Back to blogJuly 24, 2026

Let research keep moving with the question

Keep questions, evidence, research roles, and follow-up checks with the same work. Let Bots update the evidence as material arrives, so you can focus on the decision.

Get Started Download

You are deciding whether to build a service for a new kind of customer.

The first round of research goes well. You gather information, read interviews, and ask AI to help prepare a report. The conclusion: there is an opportunity, but two key assumptions need testing.

A week later, three new pieces of customer feedback arrive. The report is still there, and so is the question. You need to find the original assumptions, recall which evidence was reliable, and decide what the new material changes.

That work of picking up the thread repeatedly consumes attention.

Keep research around an unresolved question

A report captures an understanding at one point in time. The decision may come only after several rounds of investigation and experimentation.

Someone needs to look after the unresolved questions in between: which claims are still guesses, which now have support, and what new evidence would change your mind.

You can bring that responsibility into Ducoro:

Help me keep researching whether these customers are worth serving. Start with these interviews and the public sources I specify, and list the evidence for and against. When I send new interviews here, update the relevant judgments. Every Thursday, bring together the questions that still need testing. Flag evidence that could change the direction when you find it.

Example · New interviews change the pilot; willingness to pay remains untested

Once the scope and rhythm are agreed, the question has a place to keep developing. You provide new material and adjust the question; the Bot handles organization, comparison, and further checking.

Let different Bots contribute to the same question

One Bot can take responsibility for progress within the Space and involve helpers with different strengths.

One organizes interview transcripts, another checks external sources, and a third looks specifically for counterexamples. They can use different models and tools suited to their roles, returning results to the same place.

The work needs a common question and traceable material. When you see a conclusion, you should be able to find the interviews, dates, and conditions behind it. When two Bots disagree, the basis of that disagreement should also be visible.

The Bot responsible for the work can assemble a developing research record: the current judgment, supporting evidence, opposing evidence, and what is still missing.

The value of several Bots should show up in clearer evidence and better-defined disagreements.

Show which judgment new material calls into question

Perhaps the first interviews suggested that customers lacked a particular feature. Two later interviews suggest they care more about how easy their existing process is to carry out.

A useful update would place both groups of evidence together and explain what each supports. It could then propose a testable question: if the process improves without adding features, would these customers keep using it?

The Bot can prepare the next interview outline or a small experiment. You decide which assumption to test and how much time to invest.

Your decision stays with the work. The next round has a direction and a record of why that direction changed.

When there is little new evidence, an update can be short: what was checked, which conclusions still lack additional support, and which questions remain open.

Several questions need a shared management view

You may be researching customer demand, pricing, and a new distribution channel at the same time. Each question can have its own Space, materials, and participating Bots.

Views let you bring research for a particular product together, review priorities, or look across upcoming checks. Open one responsibility when you need its specific evidence and discussion.

Example · The demand pilot has direction; pricing continues and channel research rests

Different topics retain their context, while a shared management perspective helps you decide which to advance this week, which already has enough evidence for a decision, and which is worth pausing.

The View provides a way to look across the work. Research continues within each Space.

Put research to work for your next decision

If each request to an assistant begins with a question and ends with a report, you still maintain the research thread. Someone has to connect new material, bring several assistants' findings together, and track which assumptions you have already set aside.

Ducoro organizes that work as an ongoing responsibility. A Space keeps the question and its progress together, several Bots contribute, and Views help you oversee different investigations. A scheduled check is one action within it. Material, judgments, and follow-up arrangements make the research continuous.

Begin with a question you have not yet decided. Give a Bot the first set of material and ask it to distinguish support, opposition, and uncertainty. Then agree on what new evidence should drive the next round.

More of your attention can stay with the important question: given the evidence I have now, what am I going to do?

Series introduction: Your Bots are getting better. Who keeps the work together?

Next: Turn your ideas into content that arrives on time

Keep the next job moving

Manage your Bots in Ducoro, and hand over the next job in plain words.