Your context
is cooked.
The latest brief is in a shared document. The reasoning is buried in a chat. Someone saved a summary as a Markdown file. Bring in a teammate or open a different AI, and you’re piecing the story together again.
Project spaces in Claude and ChatGPT help within those apps. Your team’s work moves between them, and beyond them.
What’s current?
A shared file helps—until there are three versions in three conversations.
What happened when?
A memory summary can tell you where you are. Reconstructing how you got there is another job.
Why did we decide that?
The answer is in an old chat. The evidence is somewhere else. The next person has to catch up.
Different minds.
Different AIs.
One project.
One person is reading research. Another is testing a hunch. Someone else is preparing the next workshop. Work on the same project at the same time, each in your own AI client and your own direction.
Share the brief and source material. Explore separately. Bring useful findings back when they’re ready for the team.
Use compatible AI clients with your existing accounts, including personal plans.
Shared goal: a better solution
Maya + Claude
Reading research
What does the evidence say?
Alex + Leonardo.Ai & GPT
Developing ideas
What else could we try?
Jo + Gemini
Challenging the brief
Are we solving the right problem?
Lee + GLM
Planning the next step
What should we test together?
Decide together what to test next
Read widely.
Keep the thread.
A report raises a question. That question sends you to new research. A finding changes what you want to explore. Keep those connections, not just a folder of downloads.
Ask your AI to compare sources, pull out what matters to the brief, and separate evidence from things you still need to test.
“What does this report change about our thinking?”
Keep the source
The original report, with a linked summary.
Connect the finding
What it supports. What it calls into question.
Name the next question
What we still need to learn from people.
Follow a hunch.
Keep the brief.
Start with the same question and take it somewhere different. Keep a running list of hunches, sketch an idea, or challenge an assumption in your own direction. The shared brief stays intact while you work it through.
Disagree with a teammate? Keep both views and their evidence together, and mark the finding as disputed. Akagi keeps it flagged for the team until you resolve it.
The shared question
How could this service work better?
Your working hunch
What if we helped people before they needed the service?
Manage context.
Keep the history.
Know what changed, how a decision was made, and which evidence led there.
Akagi keeps the project’s current knowledge alongside its history: who changed what, why they changed it, and the sources behind it. Ask how an idea evolved, revisit an earlier version, or trace a decision back to the research.
Keep old assumptions and discarded ideas from becoming tomorrow’s facts. That’s the context poisoning Akagi helps you avoid: your AI working from stale or unproven thinking.
One person. Many conversations.
Review sources in one chat. Work through ideas in another. Return later to prepare your questions for a meeting. Akagi holds the thread even when the whole team is you.
Your files.
In the conversation.
Upload a report or a PDF of your working board. Ask your AI to read it alongside the project’s research. Keep the original, the summary and the thinking it led to together. Download the document you need for the next meeting.
“Save this report and summarise what’s relevant to our brief.”
Research report.pdf
Original file + linked research notes
“Give me our latest workshop questions to download.”
Workshop questions.md
Ready to download
Stay in the work.
You shouldn’t need a second job organising Notion pages or remembering which chat has the latest thinking. Your AI can file the useful work in Akagi, with its sources and reasons, so you can keep creating.
Bring into the project
- PDFs and documents
- Web research and links
- FigJam board exports
Work in your AI
Any compatible AI client
Akagi
Your project, kept organised
Take the thinking forward
- Ideas for your FigJam board
- Workshop notes and questions
- Files to download and share
Know where
the project stands.
Turn scattered research and conversations into a working view of the project: what’s ready, what needs checking, and what to do before the next meeting.
“Where are we up to, and what do we need for the workshop?”
- Ready to discuss
- Research summary and early ideas.
- Still to check
- Assumptions we need to test with people.
- Next
- Compare the team’s directions and agree the workshop questions.
When the questions change, keep the reason with the update. Your AI can pick it up when it checks Akagi, and you can revisit the earlier version.
A concept from
)%20%2F*%20%7B%26quot%3Bname%26quot%3B%3A%26quot%3BWarm%20Black%26quot%3B%7D%20*%2F%22%2F%3E%3C%2Fsvg%3E%0A){kind=small}
Akagi is a concept from Nakatomi, based on how we work when creating new companies: different disciplines, independent thinking and a shared understanding of the project.
Nakatomi is the company that makes companies. A team of technologists, creatives and entrepreneurs turning big ideas into businesses.
Bring a real project.
See where it goes.
We’re inviting creative teams and independent minds to help shape Akagi. Try it on something you care about, and tell us what makes the work better.
A quick chat with Andy, then a free trial. No obligation.