Partnership brief · Confidential · 2026
Operations go in raw
and tangled. They come out clear.
Clearwell is the foundation layer that makes a business ready for AI — before a single agent is built. The market is funded, the timing is now, and the product is ~80% done.
Why now — the wave
Every board has mandated AI.
Every budget is funded.
2026 is the inflection year. The capital is already moving — at a scale software has never seen.
The money is committed. The question every executive is now asking: why isn't it working?
The evidence
The whole market is betting on this — and needs it to work.
Not a trend to ride. It's the boardroom's number-one bet, and careers are riding on it. Every figure below links to its source.
What the smart money sees
The prize was never the software budget. It's payroll.
The top funds are all pointing at the same thing — AI is going after the cost of labor, a market 20× bigger than software.
"Software is eating labor. The US software market is ~$300B. The labor market is ~$13 trillion."
White-collar services ≈ $6T — about 20× US enterprise software spend."Vertical AI is a 10× larger opportunity than vertical SaaS."
It competes for the 13% of GDP spent on labor — not the 1% spent on IT."2026 is the year of the agent employee — AI moves from a tool you use to a workforce you hire."
Services, not software, becomes the addressable market."Don't sell software. Sell the service. Do the work."
Spend on services dwarfs spend on software.But you can't put AI to work on operations that don't exist on paper. Someone has to lay the foundation first.
The trap
The spending is historic.
The results are not.
95%
of enterprise generative-AI pilots deliver no measurable business impact. — MIT, The GenAI Divide, 2025
60%
of AI projects abandoned through 2026 — for lack of AI-ready data and process foundations. — Gartner
63%
of organizations don't have, or aren't sure they have, the data practices AI needs. — Gartner
Root cause
It isn't the models.
It's the foundation.
You can't automate a process that was never written down, never stable, and owned by no one. MIT calls it the learning gap — companies can't fit AI into workflows that don't exist on paper. The technology is ready; the ground underneath it isn't.
The wall every growing business hits
Adapt and scale — or get out-executed by someone who did.
Every service business grows until it hits the same wall: the work lives in people's heads, so it can't scale without breaking. AI didn't soften that wall. It removed the middle ground.
People & dependency
- The owner is the bottleneck — the business runs them.
- When someone leaves, their knowledge leaves overnight.
- One key person can break a critical function.
- New hires take months to get productive.
- Seniors burn hours re-answering the same questions.
Quality & consistency
- The same task is done differently by everyone.
- Mistakes repeat — the org never actually learns.
- SOPs exist but nobody reads them; docs die on arrival.
- You can't delegate without things breaking.
Scale & cost
- Scaling means re-teaching everything from scratch.
- No idea what each process actually costs.
- Can't see which processes are followed vs ignored.
- Rework silently bleeds 5–15% of revenue.
The AI cliff — new, and fatal
- No documented processes = nothing for AI to run.
- Rivals who systematize then automate get a cost-and-speed gap you can't close by working harder.
- There is no third option: become AI-leveraged, or compete with someone who is.
The market's two wrong answers
Everyone is selling the layer above the problem.
Sell you an AI agent.
- Bolt intelligence on top of chaos.
- The agent automates the mess — then breaks the moment the process shifts.
- Nothing underneath is documented, so nothing is maintainable.
Send in developers.
- Expensive, slow, one engagement at a time.
- The moment they leave, the knowledge leaves with them.
- You're left with a custom build and still no foundation.
Both treat the symptom. Neither fixes the ground the business runs on. That's the gap Clearwell owns.
The insight
You never automate a mess.
You automate a process that is clear.
MIT's own data: teams that buy the foundation from a specialist succeed 67% of the time. Teams that build it themselves succeed about a third as often. The winning move is to buy the foundation. Clearwell is that foundation.
What Clearwell is
The foundation layer for AI-ready operations.
Your operations go in raw and tangled. Clearwell holds them clean, structured, and ready to run — so the AI you build on top actually works.
Why we're called Clearwell
Named after the outcome, not the chaos.
In water infrastructure, a clearwell is the final reservoir — where water has been fully treated, organised, and made ready, held in clarity before it flows out exactly where it needs to go.
Most operations software is named after the mess it manages. We're named after the transformation: raw in, clear out.
"Nobody wants to read an SOP. Everyone already knows how to follow a river."
The mental-model shift is the product. Your team doesn't open a manual — they step into a system that shows where work comes from, where it's going, and how it flows.
The architecture
A living water system: Streams, Flows, Assets.
The major arteries of your business.
Client onboarding. Campaign delivery. Each Stream is the parent that gives shape and context to everything inside it.
The executable, ownable units your team runs.
Where the doing happens — the steps, the owner, the tools, the templates. The unit a person actually picks up and runs day to day.
The sources beneath every Flow.
Reference materials, guidelines and templates that keep execution consistent. Water drawn from the ground up.
From the high ground of a Stream, down through the Flows, drawing from the Assets beneath — the system mirrors how water actually moves.
The product · ~80% built today
Map it once. Run it forever. Hand the right parts to AI.
Draft a Flow in minutes
Paste a Loom, an old doc, or just talk it through. The ✦ copilot writes the steps with you.
See the whole map
Every process rendered as a clear flow map — not a wall of text nobody reads.
A guide on every Flow
A chatbot trained only on your content answers "how do I do this?" so seniors stop repeating themselves.
Everyone's own book
Each person sees only the Flows for their role. Onboarding drops from weeks to hours.
Cost and usage, visible
See what each process costs and which ones are actually followed — not guessed at.
Automation, ranked by ROI
Clearwell tells you which processes are ready for AI, ordered by payback — on your own numbers.
Proof — it's built
Not a concept. A working product.
Click any screen to open it full-size. Working name shown is a placeholder.
The journey
From a tangled doc to a Flow the team actually follows.
Create on the left, consume on the right. The same clarity, end to end — and the foundation every future agent is built on.
Why people say yes
The value equation — maximized on every lever.
Scale without chaos — then AI leverage.
The exact thing every owner wants and can't buy off a shelf.
~80% built, proven architecture.
A working preview on the first call, plus a founding guarantee. They believe it because they see it.
First Flow in under a minute.
Full build compressed to 24 hours–5 days. Value is felt in the first session.
A DIY SaaS, not a project.
Replaces a 6–7-figure consulting engagement with software they run themselves.
Push all four levers at once and the offer becomes hard to say no to.
The value
The real cost isn't software. It's payroll running on chaos.
A mid-sized service firm doesn't lose a few thousand a month — it burns millions a year paying expensive people to re-explain, re-do, and hand-run work that documented, AI-leveraged operations would do faster, better, and for a fraction of the cost.
Clearwell costs under 2% of the leak — and turns the rest into margin.
Annual cost of "people-in-heads" operations · example: a 120-person, ~$25M services firm
The market — the calculation
We don't price against software. We price against payroll.
$7.2B on base subscriptions alone is the floor — plan tiers to $50K/mo, $30K+ builds, and recurring agent rentals push real revenue-per-account several times higher, all beneath a $6T services prize. We only need ~20 customers to reach $100K MRR.
The money model
Three revenue layers. One compounding account.
Tiered subscription · priced by scale (Streams & Flows)
Founding partners lock their tier rate forever · $99 refundable deposit on the call.
$5K–$50K/mo, tiered by scale.
Predictable MRR and the retention moat — the base every account starts on.
$30K+ one-time, per AI agent.
High-margin punch revenue — and each build feeds the moat.
$1–2K/mo each — companies hire several.
Recurring revenue stacked on top of subscription, and a marketplace other companies rent from too.
The moat
Every automation we build becomes an AI employee others can hire.
Clients fund the R&D. Each build is productized into a named agent — and rented out, monthly, like staff.
A client funds a $30K+ custom agent for their process.
We productize it as a named AI agent — with a defined job.
Other companies hire that agent for $1–2K/mo — like an employee.
Each agent is client-funded R&D and recurring revenue.
The library compounds into a marketplace of hireable AI staff.
Sequoia calls 2026 "the year of the agent employee." We own the foundation they get hired from.
To match it, a competitor needs the foundation, the agents, and the marketplace — years of work, not one feature.
Go-to-market
Validate the positioning. Win by outbound. Then pour fuel.
The product is ~80% built — we're not pre-selling vapor. We validate the positioning on live calls, then finalize the product around what the market already said yes to.
Fine-tune on live calls.
Outreach conversations lock the ICP, hook, offer and price — then we finalize the product to match.
100% outbound to $100K MRR.
Outbound is the engine. Content is support that makes outreach land — not the primary channel.
VSL + paid ads.
Once beta is polished — bugs fixed, gaps filled — we build a VSL and pump ads to scale past $100K.
Founding customers from outreach.
Positioning proven, repeatable.
~20 accounts, outbound-driven.
Pour fuel on a proven funnel.
Who runs it
Three lanes. No overlap. Each one covers a constraint.
Builds the product, plans the go-to-market, sets up the marketing funnel, and steps into final closing when it counts.
Strength: deep product + closingOwns the outreach engine that drives the first $100K MRR — the volume and the conversations that prove the positioning.
Covers: the outbound gapCreates content and orchestrates others to create more — fuel that makes outbound land now, and the raw material for the ads phase later.
Covers: reach & air-coverAaron's constraints are time and outbound. This team is built to solve exactly those — so each partner owns the lane that matches their edge.
Why this is hard — and why it's a moat
The hard part isn't the app. It's the architecture.
Fixing operations, making a business actually scale, then layering AI on top is normally a 6–7 figure consulting engagement. Turning that into a do-it-yourself SaaS — one a 40-person agency can run itself — took years of design. That compression is the moat.
Anyone can ship a prettier SOP tool. Almost no one can compress an entire operational transformation — fix, scale, then automate — into a product. That's the part that's already done.
Years went into making this simple.
Now it's time to take it to market.
I've spent years turning a 6–7 figure operational transformation into a product a business can run itself. It's ~80% built. What it needs now is an outbound engine and content — your edges. I'm offering 50% of the project, split between the two of you, to build this with me. Let's get to $100K MRR — then scale it with ads.
Sources
Every number on the record.
$2.59T 2026 (+47%) → $3.49T 2027 · software $453B→$638B · agents $206B→$376B · services $585B→$759B
MIT — The GenAI Divide, 202595% of pilots deliver no ROI · buying from specialists succeeds ~67% vs ~⅓ for internal builds
Gartner — Lack of AI-ready data60% of AI projects abandoned through 2026 · 63% of orgs lack the data practices AI needs
a16z — Software is eating laborUS software ~$300B vs US labor ~$13T · white-collar services ~$6T (~20× software)
Sequoia / Bessemer — agents & vertical AI2026 = "year of the agent employee" · vertical AI ~10× vertical SaaS (labor 13% of GDP vs IT 1%)
Grand View / US Census — market sizingBPM $61B by 2030 (20.3% CAGR) · ~200,000 US mid-market firms (Census SUSB / NCMM)
McKinsey — The State of AI, 202588% of orgs use AI (+10 pts) · 72% use gen AI · only 7% have fully scaled it
WEF · KPMG · Conference Board — boardroom demand 2026~80% of CEOs committing 5%+ of capital to AI · half say their job depends on it · AI = #1 investment priority
FactSet — S&P 500 earnings callsAI cited on 68% of Q4 calls (331 — a 10-year high) vs a 10-year average of 86