The first 90 days of becoming an intelligent business

Ask any leader what's stopping their business becoming more intelligent and you'll rarely hear "the technology". You'll hear "we don't know where to start", "we've tried and it didn't stick", or "we're too busy running the business to change it". Fair answers, all three. The good news is that the early stages of becoming an intelligence-led organisation don't have to be dramatic. Done well, they're quiet, structured and quick to show progress.

Here's what a realistic first 90 days looks like, before any big deployment and while the business is still running as normal.

Before any technology gets introduced

The single biggest mistake is to start with a product. Nothing derails a programme faster than a licence bought before anyone has agreed what "better" looks like. The right first move is to spend a couple of weeks with the business as it actually is. That means:

• Sitting with the teams whose work matters most to customers, and watching what a normal day looks like.
• Listing the tools already in use, and who uses them for what.
• Mapping the top ten processes end to end, not from a diagram but from reality.
• Understanding where information lives, and whether anyone trusts it.
• Capturing the frustrations people have stopped complaining about because they've become normal.

Discovery like this is quick, cheap and enormously informative. It's also the moment where the business case starts writing itself, because the friction points that come up will be the same ones that appear on the roadmap.

Weeks one to three: discovery and framing

The first three weeks are for gathering evidence and shaping a shared view of the current state. The output shouldn't be a hundred-page report. It should be a short document with three sections: what the business is trying to achieve over the next twelve months, where friction is stopping that, and a shortlist of the processes worth improving first. If the leadership team can agree on that, most of the hard thinking is done.

Weeks four to six: pick the first wins

The best first projects share a few characteristics. They touch a process people already care about. They can be delivered in weeks, not quarters. They produce numbers you can put on a slide. And they don't require heroic change management to land.

Common candidates include automating a chunk of manual reporting, deploying Copilot to a team that produces a lot of documents, sorting out permissions and oversharing before a wider AI rollout, or consolidating two disconnected systems that keep causing rework. Whatever you pick, aim for two or three - enough to build momentum, not so many that any one of them stalls.

Governance, ownership and stakeholder buy-in

Programmes drift when nobody owns them. In the first month, name an internal sponsor with the authority to make decisions and unblock things. Pair them with a working group small enough to move quickly - typically representation from IT, operations and the business function most affected by the first projects. Set a rhythm: a short weekly stand-up to remove blockers, a monthly review to check the numbers, a quarterly steer with the leadership team.

Governance should feel light but visible. Its job is to keep the programme honest, not to slow it down.

Weeks seven to ten: deliver the first project

With the sponsor in place and the shortlist agreed, weeks seven to ten are for delivery. Keep the scope tight. Get something in the hands of real users as early as possible. Measure the baseline before you start and the outcome as you go. Expect surprises - they're the point. The purpose of a first project isn't to be flawless, it's to be finished, useful and learned from.

Communicate as you go. A short update every fortnight to the wider business does more to build appetite for the next phase than any formal launch would.

Weeks eleven and twelve: consolidate and plan the next quarter

The final fortnight is for taking stock. What did the first project deliver, in numbers people can trust? What did you learn about how the business responds to change? Which teams asked to be next? Which stakeholders became advocates?

That evidence becomes the foundation of the next quarter's plan. The second quarter tends to be much easier than the first, because the story has stopped being theoretical.

Building momentum without overwhelming teams

The most common failure mode isn't going too slow. It's going too broad. Trying to modernise every process at once burns out the teams doing the work and dilutes the leadership attention that any single project needs. A better shape is a rolling programme: one or two focused improvements in flight at any time, delivered by people who can see the finish line.

What success looks like at the end of the first 90 days

• Leadership shares a single view of where the business is losing time and money.
• Two or three quick wins are live, with measurable outcomes.
• An internal sponsor and working group are in place and functioning.
• The next quarter's shortlist is agreed and funded.
• The business is talking about intelligence in terms of outcomes, not products.

That's a strong first 90 days. It's not a transformation, but it's the beginning of one - and it's the foundation everything that follows will build on.

Where to start

If you'd like a partner to run discovery, shape the first shortlist and deliver the initial wins, that's exactly what our AI consultancy and wider intelligence services are built for. The first 90 days matter more than most, and getting them right makes everything after them cheaper and easier.