AI Go-To-Market Strategy: The ICP-First Playbook for B2B AI Companies, with Jon Cheney

Jon Cheney built his company in three days. Then one phone call showed him he had built it for the wrong buyer. We had Jon on Edbound With Kinner, hosted by Kinner Sacchdev, to talk about it. His AI go-to-market strategy is easy to say and hard to do: decide who buys, and why, before you write any code.

Jon runs the Gen AI Proficiency Institute. Before that he founded Seek and raised about $13 million for it. He says this company hit a million-dollar run rate in about six months.

One thing to know up front. Jon sells AI assessments, training and advisory, not software. His order of decisions still applies if you sell an AI product to businesses, and that is what this post covers.

Why AI go-to-market strategy starts with the buyer

An AI go-to-market strategy is your plan for who buys your AI product, why they buy it, and how you reach them. Jon's version has three parts: settle the buyer first, test with real conversations, and price the outcome before you scale any channel.

He starts with one question: "Who's gonna buy this and why are they gonna buy it?" Asking it after the build, he told Kinner, is "totally backwards." 

It also changes how you build. Say you know the product will be freemium. Then the AI tool sets up your database and onboarding that way from the first prompt. Decide later and you rebuild.

How to define your ICP with an AI interview, one question at a time

Jon does not ask the AI for a business plan. He asks it to interview him.

1. Write two or three sentences about the idea.
2. Tell the model: "interview me one question at a time."
3. List what it should cover: name, brand, ideal customer, marketing, monetization and MVP.
4. When the interview ends, ask for a business plan.

One question at a time works because ten at once feels like a test, he said. A single question feels like a conversation, and the follow-ups get better. His first interview took two to two and a half hours and ended in a 2,000-word plan.

Then he pastes the plan into Grok, which can search X, and tells it not to sugarcoat. The feedback goes into a second draft. He says the whole loop now takes about 15 minutes.

Kinner runs a similar process in Notion. He puts a plan on a page and asks different models, Grok one minute and Gemini the next, to comment on it.

For more on testing demand, see our startup idea validation framework and the Jobs To Be Done framework.

Why a free quiz funnel stalled and one LinkedIn message changed the model

Jon's first plan targeted individuals. A free "AI Q" test, gated by email, scored visitors. High scorers could buy a $59 certification. Lower scorers were sent to courses at $19 or $29.

He launched it with a Facebook ad around 3 a.m., right after the three-day build. By Monday, 140 people had clicked and nobody had bought. His Stripe account was still in test mode. It took two minutes to fix.
With checkout working, he started messaging people on LinkedIn. About five of them. One replied, and one call became a $15,000 contract for a company of about 500 employees. "I just sold 500 employees all at once in one call," he said.
So he had a B2C funnel and a B2B buyer. Six weeks in, he had booked $180,000, none of it recurring, from trainings and courses.
Kinner made a similar move with his own company. His team served coaches and creators first, then shifted to selling to businesses. His take: the size of the deal changes everything.

A five-step AI go-to-market strategy for B2B AI companies

Here is Jon's sequence, in order.

1. Name the buyer before the build. Use the AI interview to settle your ICP, your pricing model and your first offer.
2. Get outside feedback on the plan. Run it through a model with live market data and ask for blunt answers.
3. Launch something small. Jon went from idea to live in three days, with a working privacy policy, real social pages, and a domain and name that looked credible to buyers.
4. Message real buyers within days. Five direct messages gave him his first customer. Our guide to buying intent signals shows what to look for. Let the first call tell you whether you picked the right buyer.
5. Price what the client gets. A former neighbor who ran a pool construction company asked what happens after the training. Jon made up a fractional Chief AI Officer retainer on the call. It was $15,000 a month, in a range he gave as $10,000 to $25,000.

On step 5, Jon said: "to get the second fifteen thousand dollars, you gotta deliver." A second client signed the same day. About three months after the first pitch he had six clients and a million-dollar run rate. He hired because he couldn't serve six clients alone, and the company now has 28 employees.

For another look at large deals, read our enterprise selling playbook.

How Jon gets inbound demand with no paid marketing

Jon spends nothing on paid marketing. He posts on LinkedIn every day (about 26,000 followers), speaks at events, and goes on podcasts. For the posting side, see our LinkedIn personal branding strategy and founder-led marketing strategy.

He writes every post himself. "AI scales, but it looks fake and people can feel it," he said. People who find him this way already want what he sells, so the calls are easier.

He also works on getting found in AI search. GEO tools run about $1,000 a month, he said, and he did not want to pay that. So he asks ChatGPT, Grok and Claude to research how those tools work, has them check each other, then gives the result to Replit to audit his site. He repeats it until the gaps close.
His one concrete tip: put a date in your headings. LLMs read it as a sign the content is current. Some buyers now tell him, "ChatGPT said you guys were the best."

Kinner told Jon the process could save a team ten to fifteen thousand dollars in tool costs, and asked him back for a longer session on it.
To go deeper, read our guides to AI search visibility, LLM SEO for startups and the inbound-led outbound strategy.

What B2B AI product teams can take from this

Jon's numbers come from a services company, so they won't match yours. Three things carry over. Settle the buyer before the build. Test that choice with direct messages in the first week. Pay attention to what your first customer asks for next. If your first deal shows the buyer is a team or a company, change how you sell. The product may be fine.

What to do next

• Write down your ICP and the first question that buyer asks. If you can't, run Jon's interview this week.
• Message five people who fit that profile before you build a funnel.
• After your first paid deal, ask the client what they need next.