How to choose an AI automation partner.
Pick the AI automation partner who can show you a live working deployment, gives you a fixed-scope quote (not hourly), tells you the monthly running cost in writing, and offers a 30-day fix guarantee. Match the size of the partner to your size: solo freelancer for one-off needs, specialist agency for ongoing builds, enterprise consultancy only if you're past $10M in revenue.
Last updated
| Partner Type | Typical Cost | Best Fit |
|---|---|---|
| Solo freelancer | $1K – $8K / project | Cheap, but handoff & maintenance risk |
| Specialist agency (us) | $1.5K – $15K / project · $3.5K/mo retainer | Built for service businesses $1–5M |
| Enterprise consultancy | $50K – $250K+ / project | Overkill for sub-$10M businesses |
| In-house hire | $140K – $220K / year + tools | Worth it only past ~$10M revenue |
Why choosing the right partner matters more than the tech
AI automation is now commoditized at the tool level. Anyone with a credit card can spin up OpenAI, Claude, Zapier, and a CRM. What separates a successful project from an expensive failure isn't the underlying technology — it's how well the partner scopes, handles edge cases, and hands off the system.
We've seen $40,000 builds fail because the team never asked the right scoping questions. We've also seen $3,000 builds that pay for themselves in 30 days because the partner understood the business first.
The 9-step buyer's checklist
- Define the outcome, not the technology. Write down the business outcome you want — hours saved, leads recovered, customers responded to. Not "I want AI." Without this anchor, every partner sounds the same and you'll buy on personality.
- Check for operator background. Has the partner ever run a real business? Have they handed off systems before? AI engineers without operator experience often build technically correct systems that don't fit how you actually work. Operator-led shops tend to scope tighter and walk away from the wrong projects.
- Ask for a fixed-scope quote. Avoid hourly billing for the initial build. A serious partner can scope and quote in fixed terms after a 30–60 minute call. Hourly is fine for ongoing tuning. Hourly for the initial build means open-ended cost and no incentive to ship fast.
- Demand a live working example. Ask to see one running deployment with real data — not a screenshot, not a case study. A partner who can't show you something live either hasn't shipped recently or is hiding something.
- Understand the handoff. What do you receive when the project is done? A walkthrough call, written documentation, admin access to all systems, recovery procedures for when something breaks. If the answer is "just trust us" or "we hold the keys," walk away.
- Get monthly running-cost numbers in writing. Demand a line-item estimate of monthly API/model costs, integration platform fees, sending fees (email/SMS), and any new tooling. Most agency-build complaints come from surprise running costs nobody warned the client about.
- Confirm a fix guarantee. Reputable partners offer 30–60 day fix guarantees on shipped scope. Read the conditions. "Bug-free" guarantees are red flags (nothing is bug-free); "we fix what we scoped within X days" guarantees are honest.
- Watch for AI maximalism. If the proposal says "AI" for every step including the boring ones (saving a row to a database, sending a confirmation email), they're optimizing for impressive-sounding language, not for your outcome. Good partners use AI where it matters and plain code or no-code where it doesn't.
- Trust your gut on communication. If the early calls feel slow, evasive, or full of unexplained jargon, the project will feel the same. The best partners are direct, explain in plain English, push back when your idea is wrong, and tell you when something isn't worth building.
Red flags: walk-away signals
These come up across enough disappointed buyers that they're worth memorizing.
| Red Flag | What It Actually Means | What to Ask Instead |
|---|---|---|
| "We use proprietary AI models" | Marketing language, not a real advantage | Ask what they actually run under the hood |
| Hourly billing for the build | Open-ended cost, no incentive to ship fast | Insist on a fixed-scope quote |
| No live demos, only case studies | You can't verify it actually works | Walk away or get a 1-call demo |
| "We can't share running costs upfront" | You'll be surprised by API bills | Get a written monthly estimate |
| Vague handoff plan | You'll be locked in for tuning forever | Require docs + admin access at delivery |
| Promises "100% AI, no humans needed" | Edge cases will burn you | Insist on human checkpoints for high-stakes outputs |
| 12+ week timelines for one automation | Either over-scoping or under-resourced | A focused automation should ship in 1–4 weeks |
Questions to ask on the first call
We've been on both sides of this call. Here's a tight set of questions that surface the answers you need within 30 minutes:
- Can you walk me through one live deployment, screen-shared, right now?
- What's a project you turned down recently, and why?
- If our automation breaks at 2am on a Saturday, what happens?
- What's the monthly running cost for a typical lead follow-up build at our volume?
- What's your fix-guarantee window and what does it cover?
- If I asked your last 3 clients what they'd change, what would they say?
- Can I keep all the credentials, prompts, and documentation when we're done?
A confident, experienced partner answers all of these without hedging. If most of them get fuzzy, you have your answer.
Match partner size to your business size
One of the biggest mistakes we see service businesses make is over-buying. A $2M home services company hiring a $250K enterprise consultancy will spend 6 months in discovery and come out with a system nobody can maintain.
- Under $1M revenue: a freelancer or DIY (Zapier/Make) usually beats hiring an agency. Your needs are simple; the overhead isn't worth it.
- $1M – $5M revenue: specialist agency or a small operator-led shop. The sweet spot for scoped automation work.
- $5M – $20M revenue: specialist agency on a retainer, plus eventually a part-time in-house ops automation lead.
- $20M+ revenue: in-house team plus enterprise consultancy for governance-heavy projects.
