• 

  An AI platform for long-term wealth planning that led to a cognitive architecture for memory, temporal reasoning, reflection, and strategy.

FamilyBrain

Inspiration

FamilyBrain was created to help families make better long-term wealth planning decisions by integrating legal, tax, accounting, financial, corporate, international, and estate planning into a single AI platform.

While redesigning the user experience into a single conversational interface, I discovered something unexpected.

The biggest challenge was no longer building specialist AI agents.

The biggest challenge was preserving continuity across long-term conversations.

The Discovery

As conversations became longer, previous decisions, objectives, and context gradually became disconnected.

I realized that memory alone was not enough.

An AI also needs to understand when information happened and where it belongs in the timeline of a relationship.

That insight transformed what initially seemed to be a user experience problem into a cognitive architecture challenge.

How It Was Built

FamilyBrain combines multiple specialist domains into a single conversational experience, including:

• Legal planning
• Tax planning
• Accounting
• Financial planning
• Corporate structures
• International assets
• Estate and succession planning

During its development, I began documenting a reusable cognitive architecture called CogniLab.

Rather than creating more specialist agents, the architecture focuses on four complementary capabilities:

• Memory
• Temporal Reasoning
• Reflection
• Strategy

Challenges

The greatest challenge was not retrieving information.

It was maintaining continuity over time.

Traditional AI systems remember previous messages, but they often lose the chronology of decisions, changing objectives, and the temporal context that gives meaning to long-term collaboration.

What I Learned

Building FamilyBrain taught me that the next generation of AI should not only remember more information.

It should understand time.

I believe AI systems that preserve continuity, reason across timelines, and reflect before acting will be able to collaborate with people over months, years, and eventually generations.

FamilyBrain is the practical application of these ideas.

CogniLab is the cognitive architecture emerging from that journey.## Inspiration

What it does

How we built it

Challenges we ran into

Accomplishments that we're proud of

What we learned

What's next for FamilyBrain

Built With

• anthropic
• api
• codex
• css
• docker
• fastapi
• github
• gpt-5
• next.js
• openai
• postgresql
• python
• sqlalchemy
• tailwind
• typescript
• vercel