Your first AI service offer — a decision framework

This is the version to keep open while you put together your first AI services offer. Each section is one decision: a question to settle, the criteria from the article, the options compared, and what disqualifies each one. At the end you get a one-page decision path to work through.

The rule that runs through all of it, in one sentence: step onto the ladder at the lowest rung you can already handle, not the one you'd like to jump to.

Decision 1 — which rung to start on

Question: which rung of the offer do you sell first?

The article frames an AI business as a ladder. You earn each rung by standing on the one below it — you don't skip straight to the top.

Two definitions that come back later. An audit is a review of the business aimed at finding what's worth improving. A workflow is a repeatable sequence of steps — for example "take the inquiry, quote it, send the offer" — that software can partly take over.

Selection criterion: start on the rung you've already stood on in practice. The third rung (retainer) is the best place to end up — steady revenue from a few clients — but it's where the path ends, not where it starts.

What disqualifies starting at the top:

• Trying to open with a project or a retainer without proof or a relationship behind it — this is where the industry usually starts, and exactly why it stalls.
• You've built impressive things but never worked with a client.
• Jumping to a project produces a feeling of being a fraud. At rung zero the environment is calmer.

You feel like a fraud because you're trying to sell something you've never done. Drop a rung, onto ground you actually know, and the feeling starts to fade.

Decision 2 — is a paid hour the right first step

Question: does selling a paid hour fit your situation at the start?

The article gives four reasons it does. Check them like criteria — if they hold, you have your confirmation.

What disqualifies this framing of the hour:

• Treating the hour like tech support and a side income — something separate from "real" work. The article calls this the mistake.
• Treating the hour as an endpoint rather than the first step of a longer relationship. Framed that way, you'll pull fewer engagements out of the conversations.

Put it this way: the hour isn't tech support, it's the first installment on a partnership.

Decision 3 — what you're actually selling in the session

Question: how do you frame what you give the client during the hour?

This is the core reframe. The hour isn't "I'll teach you to use an AI chatbot." It's help building an AI operating system — one place where the business collects its data, its expert knowledge, and its processes. That's what stops the business depending entirely on the owner's head.

Criteria that hold this framing together:

• Only the owner can build the system — they're the one who knows the customers, the processes, and the business's knowledge. You bring tool fluency.
• You're selling a combination: their industry knowledge meets your ability to run the tool.
• During the session your job is managing overwhelm, not showing off knowledge. The owner will be excited and overloaded at the same time.

When questions come up about concepts like an agent — an AI program that carries out tasks step by step on its own — or data connections, explain them plainly, slow down when someone loses the thread, and ask whether it's still clear.

What disqualifies your approach in the session:

• Posing as a search engine instead of a person who solves problems.
• Pretending to know when you don't. Say it plainly: "I don't know, I'll check and get back to you." That's not weakness, it's the role.

Decision 4 — how to land your first clients

Question: which channel do you start with, and in what order?

The article lays this out as seven steps, each following from the one before it. The order itself is the criterion here — the "outward" channel only comes once you have proof.

1. Teach people you know first. Show three trusted people how the AI tools work. No selling, no fee — safe practice, and a chance to notice where your explanations are unclear.
2. Reach out to business owners you know. Offer an hour "at your own expense" to build experience and show them how to set up their system.
3. Ask for referrals. After a few sessions, ask whether the client knows someone else this would help.
4. Join communities. In AI and business groups, someone is asking for help every day — answer, and offer a call.
5. Show what you're doing. Write up a case on LinkedIn or your own site. You don't need to be an influencer — this is about the trail that builds trust.
6. Grow relationships toward bigger engagements. No pressure. At some point the conversation will drift on its own toward "we could build this together."
7. Only go outward once you have proof. Local businesses, industry meetups, events. This is a step for someone who already has a few sessions behind them.

Every one of these steps can start this week.

What disqualifies rushing the order:

• Going outward (step 7) before you've collected a few sessions and some proof.
• Pushing for a bigger engagement instead of waiting for the relationship to mature into one on its own.

The limits of this path

Three limits the article states outright — keep them in mind while making the decisions above.

• Hours don't scale indefinitely — it's still time for money. Selling hours is a way onto the ladder, not the destination.
• Early on you're not optimizing for cash, you're optimizing for the number of conversations you've had and your own confidence. The money shows up in six months if you put in the reps now.
• The feeling of being a fraud never fully disappears — for many experienced people it comes back almost daily. What counts is reps: a business owner talking to someone after twenty consultations can tell the difference, even if you can't feel it yourself.

One assumption holds the whole plan together: build your own system first, because it's hard to teach someone something you've never done yourself.

How the path to a project usually happens

Follow-up question: where does the first project come from, if you're selling hours?

The typical path: you build an automation live with the client, session after session, until by the third meeting the client asks whether you'd just build it for them as a project. There's no pitch — there's presence, help, and trust.

Not everyone will say it outright. Once you know the business's operations, you can gently steer the conversation:

I can see this problem in your operation, and I know a workflow that could solve it — let's talk it through in the next session, and I'll come back with a full proposal.

---

Decision path

One page to work through before you send an offer.

1. Rung. Start at the lowest one you can already handle — usually rung zero (hours). Don't jump to a project or a retainer without proof.
2. Fit of the hour. Check the four criteria: trust threshold, proof threshold, paid discovery, hard to replace. Don't treat the hour like tech support.
3. Framing of the offer. You're selling help building an AI operating system, not a course. In the session you manage overwhelm and say "I don't know" when you don't.
4. Landing clients. Go through the seven steps in order: people you know → businesses you know → referrals → communities → showing your work → growing relationships → going outward. The outward channel only comes with proof.
5. Limits. Hours don't scale; early on you optimize for reps, not cash; the fraud feeling fades with the number of conversations.
6. Base assumption. Build your own system first.