Product Research · Example Report
How IntelAnvil combined research with potential users, AI-assisted analysis, product judgment, feature prioritization, and commercial-interest testing to develop the product direction for ThoughtTapestry.
A new consumer product can have dozens of plausible features before there is good evidence about which ones potential users actually value.
ThoughtTapestry faced this problem. The basic concept was established: a system for recording thoughts, ideas, observations, and questions over time and using AI to understand them together. But that still left major decisions about what the product should actually do.
IntelAnvil used a staged research process to move from this broad possibility space toward better-supported product decisions.
The research was conducted with UK-based participants recruited through Prolific. Different stages examined what potential users wanted from the concept, how they prioritized competing capabilities, and their interest in the resulting product proposition.
The first stage deliberately avoided presenting participants with a predetermined feature list.
In the initial study (n = 40), participants were given the underlying product premise and asked what they would want the system to do. This revealed recurring needs before participants were exposed to our own product ideas.
Participant Question
Imagine that over the course of a year you had recorded hundreds of ideas, observations, questions, and thoughts about things you were interested in or working on. If a system could understand all of them together, what would you want it to do with that information?
Percentages indicate the share of responses that explicitly mentioned each theme. One response could contribute to several themes.
The responses were analyzed using Text Response Hub, an AI-assisted qualitative analysis system developed by IntelAnvil. The analysis identified recurring needs while preserving less common responses that could point to valuable product opportunities.
The most common responses were not simply converted into a feature list.
We combined three sources: recurring participant demand, interesting individual ideas, and independent product hypotheses worth testing.
AI helped structure the evidence. IntelAnvil used human product judgment to determine which ideas represented meaningfully different capabilities, where apparently similar suggestions could be combined, and which possibilities deserved inclusion in the next experiment.
| Capability | Popularparticipant demand | Interestingindividual ideas | Independentproduct thinking |
|---|---|---|---|
| Organize my thinking | ●●● | ● | ● |
| Show me the big picture | ●●● | ●● | ●●● |
| Discover connections I missed | ●●● | ●●● | ●●● |
| Bring back relevant forgotten thoughts | ●● | ●●● | ●●● |
| Show how my thinking has evolved | ● | ●●● | ●● |
| Develop my ideas further | ●●● | ●●● | ●●● |
| Generate new ideas from my existing thinking | ● | ●●● | ●● |
| Find the ideas most worth pursuing | ●●● | ●●● | ●● |
| Turn my thinking into next steps | ●●● | ●●● | ●● |
| Find gaps in my thinking | — | ● | ●●● |
| Find contradictions in my thinking | — | — | ●●● |
| Challenge my assumptions | — | — | ●●● |
The indicators show why each capability entered the next stage. They are qualitative provenance indicators, not participant ratings.
This avoids two common shortcuts: building only what potential users already know to request, and using research merely to validate ideas the product team already prefers.
Once the possibility set had been constructed, the competing capabilities were tested directly against one another.
In the prioritization study (n = 36), participants repeatedly chose which of two capabilities they would personally find more valuable. Pairwise comparison forced participants to make trade-offs between plausible features rather than allowing every attractive capability to receive a high rating.
Capability preference
Relative preference strength reconstructed from participants' pairwise choices between competing product capabilities.
The strongest preference was not for organizing notes, but for turning accumulated thinking into practical next steps. Several other highly ranked capabilities also went beyond conventional note management: developing ideas, identifying which ideas are worth pursuing, and generating new ideas from existing thinking.
Percentages are relative preference-strength estimates reconstructed from participants' pairwise choices, rather than the percentage of participants selecting each capability.
Feature preference does not automatically translate into demand for the overall product.
In a separate study (n = 30), participants rated their interest in using an app with capabilities like these and were asked whether they would consider paying for it.
Overall product interest
Participants rated their overall interest on a seven-point scale.
The distribution leaned toward the positive end of the scale, while also showing that the concept did not appeal equally strongly to everyone.
Willingness to pay
Participants who were willing to pay entered their monthly value as an open-ended amount rather than selecting from predefined price ranges.
Median stated monthly value among respondents willing to pay
£10
The value of the process does not come from any single research method. It comes from combining different kinds of intelligence around the product decision.
Potential users contribute evidence that cannot simply be manufactured inside the product team. Their responses reveal recurring needs, unexpected possibilities, preferences between competing capabilities, and reactions to the overall product proposition.
AI helps turn individual responses into usable structure, identify recurring themes, preserve less common observations, compare alternatives, and support analysis across larger sets of evidence.
IntelAnvil determines what to investigate, which observations deserve further attention, which independent hypotheses should remain in play, how competing possibilities should be tested, and what the resulting evidence means for product development.
Human intelligence enters the process in two distinct ways.
Participants contribute human judgment as evidence. IntelAnvil contributes human judgment through research design, synthesis, experimentation, interpretation, and product direction. AI is allocated alongside both where it can improve the quality or efficiency of the process.
The result is not simply a survey or a feature ranking. It is an evidence base for deciding what to build, what to emphasize, what to test next, and where further investment is most justified.
If you have a product concept but are uncertain which features people value, how the product should be positioned, or whether there is enough interest to justify further development, IntelAnvil can design and run product research around those questions.
The research is built around the decision you need to make. It can include early discovery with potential users, feature and capability prioritization, concept comparison, positioning tests, willingness-to-pay research, prototype evaluation, or a combination of these.
Useful product research does not necessarily require a large study or research budget. A focused experiment can answer a specific question: which of several features people value most, whether one product concept is preferred over another, how potential users understand a proposition, or whether a prototype solves the problem it is intended to solve.
Focused studies can start from around €100, with broader product-research projects typically costing several hundred to €1,000+, depending on participant recruitment, sample size, research design, analysis, and reporting.
Not every useful question can be answered well in a single experiment.
The ThoughtTapestry research on this page is an example. Before testing product capabilities, we first had to find a way to present an unfamiliar concept clearly enough that participants understood what they were being asked to consider. The early research informed the next stage, and the resulting capability set could then be tested more directly.
This kind of iteration is often where much of the value comes from: run a study, learn from the results, improve the question or product hypothesis, and use that knowledge to design the next experiment.
A project can therefore begin with one focused study and expand only when the findings justify another round.
IntelAnvil can also conduct the research independently and produce a standalone report documenting the research question, participant sample, methodology, results, analysis, and conclusions.
The report can be delivered privately, made available through a password-protected page, or published on the IntelAnvil website.
This can be useful when the research is intended not only to guide internal product decisions, but also to provide an independent analysis that can be shared with investors, partners, management, or other stakeholders evaluating the product or opportunity.
IntelAnvil can help turn it into a research design, recruit suitable participants, run the study, analyze the evidence, and deliver the results.