Client-Side Lessons for AI Adoption: What Thirty Years in Consumer Insight Teaches Me

By Bolt

• article

• AI
• Artificial Intelligence
• Qualitative Research
• Automated Reporting
• Market Trends
• AI Moderated Interviews

This article is based on the webinar “How Leading Insights Teams Deploy AI to Thrive”. You can watch the full webinar for free at the link below:

I’m about three months into a new role at Bolt Insight, overseeing global client partnerships. Before that, I spent my entire career on the client side, most of it at Unilever, leading different consumer and market insight teams across different categories. I mention this because I think it matters where a view like this one comes from. I’m not a leading-edge AI specialist, and I’m not here to predict what’s going to happen in technology over the coming years. 

What I can offer is a perspective from someone who worked client-side for thirty years, tracked a lot of these changes as they happened, and has now jumped the fence to join the disruptors leading the transformation in our industry.

AI is Changing More Than the Toolkit 

AI adoption has already transformed the systems that insights and market research teams work within. Many people are tracking this on a daily basis already, in one form or another. But too often, the conversation still starts around toolkits. Should I be using AI for analysis? Should I be using AI for transcription or report generation? These are reasonable questions. People want to know what to actually do differently. But I think they’re the wrong starting point, because they treat AI as an addition to the existing toolkit, when actually it’s bigger than that.

What I mean is this. AI has restructured the systems through which things get discovered and evaluated. That’s not the same as adding a new piece of software to a process that otherwise stays the same. It’s a change to the environment the process sits inside. And if that’s true, then starting with “which tool should I use” skips over a more basic question: what has actually changed around you, and does your existing process still make sense given that change?

That’s the perspective I want to explore here. I come to this from the client side rather than the technical side, having spent most of my career inside organisations trying to make sense of new capabilities and land them properly. 

The Harder Question is Systemic 

This is really where the client-side view earns its place in the conversation. In my experience, the toolkit question is an easy one to reach for because it feels manageable. You can pick a tool, trial it, and report back on whether it saved time or improved a specific output. That’s a tidy exercise. 

The systemic question is much harder, because it asks you to look at the whole environment your insights function operates within, and to consider whether the assumptions built into your process are still valid. That’s not tidy. It doesn’t have a clear start and end point. But I think it’s the more honest question to be asking, and my experience has left me more comfortable sitting with that complexity than a quick toolkit answer.

Technology Only Works When It Fits the Organisation 

I saw something similar, on a much smaller scale, during my time working with agency partners. On one side, I worked with long-established agencies we’d partnered with for years, upgrading our methods together over time. On the other, I worked with a growing number of tech-driven startups, who often had genuinely cool technology, but sometimes lacked the market research expertise to go with it. 

In both cases, the real work wasn’t picking the tool. It was working out how a new capability actually fit into the organisation, what it meant for the people using it, and what needed to change around it for it to be useful. That, for me, is why Bolt felt like a sweet spot to step into, as the best approach marries deep client-side expertise with genuinely capable technology, and it’s only when you combine the two that you get success.

The toolkit framing misses that entirely. If you ask “should I use AI for analysis,” you get an answer about analysis. You don’t get an answer about whether the way your function discovers and prioritises what to analyse in the first place has already shifted. Based on what I’ve said about AI restructuring the systems through which things get discovered and evaluated, that shift is the more important one to notice first.

Four Principles for Approaching AI Adoption 

So if I were to recommend how insights professionals should actually approach this, drawing on client-side experience and the systemic view, it would look something like this.

1. Resist starting with the toolkit question, however tempting it is. Ask instead what has actually changed in the systems your function relies on. Has the way people discover and evaluate things already shifted, whether or not you’ve adopted any new software yourself?

2. Once you’ve answered that, look at which tools might make sense in response. The tool question is real and worth asking, but it should come second, not first.

3. Think about who is going to be responsible for landing them properly inside your organisation. In my experience, the technology on its own rarely fails. What tends to go wrong is the landing of it: whether the people bringing in the change understood the organisation well enough, and whether there was enough client-side expertise involved to make the new tool trustworthy and usable for the people who have to rely on it every day.

4. Look for that same combination when you’re choosing who to work with, whether that’s an established partner or a newer one. The value comes from combining deep client-side expertise with genuinely capable technology, not from having one and hoping the other follows.

Put the Toolkit in the Bigger Picture 

I don’t say any of this to be dismissive of the toolkit questions. They’re worth asking, and I’d rather people ask them than avoid AI altogether. 

But I think they should sit inside a bigger frame, one that starts with the system you’re operating in and pays proper attention to how change actually lands inside a real organisation, with real people who have to trust and use whatever comes out of it.

This article is based on the webinar “How Leading Insights Teams Deploy AI to Thrive”. You can watch the full webinar for free at the link below:

Author

Learn more about