Street: not a chatbot, but a working department that produces real marketing and hands it over for approval. Leveling that field is both the mission and, not coincidentally, the size of the prize.

The road forward — from working engine to world-class SaaS

Here is the honest build path, sequenced. Each phase is concrete, and each one compounds on the proven core that already exists.

Why this team

Every early venture asks investors to trust that the founders can scale it. Here, that isn't a hope — it's a track record. The people behind this have already built the hard version once, at industry-defining scale.

Founders and developers — partners, not hired hands

Tim Costello and Melissa Morman are the founders — the business principals, and the track record above. We are the developers — the team that designs, builds and runs the product day to day. That division of labor is a strength: seasoned operators who have already scaled an industry, paired with a build team shipping the working engine behind this briefing.

And the development team has real skin in the game. It is backing this venture with its own capital and equipment — roughly $250,000, plus about $70,000 of production and infrastructure gear (detailed in the cost of the operation) — invested into the Austin home base and the in-house content engine. That is money and gear on the table from the builders, not the founders — and the founders know it and value it. The people writing the code are financially invested alongside the people whose names are on the venture: partners, not hired hands.

Building the audience with our own engine

A marketing company has to be its own best marketing. So the plan for demand isn't a separate discipline bolted on — it's the product pointed at ourselves. We build our audience the same way we build a customer's: with the engine. That makes the strategy unusually credible, because we're customer zero — the living proof that the thing works. What follows is the plan, staged honestly: the engine and the proof exist today; the channel and the full social engine are the build-out ahead, funded and sequenced like everything else in this part.

A dedicated YouTube channel — the proof reel

YouTube is the evergreen, searchable hub: where a skeptical prospect goes to watch the engine actually work. The plan runs on two formats that feed each other.

A well-rounded social presence — every platform with a job

Each channel earns its place with a distinct role — and because our customers live everywhere, we cover the long tail, not just the majors.

Content pillars — five reasons to follow

Every post ladders up to one of five pillars, so the channel is coherent instead of scattered.

The flywheel — why this compounds

The reason distribution is cheap here isn't discipline — it's structure. One engine produces the work, and one unit of that work becomes content on every surface.

It's self-reinforcing, and it's inexpensive to run because the engine and AI do the production. A single audit or working session repurposes into a long-form video, a batch of Shorts, a handful of clips, and native posts across every platform above — one unit of work, many surfaces. The content isn't a cost center bolted onto the product; it's a byproduct of using it.

Why it matters to investors

• A low-cost organic distribution moat — compounding reach that isn't rented from ad platforms and gets cheaper per impression as the library grows.
• Brand and credibility — the proof reel is the sales pitch; watching the engine work is more persuasive than any deck.
• Our own SEO / AI-search footprint — we practice what we preach and rank for it, which is itself the case study.
• An owned audience — a newsletter and subscribers we control, not an audience borrowed from an algorithm.

All of it produced at low marginal cost by the very product we sell — the cleanest possible alignment between what we make and how we grow.

The cost of the operation

A vision is only credible if the economics are. This closing chapter lays out the real cost of building, running and scaling this — what it takes to keep 200+ live websites and the engine behind them running today, who runs it, and how the cost curve behaves as it grows. The honest headline: the whole thing runs on a surprisingly lean, mostly-fixed cost base, and the unit economics get better with scale, not worse.

Part 1 — Hardware & infrastructure

The engine and the 200+ sites it audits and serves run on a compact, deliberately un-exotic stack. Below is the real breakdown, with directional cost ranges — these are representative, order-of-magnitude figures to show the shape of the spend, not exact private numbers.

Ranges are representative / order-of-magnitude, shown to convey structure — not exact figures. One-time hardware amortizes over years.

Part 2 — Manpower

Today, it's a lean human-plus-AI operation. AI agents do the bulk of the building, the content and the engine work; humans direct, review and approve — the same human-in-the-loop gate that governs every deliverable in the product. A very small team runs what used to require a full agency or an in-house department. That's the labor story in one line: the machines do the volume, the people do the judgment.

As it scales into a true SaaS business, the operation adds the roles a growing platform needs — but it adds them onto a base that's already automated, so headcount grows far slower than customers:

• Engineering — to build the integrations, the multi-tenant platform, and keep it reliable.
• Content / QA oversight — humans who set the bar and spot-check the machine's output at scale.
• Customer support — onboarding and helping self-serve customers succeed.
• Sales — for agencies, franchises and the multi-location accounts that don't self-serve.
• Operations — billing, security, and the back-office of a company serving thousands.

Physical footprint stays just as lean. Today the operation is mostly home-based — a near-zero facilities overhead, which is precisely the point of an AI-leveraged company: no server room, no floor of desks. The step-up as it grows is deliberate and modest — not a headquarters, but a real Austin home base, in the right place at the right time.

The Austin home base — Capital Factory + the UT campus

The plan here is concrete. Establish a seat at Capital Factory — plugging the company directly into Austin's premier startup hub, and with it the mentorship, investors, talent and deal flow that concentrate there. The first choice for the office / residence is the UT campus / downtown-north area — beside the McCombs School of Business and the densest concentration of AI activity in the city: the university's AI research, the talent pipeline it feeds, and the cluster of AI people and young companies around it, all within reach of Capital Factory downtown. East 5th Street is the strong second choice — an office / residence combo in the heart of the east-side creative and tech corridor, a short walk from Capital Factory and downtown — held as the fallback if the campus setup doesn't come together.

Why Austin — campus, downtown-north, and the east side

Austin — "Silicon Hills" — is one of the most dynamic AI ecosystems in the country, and the momentum isn't subtle. Roughly $7.19B of venture funding flowed into Austin startups in 2025, up about 65% year over year; AI / ML now accounts for ~32% of local startup funding; and the metro is on track to add on the order of 8,300 net new tech jobs in 2026, concentrated in AI infrastructure. (Figures are directional, drawn from widely-cited public reporting — the shape matters more than the decimals.)

And within that, the UT campus and the downtown-north corridor are the center of gravity for AI talent. The McCombs School of Business and the university's research labs feed a dense pipeline of AI people and young companies, and Capital Factory sits right downtown at the edge of it — which is why the campus / downtown-north area is the first choice for the home base. East Austin remains a strong second. Oracle built its waterfront campus on the east side; Cloudflare took a building in East Austin; and the East Cesar Chavez / Mueller stretch has become a walkable, creative tech corridor minutes from downtown and Capital Factory — with East 5th Street squarely in it. Either location is a considered bet — supported by the trend — on planting the home base where the talent and the energy are gathering.

Either way, the facilities line stays small. A Capital Factory membership and a modest live-work lease are a fraction of a traditional company's facilities cost — added deliberately, funded in part by the development team's own capital, and only as growth calls for it.

And the lean footprint is a deliberate echo of how the founders built before — from humble North Austin beginnings to an industry-defining platform, the full story of which is in The road forward. The difference this time: AI carries the operational load that once required a large team, so the operation can start leaner and stay that way.

Part 3 — Expansion as we grow

Here is the encouraging part of the math, and the reason this is attractive to fund: because the engine is software, the marginal cost of each additional site or customer is low. The fixed infrastructure — the droplet, the NAS units, the backup, the DNS, the platform subscription — amortizes across the entire network. Add the 201st site, or the 2,001st, and it costs a sliver of what the first one did.

The costs on the right — compute, storage, the content budget, multi-tenant infra, support and team — do grow with the business. But they grow far slower than revenue, because each new customer pays a recurring fee while adding only incremental load to an already-amortized base. That's the definition of healthy SaaS unit economics, and it's visible here already. It holds for the AI cost too: when the full Claude-powered assistant ships, each customer powers it with their own Claude subscription, so the per-question AI cost sits with the customer rather than on the operator's cost base — keeping marginal cost per customer low even as the product grows smarter.

The best argument for the product is that it's already ours

One thing to carry out of this briefing above the rest: we are our own first customer. The engine isn't a demo built to sell and set aside — it runs the marketing for 200+ of our own live websites, every day. We didn't build a marketing department and hope it worked; we built the one we wished existed, and then we pointed it at ourselves. Customer zero is us.

That is also the proof. Our own growth is the live case study — not a claim in a deck but a business actually run on the thing it sells. The flywheel is simple and self-proving: the same engine that markets a customer's business markets ours, and every improvement we ship for them we ship for ourselves first. Expanding our own footprint is the demonstration that it works.

Which is why this briefing is what it is — at once a product walkthrough, an operating manual, a business plan, and a blueprint for forming the company — how the operation stands up, how it finds a home (from working out of our homes today toward, as it grows, a small East 5th Street office and a seat at Capital Factory), and how we use the product itself to establish and expand the business. The road forward and the cost of the operation aren't appendices to the product — they're the same story told forward.

Appendix — every figure and factual claim in this briefing is backed by a real source: see the full Sources & references. And for the "why now" behind the moment we're building into, watch: why now.