More services. Still the same clients.

Suppose you run a small web agency. Projects are ending, and new work hasn't come in. You're considering training the team to offer AI workflow services, starting with past clients.

It's a reasonable place to start. They know you. They've bought from you. What you learned about their businesses while building their websites may still be useful.

But what do you want that investment to buy? More work from existing clients, or revenue that no longer depends on just those few businesses?

That changes how much you can justify spending on the new service.

A Promethean Research interview roundup describes Fountain City's experience. Work from existing clients was expanding, while new-client acquisition received less attention. Then a major client of six years completed its digital transformation project. The engagement ended, leaving a substantial revenue gap.

Client projects end. Six years together doesn't mean another six years of work are waiting.

Back to the web agency. Suppose the AI service sells, mainly to the same clients. Revenue may recover. But if one client cuts its overall budget for outside services or stops working with you, both website and AI revenue may be affected.

More services. Still the same clients. You may earn more from them without gaining another group of clients.

If the revenue gap came from projects ending without new clients coming in, how much the new service can replace depends on the work those remaining clients can buy. When I review a stalled business, I look at project revenue alongside why the work ended. Can another project from the same client fill the gap, or do other clients need to replace that account?

A list of past website buyers can't answer that. An AI service needs current work to discuss, and the person buying it may not be the person who bought the website.

If that work exists, you can reach whoever buys it, the return after delivery costs and the additional investment is worthwhile, and someone can deliver it, expanding the relationship may make sense. It could even fill the immediate gap. You don't have to fix the whole old business first.

But restoring revenue from the same clients leaves you dependent on them. Selling those extra projects doesn't mean the next clients are lined up.

If you want the AI service to bring in new clients, you still need to work out where they are and how to enter their buying process. Training builds delivery capability. It hasn't built that route to new buyers.

With specific projects to discuss with existing clients, you can judge the training and delivery investment against that work. If you're counting on the service to replace ending projects over time, but have no route to new clients, you can't treat the whole expansion as if it already has a market.

Adding a service can be good business. Be clear whether you're investing in more work from existing clients or also giving the next clients a way in. Training isn't the whole cost of both.