It is an inevitable question, and an important one. A single study rarely exists in isolation. The challenge is that the wider context often sits somewhere else. At an agency, that other project might have been run by another team or even another agency. In-house, it might be sitting in a deck on someone’s drive, alongside dozens of other studies that were incredibly useful at the time and have barely been opened since.
There is a lot of energy and money spent on getting more data, faster. Faster fieldwork, bigger samples and more frequent studies all matter. But some of the most expensive waste comes from knowledge organisations have already paid for that now sits in silos and is difficult to bring back into the decisions being made today.
Finding the insight is only half the job
Finding the relevant evidence is only one part of what a good researcher does. The other is interpretation: understanding the question being asked, placing the evidence in the right context and translating it for the person who needs to use it.
A head of marketing, a brand manager and a sales lead can be trying to answer the same underlying question, but they may ask it in three very different ways and need three very different answers. Good research requires interpretation in both directions: understanding what someone is really asking, then deciding which evidence matters, how much context they need and how to explain it in language that makes sense for the decision in front of them.
Finding the insight is only half the job
Finding the relevant evidence is only one part of what a good researcher does. The other is interpretation: understanding the question being asked, placing the evidence in the right context and translating it for the person who needs to use it.
A head of marketing, a brand manager and a sales lead can be trying to answer the same underlying question, but they may ask it in three very different ways and need three very different answers. Good research requires interpretation in both directions: understanding what someone is really asking, then deciding which evidence matters, how much context they need and how to explain it in language that makes sense for the decision in front of them.
That is where research expertise becomes so valuable, and where technology has an opportunity to make it more accessible. The principles that guide good research, from framing the right questions and challenging assumptions to interpreting evidence in context, have traditionally lived with experienced researchers. AI gives us a way to build more of that expertise directly into the tools people use, helping anyone across an organisation approach questions with the rigour and discipline of a good analyst.
For me, that is the real opportunity. To take the knowledge, methods and judgement that make research valuable and translate them into technology that helps more people ask better questions, understand the evidence and make informed decisions. The more of that expertise we can build into the platform, the more accessible good research becomes.
AI has made this easier, and riskier
Language models have made it dramatically easier to ask questions of information and receive an immediate response. That creates opportunity across almost every profession, but it also creates a new problem: a plausible answer can sound authoritative even when the evidence behind it is weak, incomplete or missing.
That risk becomes especially important in research. If an AI tool fills a gap in your evidence with general knowledge or a confident guess, it can create false certainty at exactly the point where someone is trying to make an evidence-based decision.
The standard therefore has to be higher than whether AI can produce a convincing answer. In research, it needs to show where the answer came from, point claims back to the evidence behind them and be explicit when the available research cannot answer the question.
The goal is not to replace judgement. It is to give people an answer they can interrogate and understand, while keeping the evidence visible enough for a researcher to challenge, contextualise or take the question further.
AI has made this easier, and riskier
Language models have made it dramatically easier to ask questions of information and receive an immediate response. That creates opportunity across almost every profession, but it also creates a new problem: a plausible answer can sound authoritative even when the evidence behind it is weak, incomplete or missing.
That risk becomes especially important in research. If an AI tool fills a gap in your evidence with general knowledge or a confident guess, it can create false certainty at exactly the point where someone is trying to make an evidence-based decision.
The standard therefore has to be higher than whether AI can produce a convincing answer. In research, it needs to show where the answer came from, point claims back to the evidence behind them and be explicit when the available research cannot answer the question.
The goal is not to replace judgement. It is to give people an answer they can interrogate and understand, while keeping the evidence visible enough for a researcher to challenge, contextualise or take the question further.
What we built
That thinking is behind Insights Assistant, the newest part of our Assistant suite, which we launched this month to every Ideally customer, on every plan. The suite already helps customers build studies and draft presentations. Insights Assistant extends that experience across the research they have already run.
You can ask a question of your research in the same way you might ask a colleague. Insights Assistant searches across the studies you have access to, brings the relevant findings together and lets you keep the conversation going. You might ask, “Was that true for older consumers too?” or “How did these two ideas compare?” and explore audiences and comparisons wherever the research supports them.
Every answer links back to the studies behind it, so you can see where the evidence came from. If your research does not cover the question, it tells you rather than defaulting to the general knowledge or plausible-sounding answer a typical LLM might provide. When a chart explains the answer more clearly than words, it can build one from a single source with the citation attached.
For teams using Ideally Canvas, category data is also part of that same conversation. You might ask, “Which of our concepts fits the biggest demand zone?” with study evidence and Canvas evidence separately cited so it is always clear what came from where.
We built Insights Assistant with people who work in insights, and I spent a lot of time shaping the research approach behind how it responds to questions. It incorporates the research methods I use every day, alongside the marketing science thinking I believe people working in insights should understand.
The ambition was bigger than creating a search engine for old reports. We wanted something that could approach a question with the discipline of a good analyst, while keeping the evidence visible and leaving room for human judgement.
Where this goes next
The version of this I’m most excited about is the one we’re building towards: an assistant that holds the full context of what an organisation knows about its consumers. This month brings three steps towards it.
The biggest is the Ideally MCP. Teams can now ask questions of their research from inside Claude, ChatGPT and the other AI tools they already use, with no exporting or copying and pasting.
The second is brand context. Teams can upload the briefs, decks and brand plans that usually sit outside the platform, so answers reflect how the business actually works. These sources are treated as background, not evidence, which keeps a clear line between what the research shows and what the organisation already believes.
The third is depth. Ask a bigger question and the assistant builds a full presentation across your studies, Canvas and brand context, with findings, significance and citations on every slide.
For me, that gets at a frustration I have had throughout my career. There have always been pieces of information you know exist somewhere, but cannot access at the moment you need them.
We are not there yet, but for the first time in my career, I can see a path towards it. It starts with treating research as a body of knowledge that should keep growing and becoming more useful, rather than a stack of finished projects we slowly forget.
