A Principles Memo for Investors — July 2026

The trades run on trust and gut feel. We are building the machine that runs them on truth.

1OAK is the operating system for the outside trades — installed, managed, and financed. What follows is how we think, what we believe, and what the numbers say.

By Harvey Jenkins & Marcus Phillips, Founders, ENGAGE AIS

Scroll
01
The Problem, Stated Plainly

Principle 1. You cannot fix what you refuse to see clearly.

Small outdoor contractors run six-figure construction projects on paper and instinct. They lose jobs they should win because estimates are slow and ugly. They win jobs they should lose because pricing is a guess. And when they buy software, they abandon it — not because it lacks features, but because nobody sets it up and keeps it running. The failure is not effort. It is architecture.

$0B
U.S. Home Services Spend
0 Years
Construction productivity flat since 1945 — manufacturing grew 10–15x
0
Net new workers needed in 2026 — a majority just to replace retirees
Sources: Houzz · McKinsey Global Institute · Associated Builders & Contractors
02
The Machine

Principle 2. Every business is a machine — inputs, cause-and-effect, outputs.

We built the machine first, inside a real contractor. An address enters the machine. Out comes a parcel map, a 3D site model, a pre-visit brief, a measured takeoff, an auditable estimate, a photorealistic design the homeowner can fall in love with — and then deposits, draws, credits, and payments, all on rails.

Diagram A · The Pipeline
01
Address
02
Geocode
03
Parcel
04
Footprint
05
Satellite
06
Roof
07
Terrain
08
AI Site Brief
09
Human Takeoff
10
Deterministic Estimate
11
Design Mockup
12
Payment Rails
Legend
Machine
Human
03
Responsible AI, By Design

Principle 3. Trust AI to recommend. Never let it do the math.

The question everyone asks is how to put AI into the daily cycles of a trades business responsibly. Our answer is structural, not aspirational: AI recommends — products, assemblies, language. Deterministic formulas compute — every quantity, every dollar, with the math shown on every line. Humans decide — the estimator confirms every shape; the owner approves every send; recording a consultation requires consent; crews and customers can never see data that isn't theirs. A model's opinion never becomes a homeowner's invoice.

MIT field research found that the overwhelming majority of business AI pilots deliver no measurable return — not because the models fail, but because unmanaged deployment fails. Our architecture, and the human layer behind it, exist to prevent exactly that.

Diagram B · The Division of Labor
AI Recommends
  • Products
  • Assemblies
  • Language
Formulas Compute
  • Quantities
  • Dollars
  • Every line, shown
Humans Decide
  • Every shape confirmed
  • Every send approved
  • Consent required
Source: MIT Project NANDA, State of AI in Business, 2025
04
The Human Layer

Principle 4. Software that nobody runs is a liability wearing a subscription.

Every 1OAK operator gets a named account manager, in perpetuity — we seed the catalogs, tune the labor rates, wire the lead sources, and review the numbers monthly. This was never economical before. AI changed the ratio: one account manager can now credibly run 40–60 operators, at a cost of roughly $120–$180 per operator per month inside a ~$940 blended monthly revenue. The human layer is not a cost problem. It is the moat — it turns the industry's #1 churn driver, abandonment, into our retention engine.

The largest field study of AI at work found productivity gains of 15% on average — and roughly 34% for the least experienced workers — because AI packages the tacit knowledge of top performers. That novice effect is the engine of our account-manager ratio and our bilingual crew instructions.

40–60
Operators per account manager
70%+
Blended gross margin
95%+
Best-in-class net dollar retention
Source: Brynjolfsson, Li & Raymond, Quarterly Journal of Economics, 2025
05
How the Machine Learns

Principle 5. Pain plus reflection equals progress. Machines can be built to reflect.

Every completed project compares estimate to actual. The variance retunes the labor rates. The next estimate is sharper. Multiply that by hundreds of operators and you get something no competitor can copy: a living, regional database of what outdoor work truly costs — and, eventually, the underwriting engine for financing the trades.

Diagram C · The Compounding Data Loop
EstimateBuildActualsRate RetuningSharper EstimateThe CompoundingData Loop
06
The Economics

Principle 6. Underwrite the money flow, not the feature list.

We monetize like the best vertical platforms: a subscription earns the right to sit in the money flow; the money flow pays for years.

Chart 1 · ARPU Per Operator Per Year
≈ $0
Illustrative planning assumption
SaaS
$5,988
Managed service
$625
Payments (50bps · $450k GMV)
$2,250
Financing rev-share
$1,500
AI usage
$480
Channels
$500
Chart 2 · 3-Year Illustrative Model
ARR ($M) & Operators
GMV ($M) — flowing through the platform
Year 1
40 operators
$0.3M run-rate
$25M GMV
Year 2
175 operators
$1.7M run-rate
$115M GMV
Year 3
500 operators
$5.8M run-rate
$340M GMV

By Year 3 the platform sits in the flow of a third of a billion dollars of project value.

Illustrative planning assumptions.

Proof of Category

The category is proven: the leading indoor-trades platform reached ~$961M revenue (+24%) with ~1/3 from fintech products and retains 95%+ of dollars. The outdoor trades — bigger tickets, estimation-first, zero incumbent depth — remain unclaimed.

Sources: ServiceTitan FY2026 Results · Andreessen Horowitz · BCG/Adyen
07
What Could Kill Us

Principle 7. Name your weaknesses before your critics do.

One flagship customer today — the contractor we built inside; the raise funds the next fifty. Parcel data is statewide in North Carolina today — the provider abstraction for national coverage is already built. The platform was built at startup speed on modern low-code rails — the funded plan hardens it: senior engineers, multi-tenancy, SOC 2. We would rather show you the risk register than have you find it.

Risk 01
One flagship customer today
Mitigation

The contractor we built inside. The raise funds the next fifty, in a repeatable playbook.

Risk 02
Parcel data is statewide (NC)
Mitigation

The provider abstraction for national coverage is already built. Coverage is a wiring problem, not an architecture problem.

Risk 03
Built at startup speed on low-code
Mitigation

The funded plan hardens it: senior engineers, multi-tenancy, SOC 2. Speed first, discipline second — in that order.

08
The Ask

Principle 8. Make the bet explicit.

We are raising $1.5–$3.0M in seed capital for 18–24 months of runway to reach: 150+ operators, $1.5M+ ARR, proven fintech attach, and a second region.

Use of Funds
35%
Product & Engineering
25%
The Human Layer
20%
Go-to-Market
10%
Fintech & Compliance
10%
Reserve
Request the Data Room

We built the machine inside a real business first. Now we intend to install it across an industry.