^{READ TIME: 5 MINUTES}

Right now nearly every board on earth has decided it needs an ‘AI strategy’. Some bozo went to a conference, read a report, or got spooked by a competitor, and now there is a line in the leadership notes that says "develop our AI strategy". Pretty often, that jobs lands square in the lap of a data leader like you. Bummer.

In most companies, the thing being produced under that title however, is not a strategy. It is basically just a shopping list: Tools to trial, a copilot rollout, a chatbot for the website, or a vague mandate from the exec to "embed AI across the business".

💡 Freshen up on business strategy here
💡 Freshen up on data strategy here

I’ll be shocked if you’re reading this and your head hasn’t fallen off from the hardcore nodding along that you’re doing…

To help you put your hear back on straight, in this edition, I’m going to explore why and AI strategy isn’t a strategy at all, and why, if not treated carefully, can be extremely damaging to you and your team.

When a company says it has an AI strategy, alarms bells ring in my head. When I look closely I usually find a shopping list of tools the org wants to buy, a set of disconnected pilots owned by different teams, and a few bottom up use cases.

Every one of those might involve AI. But none of them is a strategy. They describe activity the business wants to encourage, but not a direction. They tell you what the org wants to do, but nothing about what it is actually trying to win at, or how it plans to win.

Why it isn't a strategy

AI is a capability. It is a f**king powerful one, but it sits in the same place as your data platform, your tooling, and your team. Capabilities serve a strategy. They are not the strategy themselves.

Strategy is about choices. Where the business is competing, how it intends to win, and what it is deliberately choosing not to do. AI only becomes a strategic enabler when it is pointed at the answers to those questions. "Use AI to approve claims faster than anyone in our market, because speed is how we win" is a strategy doing its job. "Roll out Copilot" is not…

What this does to you and your team

Unfortunately, we’ve all been here before. If you’re old enough you probably remember the various crazy hypes that all seemed to get their own dedicated strategies, but largely went nowhere: Big Data Strategies, or Cloud Strategies, or Blockchain Strategies, or Pokemon Go Strategies (maybe?).

Just like these examples from the past, when an AI shopping list gets treated as strategy, the cost start to show up in a few predictable ways 👇🏻

1️⃣ You end up with a bunch of pilots that get demoed once and then soon abandoned. The money and staff energy has been spent but nothing actually changed and the solutions haven’t been adopted.

2️⃣ Your team gets judged on how many AI things it has tried (i.e. activity metrics) rather than what any of it changed for the business (actual business outcomes)

3️⃣ With no real choices underneath it, every new tool has an equal claim on your budget, and you lose any solid basis for saying no.

4️⃣ Someone senior eventually asks what all the AI spend actually delivered, and heads start rolling when all you have to show for it is a series of failed AI pilots.

You do not need to own the business strategy to sort this out. You need to stop letting AI be treated as a strategy in its own right, and put it back in service of the choices that matter.

Start with the business strategy you already have

You probably will not be the only person writing the AI strategy. This will be a group effort, and as a data leader, your role is to be the person in the room who keeps it tied to reality. The most useful thing you can do is refuse to treat AI as a separate plan, and instead anchor every idea to what the business is already trying to achieve in its market.

That means starting from the business strategy that already exists, rather than inventing a shiny new one with AI scribbled on the side. If AI turns out to be the right tool for a particular goal, use it. If it does not, that is a perfectly good answer too. The test is always the same: does this move the business closer to something it already cares about?

So for any AI initiative on the table, these are the questions worth asking:

🔹 Does this make us a more attractive seller to our customers?
🔹 Does this help us lean into the strengths we already have in the market?
🔹 Does this reduce the impact of a weakness that is holding our business back?
🔹 Does this improve critical processes that makes us more competitive or more profitable?

If an idea cannot answer at least one of those clearly, it is usually a solution looking for a problem, and you are well within your rights to say so. If it can, you now have a reason for doing it that your executives will actually recognise, because it is framed in the terms they already use rather and that they provided. That is how you move from contributing to the AI strategy to genuinely shaping it.

👉🏼 This is a large part of what I do with the Heads of Data I coach. Not coming up with AI use cases, but getting clear on the business direction, pressure-testing what lands on their desk, and leading the conversation as a key person of influence.

If your company is busy writing an AI strategy and something about it feels off, that instinct is right, and it is worth a chat. Grab some time with me here.

Free resources:
Data Leadership Playbook · Data Career Compass · Data Leader Assessment

Tristan Burns
Find me on LinkedIn

Last time, I asked you where are you at with public speaking?

Here’s how you responded: