Back to Insights

AI projects · Strategy

An AI project starts with your bottlenecks

Media source

The starting point is already in your day-to-day: the bottleneck that costs you time every week, long before you choose an agent. Knowing what you want to change is enough to get going; the tool comes later.

Why is “an agent” the wrong first request?

“We’d like an agent. An agent that, um…” The line comes up in almost every first meeting with SME leaders. They know AI is going to change their business, they can feel time running out, so they look for something to buy. But an agent is an answer. Without a specific problem to hand it, it has nothing to solve, and the project gets stuck comparing tools.

Where is the right starting point?

In your bottlenecks, and you already know them. The quote that takes half a day to write and three days to approve. Meeting notes typed back into the CRM by hand. The technical lead everything depends on, who had better not fall ill. A good bottleneck is easy to spot: it comes back every week and someone is waiting on it. Every company has its own list. Each line is a possible AI project, with a cost you can measure and a gain everyone will see.

Optimise or build?

Every bottleneck offers two paths. Optimise what exists: the same work, done in a fraction of the time. Or build what didn’t exist: a service, an offer, a way of working that AI finally makes possible. The first path pays off fast; the second changes your place in the market. Either way, it’s your business that evolves, and that is what gets the tools.

Choosing the tool and how autonomous it should be comes afterwards, and it is decided from that bottleneck. At NODS, the first workshop is there for exactly that: drawing up your list and picking the line that will pay off fastest.

Explore our skills →