^{READ TIME: 6 MINUTES}

Last week I rocked up to Mixpanel’s AI product launch in London to see what they’ve been cooking. (Don’t worry this isn’t an ad for them!)

While I thoroughly enjoyed the event and the hospitality was banging, I can genuinely say I walked away with a palpable sense of dread for what’s coming, and now possible, with AI in the analytics space 😱

I have a product analytics background. Having worked in E-commerce and Med tech, I’m used to being the person that Product Managers, designers and engineers come to with questions about product performance, as well as experimentation.

It’s probably my favourite thing in the entire data & analytics space.

The funny thing about this event, and my first alarm bell, was that although Mixpanel is a well known analytics tool and something I’ve always assumed was firmly in the tool kit of analytics teams, this entire event was aimed at Product Managers. 💀

WTF? Why are you doing an AI product launch for an analytics tool to PMs? In my experience, PMs are supposed to be data driven but often AREN’T, and are more likely roadmap driven than anything. They don’t do data!!

After sitting through each of the demos it became increasing clear.

If we set aside the reality that a lot of companies aren’t in a great place yet to leverage AI in their data work yet, the capabilities that Mixpanel demoed, were extremely impressive.

In a nutshell, their new product analytics tools are essentially an always on data analyst that provides users with a range of insane AI powered capabilities I’d not seen before.

I’ve summarised what they presented here so have a read!

In today’s newsletter, I’m taking a look at what it will take for data folks to survive on the AI frontier and the actionable steps you can take today to remain relevant now, and into the future.

💡 P.S. I’ve put together a free Data Leadership Assessment to help you figure out what kind of data leader you are. It only takes a few minutes so if you’d like to find out head here to take the free assessment.

For most of the history of data and analytics, the value data professionals provided was inseparable from the technical work itself. The ability to query, model, visualise, and interpret data was the product/service we sold. AI is systematically unbundling that.

1️⃣ Self-serve analytics is starting to work. "Self-serve" has always been a promise that never quite materialised. The tools were often too complex for non-technical users or the users were too ‘data illiterate’ to interpret the data correctly. That era is ending. Natural language interfaces mean Product Managers, marketers and business leaders can now get answers directly, without going through you. Mixpanel's demo was a live example of exactly this.

2️⃣ Dashboard and report production is being automated. A significant portion of what many data teams spend their time on, building, maintaining, and iterating on dashboards, is being absorbed by AI-native tooling. The output still exists. The human bottleneck (aka us) is being removed.

3️⃣ Exploratory analysis is no longer a specialist skill. The kind of ad hoc digging that used to require SQL proficiency and domain knowledge of the data model can now be replicated by AI agents given access to the right schema and context. The barrier to entry has collapsed.

4️⃣ Your seat at the table was never guaranteed. Many data professionals assumed proximity to the data meant proximity to decisions. Sadly and for reasons I speak about frequently, that was never really true. What it did was make the team a necessary stop on the way to an answer. But without that friction, there goes our leverage.

The data leaders who remain relevant won't be the ones who can run their queries fastest. It’s going to be the ones who know what questions are worth asking, why they’re worth asking, and how decision translate into business outcomes.

1️⃣ Become the person who understands the business, not just its data. Deep knowledge of how your organisation makes money, where it loses it, and what decisions actually move the needle is not something AI can replicate. That knowledge lives in relationships, context, and years of organisational pattern recognition. Invest in it aggressively.

2️⃣ Reframe your role around decision quality, not data production. The question to ask yourself is not "what did we build?" but "what decision did we improve, and what was the outcome?" Leaders who speak this language credibly will be far harder to replace than those who report on throughput.

3️⃣ Use AI to increase your own output and surface area. Rather than treating AI as a threat to your relevance, use it to do more of what used to take time. Become more efficient with it. Faster analysis, faster synthesis, faster scenario modelling. Data leaders who use AI to think at scale will dramatically outpace those who are still debating whether to engage with it.

4️⃣ Get closer to where strategy is actually made. If you are still primarily oriented toward the data team and its backlog, your exposure is high. The leaders that are going to survive this shift are embedded in commercial conversations, product strategy, and executive decisions. That means proactively creating those relationships now, before the technical case for your presence weakens further. Don’t delay!

5️⃣ Own the interpretation layer, not just the information layer. Data will be everywhere. Insight that is accurate but contextless will be cheap. What will remain scarce is the ability to synthesise what the data means for this business, in this moment, given what leadership is trying to achieve. That is a leadership skill. Build it.

👉🏼 BTW guys, this is exactly the sorts of skills and challenges I work on with my data leadership clients. I currently have more clients than I’ve ever had at the same time but I think I can squeeze one or two more in starting in July. If that’s something you’d like to explore, find some time with me for a 1:1 intro conversation here.

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

Tristan Burns
Find me on LinkedIn