^{READ TIME: 5 MINUTES}

This topic has been creeping into many of my client conversations lately. Seasoned data professionals are being instructed to just hand over their analysis to AI systems by execs. “can’t we just put all of this information into AI and just get the answer?” They’re asking.

Every single client I’ve spoken to has talked about trying this, but have been getting absolutely useless responses from their model. But do you think that’s deterred the execs? Nope, they double down and blame the data team for doing something wrong and order them to try again, properly this time.

This leaves the data team in an absolute bind. Play along and disappoint your execs when the results come back weak, or push back and risk looking like a blocker to AI progress holding the business back?

AI is where the attention and the budget sit right now. When executives brings you an AI request, they are testing whether you can turn the technology into something useful for them or the business.

Most have already decided that AI is the silver bullet is before they speak to you: a machine you point at a problem to get an answer. So their requests arrives that way. "Can't we just put all this data into AI and get the answer?" They expect the tools to work magic. And if they don’t it’s not the magic that’s broken, it the so called ‘wizard’ in the data team.

AI is good at some things and poor at others. Accept this framing from the execs and you will usually disappoint them. Reject it and you look like a sceptic trying to hold the business back. Handling this well is one of the defining skills for data leaders right now, and very few are doing are nailing it.

1️⃣ An executive instructs your data teams to put a business question to AI expecting a clean, confident answer to come back.

2️⃣ Your team does the work properly and returns with an honest read: what the analysis shows, where it is uncertain, and what the data cannot tell them.

3️⃣ The executive hears those caveats as the team making excuses or not being up to the job, rather than as a sign the work was done carefully.

4️⃣ They decide the data team could not get AI to deliver, but never stop to ask whether it was fair to expect AI to answer that question in the first place.

5️⃣ The cycle repeats. Each time it does, your business leaders trust the team's work a little less, and your credibility with them drops.

Left unchallenged, this belief keeps the blame pointed at the team, and your credibility with business leadership steadily erodes.

Find out what kind of data leader you are with my free assessment. Get your results immediately! Click here for the assessment

You can stop the silver-bullet framing from taking hold. These steps keep the expectation realistic and protect your standing with business leadership.

1️⃣ Reset what AI is for. Tell business leaders plainly what AI is good at and what it is not. It helps with generating ideas, summarising large amounts of material, and drafting. It is unreliable as the final word on a complex business question where the data is incomplete at best. Say this early, before the work starts, so the expectation is set before any results are expected.

2️⃣ Define the problem before you touch any tool. When someone asks you to point AI at something, ask what decision the answer will inform and what they will do differently once they have it. What behaviour will change? If no one can answer that, the request is not ready, and AI will not fix that. You already apply this to normal data requests. Apply it here too.

3️⃣ Put the limits in writing. Before you start, send a short note to the people who asked. Set out what you are going to do, what AI can and cannot tell them on this question, and how you will check the results. This is your early disclaimer, so that if the results come back predictably terrible, then at least you’ve set expectations early.

4️⃣ Take the lead on where AI is used. Rather than waiting to be handed AI tasks, tell leadership where AI is worth using in the business and where it is not. Reference your business strategy when doing this. The person who sets that direction is the one leadership turns to on AI, and that is the position you want.

This is exactly the types of issue my data leadership clients are struggling with. If you too are dealing with executives who expect AI to hand them the answers, and you want help managing that without damaging your credibility or your budget, then why not book some time with me for a free informal chat.

👉 Book a coaching call

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

Tristan Burns
Find me on LinkedIn