From Knowledge Store to Decision Support: What the New Mandate for Insights Leaders Really Means

TL;DR

• The insights function is under more pressure than ever — stakeholders want faster, clearer, more directive answers, and they want them embedded into decisions, not delivered afterward.
• Most organizations already own the intelligence they need. The problem is activating it at the moment it matters.
• AI raises the stakes in both directions: it accelerates what's possible, but generic tools introduce new risks around trust, provenance, and organizational context.
• Our new research report draws on 14 in-depth interviews with insights leaders and managers to map what's breaking, what good looks like, and what CMI leaders should do next.

There's a version of the insights function that everyone agrees on in principle. It connects what the organization knows to the decisions it needs to make. It turns years of consumer research, market intelligence, and strategic learning into answers people can act on. It helps the business move faster with more confidence.

Most organizations aren't there yet. And the gap between that aspiration and the daily reality — fragmented intelligence, systems built for storage rather than activation, stakeholders who've stopped waiting and started using whatever gives them an answer first — is widening.

Stravito commissioned independent researcher Natalie Delgado to understand why. Over 14 in-depth interviews with insights leaders and managers across FMCG, Retail & Apparel, Quick-Service Restaurants, and Consulting, a clear picture emerged of what's actually getting in the way, what AI is changing, and what the insights function needs to look like to stay relevant in H2 2026 and beyond.

The result is our new research report: From Knowledge Store to Decision Support: The New Mandate for Insights Leaders. Here's what it covers.

The mandate has shifted — but the infrastructure hasn't caught up

The core mission of the insights function hasn't changed. Understand the customer. Help the business make better decisions. What has changed is the expected speed and form of the answer.

Stakeholders no longer want a comprehensive report. They want the "so what" — a clear, directive recommendation they can act on before the meeting ends. They trust insights teams to handle the rigor. They just don't want to sit through it. As one insights leader in the study put it: "Tell me what to do, or give me confidence in this decision."

This creates a tension that runs through every finding in the report. Teams must move faster without sacrificing the trust, context, and quality that make their answers worth acting on. And the systems most of them rely on — built for document storage and retrieval in a pre-AI era — were never designed to help them do that.

The result is a function caught between what stakeholders now expect and what the infrastructure can actually deliver.

AI is both the pressure and part of the problem

It would be easy to frame AI as the solution here. And in some ways, it is. Better synthesis, faster retrieval, the ability to surface relevant evidence across scattered sources — these are real gains, and the report takes them seriously.

But the research also makes something else clear. Generic AI tools — the ones already embedded in Teams, Outlook, and SharePoint — are creating a new kind of risk that insights leaders need to think carefully about.

When a stakeholder can get a fast, confident-sounding answer from a general AI tool without leaving their inbox, the business case for slower, more rigorous alternatives erodes quickly. Convenience wins. But those answers often lack source traceability, organizational context, and any guarantee that the content behind them is approved, current, or even accurate. As the report notes, generic AI can contribute to a culture that demands fast answers — but not necessarily trustworthy ones.

This is the central challenge the report surfaces: AI has raised expectations around speed and accessibility to a point where insights teams can no longer compete on rigor alone. But the answer isn't to abandon rigor — it's to build infrastructure that delivers both. Trusted answers, at the speed of business.

What's actually getting in the way

The research identifies three barriers that show up consistently, regardless of organization size, industry, or the tools already in place.

• The activation gap. Organizations have invested heavily in consumer research, market intelligence, and strategic learning. The intelligence exists. But it sits fragmented across repositories, disconnected workflows, local systems, and static reports built for storage rather than continuous use. Teams spend significant time searching, rebuilding context, validating sources, and recreating work that already exists somewhere else in the organization.
• The trust gap. As AI enters the workflow, the stakes around source quality get higher, not lower. A plausible-looking answer drawn from an unfinished document or an outdated study doesn't just waste time — it carries real decision risk. The report found that many teams are so uncertain about whether what they're looking at is current and approved that they commission new research rather than act on what they already own. The organization ends up paying to recreate intelligence it already has.
• The last-mile problem. Even when the right intelligence exists and someone finds it, it doesn't reliably reach the person making the decision when they need it. Insight travels when it's pushed. Most systems are built to store, not to surface.

What the next generation of insight infrastructure looks like

The report outlines four capabilities that define where insight infrastructure needs to go — and they're a long way from a better search bar.

• Synthesis. Not a list of documents, but a coherent, evidence-backed narrative answer to a specific business question. The ability to connect studies, cluster themes, cite sources, and present a clear view — so teams can spend their time on interpretation and recommendation, not retrieval.
• Contextual memory. Capturing not just what the data says, but why decisions were made. When people leave or move roles, the reasoning behind past choices often goes with them. A system that preserves that context — why a flavor was rejected, what drove a strategic shift — prevents the organization from repeating expensive mistakes.
• Active guidance. Proactively stopping duplicate work before it starts. If an existing study already answers the question, the system should say so before a new brief is written. If a team is designing a survey, the system should surface what the organization already knows. This is where insight infrastructure becomes a strategic control, not just a search tool.
• Proactive awareness. Pushing relevant intelligence to the right people based on market, brand, topic, and role — without waiting for someone to think to look. The UK team completes a study. The US team should know about it. That shouldn't require a manual email chain.

These aren't aspirational features. They're the capabilities that separate a system that supports decisions from one that just stores research.

What this means for CMI leaders

The report closes with five practical recommendations for insights leaders navigating this shift. They're worth reading in full, but the through-line is this: the insights function's value is no longer defined by how much research it can produce. It's defined by how effectively it can activate the intelligence the organization already owns — and get it to the right person, in a form they can use, at the moment the decision is being made.

That's a different kind of infrastructure. And building it starts with understanding where the current model is breaking down.