How we work with small businesses.
Three tiers — Audit, Plan, Build. Most engagements stop at the Audit because the recommendation is “you don’t need to spend money on this.” That’s the point.
Phase 01 · 1 week
Audit
A senior engineer spends 90 minutes with you, then writes a recommendation on what to keep, drop, or swap.
- Walk through the AI tools and subscriptions you’re currently paying for.
- Pick 1–2 workflows that are slow, error-prone, or expensive.
- Honest answer on whether AI helps — and if so, the cheapest version that would.
- Sometimes the answer is “you’re fine, save your money.” That’s a valid outcome.
What you leave with: Written recommendation (5–8 pages) + 30-min walkthrough. $299 flat. Credited to a Plan or Build if you move forward.
Phase 02 · 2 weeks
Plan
If the Audit surfaced something worth building, the Plan is the prioritized roadmap with vendor picks, cost math, and a fixed-price quote.
- Ranked list of 3–5 automation candidates by impact and effort.
- Build-vs-buy recommendation per candidate — with the math shown.
- Vendor shortlist (free tier first, paid tier only when justified).
- If a Build is right, you get a fixed-price quote with the SOW already drafted.
What you leave with: Implementation Plan (10–15 pages) + working session + a Build quote if one fits. $999 flat. Credited to a Build.
Phase 03 · 2–4 weeks
Build
Fixed-price implementation. Wired into the tools you already use. Trained, documented, and handed off.
- Free 20-minute scoping call confirms the Build is a fit and locks the price (typically $2,500–$8,000).
- Built into your existing CRM, inbox, spreadsheets — no new platform unless genuinely required.
- Hard cost ceilings on any LLM usage. Kill switch ships with v1, not after the bill arrives.
- 30-minute training session + written runbook for whoever runs it.
What you leave with: Working automation, training, runbook, 30 days of email support. After that, optional small monthly tune-up retainer or use as-is.
Operating principles
Audit / Plan / Build is the how. These are the why.
Smallest version that solves it
If a free Zapier Zap solves it, that’s the recommendation. We don’t pad scope to justify a bigger invoice. Most small-business AI wins look boring.
Use what you already pay for
Most CRMs, email platforms, and accounting tools have AI features built in now — usually unused. Step one is checking whether the thing you’re about to buy is already in your stack.
Honest cost math, written down
Every recommendation includes the realistic monthly run cost — tokens, subscriptions, integration fees. No “it depends” hand-waving. If the payback isn’t obvious, we’ll say so.
Hard ceilings, not soft alerts
Every Build ships with a monthly token-spend cap that hard-stops the workflow when hit. Soft alerts get ignored. Hard stops force a conversation.
What we do not do
Four things we won’t take money to perform.
Hourly retainers
We don’t bill by the hour. The incentive misaligns the day it starts. Every engagement is a fixed price agreed before work begins.
Vendor kickbacks
We don’t take referral fees from software vendors. The recommendation is yours to act on without us; we don’t care which tool you pick as long as it’s the right one.
Building when buying works
If a $20/month off-the-shelf tool covers 90% of what you need, that’s the answer. Custom builds only when the off-the-shelf tools genuinely don’t fit.
Lock-in
We document everything in formats your team can edit without us. Workflows, prompts, scripts, runbooks — all yours, all editable. No “you need our retainer to maintain it.”
Start with the Audit.
The $299 AI Stack Audit is the cheapest way to see how we work. 90 minutes, one senior engineer, a written recommendation by Friday. If the answer is “don’t buy anything,” we’ll say it.