AI won’t solve wasted research in product development

AI can be useful in some research tasks — but there’s no substitute for researchers’ synthesis, curation, communication, and advocacy

Here comes that question again: “Won’t chatting with an AI solve this?”

Where ‘this’ is getting more done with critical research knowledge. A type of knowledge that can build up in tech organizations.

Augment yes. Solve no.

AI can be useful in certain research processes and aspects of knowledge management — but it’s not going to increase research-based planning on its own.

To start with, research content needs to be identified and consolidated in a way AI can consume. That preparation can take considerable effort. Not to mention the potential damage of ‘hallucination’ and accuracy issues of generative AI. Setting those baseline issues aside for now — there’s something more fundamental amiss.

Even if AI was technically perfect at retrieving and distilling varied research learning, there’s an underlying misconception about what it means to communicate and activate research. The role of researchers’ points of view is being heavily discounted.

If AI is acting as a form of automation, what is it automating? Or if it’s more of a back and forth, what is it augmenting?

Consider the concluding steps from an individual study. When researchers and their collaborators complete a project, they’ll typically present the learning and (re)connect with various decision makers about key points. They don’t stand up in front of audiences in silence, waiting to field any question without having shared anything. They’ve built a point of view on what people should know and what’s important to act on, and they push that perspective in an engaging and structured way.

The same can be true of community’s collected research learning. Insight-seekers who access research tooling shouldn’t show up to an empty chat box. Your research community can present a rich menu of options to explore, rather than a proverbial blank page. What’s more, your researchers don’t have to wait for a decision maker to initiate their hunt for insights — you can assemble and push out cross-study intelligence on what your colleagues should know.

AI can support finding patterns and standardizing sets of knowledge — but it’s not a replacement for researcher synthesis, curation, communication, and advocacy.

There’s just too much that contemporary AI doesn’t ‘know’ about your customers, your products, and your organization to deeply offload this effort — especially if you want to position your research community as a central input.

Because, after all, product planning is a social process.

It goes without saying that product professionals won’t fully automate the creation of their product roadmaps. There are too many difficult, multi-faceted choices to make. One of those decisions is which research could be relevant. There’s no magic tech that’s going to automagically identify and integrate insights into product plans. Augment yes. Solve no.

Much more to come on these topics in forthcoming ‘Stop Wasting Research’ book for Rosenfeld Media

Sign up for email updates — monthly, at most
Connect on LinkedIn