A B2B lead generation case study: three months into the AI Edit Method

I launched The AI Edit just over three months ago. Since then, the method I sell has been running on my own business, and this B2B lead generation case study shows what it produced: 49 leads, 37 of them qualified, 29 pitched, 21 deals closed.

Of those 21, six were high ticket. Two were under £50.

July was the first month we hit our run rate target for the year, and we've already closed enough business to hit it again in August and September.

Those are the headline figures. The rest of this post is the detail behind them, including the channels that have delivered nothing at all so far.

What counts as a lead, and why 49 is a smaller number than it looks

I've had far more than 49 conversations since April. Most of them didn't make the list.

A lead only counts if two things are true:

1. The person has a genuine need I can meet, and
2. Their business fits the profile of a client I actually want.

Everything else is a conversation. Pleasant, sometimes useful, occasionally flattering. Not a lead.

This matters because it changes every number that follows. Of those 49, 37 qualified in and 21 closed. That's a 72% close rate on the leads I pitched, which sounds implausible until you understand it's by design. If I counted every interesting conversation as a lead, the close rate would fall. The rate is high because the definition is strict, not because the selling is magic.

Most businesses inflate this number without meaning to. A downloaded guide becomes a lead. A polite reply becomes a lead. Then the pipeline looks healthy and the revenue doesn't follow, and nobody can work out why.

Deal sizes ranged from £23 to five figures

Two people bought a £23 mini course. Six deals were high ticket, up to five figure consulting projects and retainers. The rest sit somewhere between.

That spread is deliberate. A £23 course and a five figure retainer are the same method sold at different depths, and the small purchases aren't a distraction from the big ones. Some of them are how the big ones start.

Network, referral and LinkedIn produced almost every lead so far

Almost every one of the 49 came from three places.

My existing network. Referrals from people in that network. And LinkedIn, which has produced a meaningful share of them. A couple came through YouTube and similar content.

That's the honest picture at three months. The channels that work fastest are the ones where relationships already exist, and anyone launching a business who expects otherwise is going to have a difficult first quarter.

Search isn't delivering leads yet, but it's coming

Organic traffic to the site is growing consistently. Domain authority is up. We've had some traffic from AI chatbots, though not much.

Leads from either: none.

This is exactly what I expected. Search and AI visibility take time. Content needs indexing. Mentions need seeding across the web. And the models themselves lag badly, because Claude is currently working from training data that ends in January 2026. A website that launched in April simply isn't in there yet.

So I'm publishing this at three months with nothing to show from search, and I expect to publish a very different number at nine.

Ranking 11th for a keyword that gets 10 searches a month

We now rank 11th for "B2B lead generation training". That keyword gets around 10 searches a month, which by conventional logic makes it not worth targeting.

I disagree, for two reasons. Someone searching for B2B lead generation training is precisely the client I want, and what we sell is the best answer to that search. Ten of the right people beats a thousand of the wrong ones (and search volume estimates are usually conservative).

For a website that launched in April, ranking on page two for a commercially relevant term is a strong signal. I expect it to climb.

We're two months into the rollout stage of the AI Edit Method

The AI Edit Method has eight stages: Landscape, Audit, Blueprint, Ammunition, Channels, Rollout, Experiments, Review.

We've completed the first five. Landscape and Audit told us what had changed and what was working. Blueprint defined the customer, the offer and the metrics. Ammunition built the story bank. Channels set the order of play.

We're two months into Rollout, which is the stage where channels go live one at a time, each running properly before the next one starts. That's why the results look the way they do. LinkedIn is running. Search is set up and waiting. Experiments and Review are starting soon.

Everything in this post is the output of five and a half stages out of eight.

What's launching next: the AI Growth Community and the B2B Lead Accelerator

The YouTube strategy is being rebuilt, as is the Substack strategy. The AI Growth Community and the B2B Lead Accelerator both launch shortly. The B2B lead gen consulting offer is already off the ground.

The LinkedIn algorithm now knows what we're about, which took consistent posting rather than clever posting.

If you run a B2B business and this looks like the path you want to be on, then you know where to find us.

By Heather Baker, founder of the AI Edit.