Compile
Reads selected sources and updates related knowledge, with links back to the source.
Inside the BA Brain
Your context.
Connected.
See what to keep in your Brain, how it works with your AI assistant and what to try in your own analytical work.
Connected contextHuman-led
A workspace behind the scenes of your analysis.
Start with the work
What would you
use it for?
A Brain is a maintained workspace of files and instructions. Your AI assistant reads and updates it when you ask. These illustrative scenarios show how to try the approach; they are not automated integrations or measured customer results.
Before a meeting: what still needs an answer?
Example: Sales requested a faster onboarding flow; Compliance raised an unresolved identity-check question. You are preparing their next discussion.
Bring: one preserved meeting note or email thread, with its date and source.
Ask your assistant: “Using this context, which questions are still open? What should I clarify with each stakeholder? Show the sources and separate recorded facts from your suggestions.”
Work towards: a meeting brief with traceable questions. Check it against the original note before using it.
When a requirement changes: what should I revisit?
Example: A stakeholder now wants guest checkout. Earlier notes assumed every buyer would create an account.
Bring: the new request, an earlier note and the relevant project draft.
Ask your assistant: “Compare this request with the recorded context. Where do they differ? Which assumptions, people or project sections should I review? Mark uncertainty and cite the source.”
Work towards: a list of differences and follow-up questions. You validate the impact and approve any project update; the Brain does not infer a complete dependency map.
Returning to a project: where did we leave off?
Example: You return to a supplier-onboarding project after working elsewhere. You need to recover the agreed scope and unanswered questions before the next call.
Bring: the project brief and selected stakeholder and communication records already in your Brain.
Ask your assistant: “Help me prepare to resume this project. Summarize the scope, relevant people and unresolved points. Distinguish the latest recorded position from information that may need reconfirmation.”
Work towards: a project refresher you can verify, with questions to take back to the team.
How is a Brain different from an agent?
Agent · carries out a task
Your AI tool reads information, reasons about a question and proposes or makes permitted changes. It needs access to the relevant files.
Brain · organizes the context
Your profiles, source material, linked records, working instructions and project drafts remain in a workspace you can inspect and reuse across tasks.
They work together. An agent may already have memory. BA Brain adds an explicit, portable structure for BA work; you maintain it through your assistant and review what changes.
Why build my own — and when is it worth the effort?
The value is in reusing reviewed context: your way of working, the evidence behind a finding and the history of a project. You can inspect the files and choose a compatible AI environment. Your team tools can stay the source of truth for shared work.
It is worth trying if you repeatedly reconstruct the same context and are willing to set up and maintain a file-based workspace. A one-off question may need only a chat and a document. A ready-made team platform with live integrations is a different need.
Local files do not mean local AI processing: what your assistant sends to its provider depends on that tool and its settings. Use only information you are permitted to process.
01 / Your profiles
Your way of working.
The starting point.
Start with two profiles. Build them with AI and decide what you want to include.
Personal profile
Who I am
- My background, vision and goals
- My values, strengths and priorities
- How I want to work with AI
- My communication and writing preferences
Analyst profile
How I approach analysis
- My areas of expertise and preferred techniques
- How I collaborate with other roles
- My expectations for evidence and review
- My tools and working rhythm
Both profiles are reviewed and maintained by you; AI can help draft updates. Start with what is useful; you can develop them over time.
02 / Inside the Brain
Inside the
public starter.
Select a layer to explore starter 0.1.0. Findings, decisions, risks and open questions stay inside communication records; separate registers are not included.
Layer 0 / Start here
How the Brain works.
The navigation, operating rules and local settings that help you and your AI assistant work consistently.
- 0_home/
- 0_system/
03 / How the Brain works
Compile, check
and keep track.
A few mechanisms help keep your knowledge connected and your sources available.
A note or emailCompileConnected knowledge
Check
Checks the structure and links. You still review whether the content makes sense.
Indexes & logs
Help AI find information and keep track of what changed.
Preserved sources
Keep the original input so you can return to it.
04 / Working on a project
Create a project.
Develop it with AI.
Two skills support your project work. You review the drafts and approve changes.
Project create
Set up the starting structure
Creates the project with seven draft artefacts and a shared project guide.
Project update
Keep the analysis context up to date
Proposes changes to project documents for you to review and approve.
The seven project artefacts
01
Project Brief
Why does this project exist, and what is its scope?
02
Current State
What do we know about the situation today?
03
Future State
What change are we trying to achieve?
04
Business Analysis Approach
How will we organize and carry out the analysis?
Inside this approach
Project context, formality, adaptive or predictive planning, activities, techniques, deliverables, timing, dependencies and review points.
05
Stakeholder Engagement Approach
Who needs to participate, and how will we work together?
Inside this approach
Stakeholder interests, impact, influence and authority; participation, collaboration, communication and engagement review. Stakeholder analysis informs this approach.
06
Governance Approach
Who decides, approves and handles changes?
Inside this approach
Decision responsibilities, escalation, change control, prioritization and approval. Review and approval are distinct activities.
07
Information Management Approach
How will we organize, connect, share and maintain information?
Inside this approach
Information categories, ownership, detail, attributes, traceability, reuse, storage, access, confidentiality, maintenance and handover.
These documents provide context for the analytical process. AI helps you draft; you own the analysis and agreements.
05 / What is available and what comes next
What is available.
What comes next.
Start with your profiles and test the simple structure. We’ll build on what we learn together.
A simple foundation.
- Folder structure and operating rules
- Blank personal and analyst profiles
- Four skills for working with the Brain
- Seven project patterns
A starting point to try and adapt.
Sessions & your notes.
- Capturing working sessions
- Compiling session context
- Supporting your own notes
These workflows need more refinement before I share them.
More connections.
- Jira and Confluence
- Other documentation sources
- Further skills and capabilities
Direction, not a promised release schedule.
A note on imports. Email and meeting-note imports are part of my private workflow. Ready-made importers, connected accounts and schedules are outside this first starter package.
06 / Your first steps
How to start
your own Brain.
This is your personal workspace behind the scenes of your analysis. You choose what becomes a reviewed output for your team.
Try one fictional meeting note first. A useful first result is a traceable record you can inspect, plus a project draft you can review. You do not need to connect your accounts to try this.
What you need to get started
An AI assistant with access to your workspace files, Python 3.10+ and PyYAML for the supporting scripts. Tool connections and model setup are separate. Follow the starter’s installation and first-use guides.
Get your own working copy
Get the starter on GitHub ↗ and follow its installation guide. Create your private working Brain outside the public repository.
Build your personal profile
Add the context, goals and preferences you want AI to know. You choose what is useful to include.
Build your analyst profile
Describe your analytical practice, techniques and ways of collaborating. Let AI help draft; review it yourself.
Add one small source
Try a fictional or appropriately shareable note or email text. Follow the first-use guide; connectors and importers need separate setup.
Compile. Inspect. Check.
Follow the source references. Look at what is known and what remains unresolved. Check meaning as well as structure.
Create your first project
Explore the seven drafts. Adapt them to your situation and review proposed changes.
Make it yours
Try a small skill, or explore how skills and agents you already use could work within the Brain’s structure and rules.
Explore the ideas behind the Brain
01 / The reason behind it
“I wanted AI to work with the way I think — and the context I build every day.”
For many business analysts, the day unfolds between Jira, Confluence, meetings and emails, leaving little time for the analytical work itself. I started looking for a way to keep track of that information, ask questions across it and make more room for synthesis and deep work.
Could AI help me stay on top of what was happening while giving me more time to think? Even an AI agent supporting BA work needs the context gathered from people, conversations and existing information. That is where I started building my Brain.
Working with AI
Start with the context
your task needs.
A useful starting point is a recurring problem: preparing for a meeting, understanding a change or returning to a project. Choose one source, inspect what AI makes of it and decide what is worth keeping.
More automation is not always the next step. Sometimes the missing piece is a source, a clearer question or a conversation with a colleague.
Read the foundations: A Second Brain for Business Analysts (PDF) →
03 / Different shapes of a Brain
What kind of Brain
are we building?
Brains can take different forms. BA Brain starts with an interlinked knowledge wiki.
A collection
of files
A straightforward place to keep your context.
Needs: organization, retrieval and freshness.
An interlinked
knowledge wiki
Related records, source references and maintained context.
Needs: evidence, maintenance and review.
A managed
enterprise platform
Search, knowledge workflows and productized integrations.
Needs: configuration, access controls and fit.
A custom
agent environment
Your own combination of tools, memory and automation.
Needs: engineering, evaluation and operation.
Examples and further reading
Karpathy’s LLM Wiki separates sources from an AI-maintained wiki and describes ingesting, querying and maintaining knowledge.
Enterprise examples include Glean’s knowledge graph and Guru’s expert verification workflows. These illustrate different capabilities, rather than identical versions of this Brain.
04 / Putting the ideas together
LLM Wiki + OKF
+ business analysis.
Knowledge about building Brains and knowledge about business analysis — brought together in one workspace.
LLM Wiki
Connected, maintained knowledge+
Open Knowledge Format
A Markdown-based knowledge format+
Business analysis
Evidence, context and human decisionsThe BA Brain
A workspace for analytical work with AI.
An independent application of these ideas, not an official Google or Karpathy product.
How BA Brain uses Open Knowledge Format
Open Knowledge Format is published by Google Cloud. BA Brain defines a local profile of OKF v0.2 for its wiki and analytical pages. Raw sources, personal profiles and other workspace files are not all part of that profile. A valid format does not establish factual accuracy.
Your next step
Start small.
Make it yours.
Get the free starter, build your profiles and try one small source. Let your Brain grow with your practice.