We built something (that’s going to be) cool!
In short, this tool allows us to upload our research decks as a searchable research repository. Give the video a quick watch.
We want to know:
About to be cooking with oil over here!! 
So pumped about this team! Great work.
I love this guys! I always find repos helpful when you can trace back to the ‘why’? “Why did we set out to learn this” / “Why/How does this signal help us achieve outcome 1 or objective 2.5.1.2”, etc
Love that you’re trying to get signals unstuck from the deliverable they were first shared in–y’all are always discovering evergreen insights and a repo like this will help you leverage those learnings all the more.
Oh we’ve had these conversations for sure @vivekaleemelo. 
I’m curious what the most valuable piece of information for framing an evergreen insight is for you?
I second this! Would love any deets here @vivekaleemelo 
Thx for your patience y’all–got eloped last week so I’ve been all over the place.
With repos, I’ve seen them go wrong in two ways.
OR
From my POV, you avoid this by having different levels of detail for different audiences. One way of organizing this is to at least differentiating between internal repo vs client/stakeholder repo since their needs vary.
In terms of specific functionality outside of just how you organize a repo–you’d also want to personalize the tagging available–your researchers care way more about reviewing all ‘card sorts’ at once than your clients do.
Looove that perspective @vivekaleemelo . There’s a pull to find the Goldilocks point that applies to everyone, but the needs are so different. Shifting to the needs of the audience helps clarify the purpose of the data and the work it’s meant to facilitate.
And congrats!
Following up here, when it comes to AI we’re seeing some very cool capabilities for repositories.
As we begin to leverage agents to synthesize quant and increasingly qual data, there’s a real opportunity to capture those signals into the Repo as part of the overall workflow. By defining how we’re shaping what a ‘Signal’ is, we can carry that forward into our Repo, and do it while removing the intensive data-entry on the part of research or designers.
We’re excited to keep shaping these pieces, and @vivekaleemelo feedback about defining the audience is crucial. If we can train agents to pass information from one tool to another, the risk is capturing everything and sorting through it later, rather than defining what is important for who, and progressively disclosing based on the audience.
Love this… wondering if these are the most effective ways to frame research findings- thinking across workflows, I can see the need to reshape ideas based on what is important.
Coming back to this idea @vivekaleemelo! Thinking about this in terms of memory.
I’m wondering now if the problem moving forward is less about the tooling and fidelity, and more about the structure of the documentation. AI can fill in the holes if the context and depth are there. So context and structure become paramount @ben.
Hey! I hear what you’re saying and agree that the real value of AI is in the translating/transposing/reformatting and easy share out of insights (via an easily accessible ‘memory’). I see plenty of folks trying to use AI as the “maker”, to pull from the visual you used in your post, but current AI tools requires so much set up and guidance when it is asked to conduct its own data collection/analysis that I’m not seeing that hypothesized efficiency right away.
Where I do see AI shining brightly is in the tedious reformatting and rephrasing of key insights.
On my current team, our repo is manually maintained, annual reports are outlined and written by hand, and insights are slow to make it out of the UX space. AI could absolutely serve as that shared memory bank, where conversational prompts allow stakeholders with limited research skills to parse through our existing knowledge without having to book a 30-minute check-in with a researcher that may only remember half the story.
Or rather, it may not replace those human check-ins, but it could make them much more strategic. Rather than just rehashing a previous study’s findings, the conversation could focus on ‘what next’.
I think this is what turns research into a strategic asset. Using past findings to inform new decisions and reframe the problem. Unfortunately, most research disappears as soon as the immediate question is answered.
Hunches about what the future could be are typically not where most researchers feel comfortable, so I think the tooling is often created with a backward-looking focus.
