Falconer is a system of record for company knowledge. I co-founded it with Maxi Benedetto in January 2025.
What we do
We build software to keep company knowledge current as code, decisions, and the business change.
- Create documentation from code, conversations, and existing knowledge.
- Maintain and organize it as the underlying sources change.
- Give people and agents answers grounded in the company’s context.
Founders
Maxi and I worked together at Uber. I later led docs at Stripe; Maxi worked on AI at Meta. We started Falconer after seeing the same problems with internal knowledge across the companies we worked with.
Stage
Seed. We launched publicly in February 2026.
Investors
The General Partnership, BoxGroup, SV Angel, Coughdrop Capital, and Xtripe Angels.
Our angel investors include Ben Lang and Gokul Rajaram, alongside operators from Ramp, Lovable, and Graphite.
If you’re an investor
We get a lot of requests for meetings.
A warm intro obviously helps, but responding to our beliefs with your agreement or disagreement is also a great way to get a meeting.
Above all else, we value knowledge and conviction about our space. Full agreement on our approach will come with time:
- Enterprises demand knowledge and intelligence sovereignty.
- We’re the only true system of record for knowledge: centralized, structured, correct.
- We think documents are the king of context for agents.
- Deeptech is the most important customer opportunity.
- We’re going to kill Atlassian.
If you can share a few bullets on where you agree or disagree with our beliefs and approach, I’ll have a better idea if there’s a fit.
Seed memo
Falconer is building a knowledge platform for technical teams to write, find, and share knowledge 10x faster.
Engineers are the most critical, expensive resources for R&D-focused companies. But they have the worst tools to create and share knowledge.
If they forced us to go back to Confluence again, we would riot.
— From Falconer UXR interview
Problem
Knowledge and docs are out of date, hard to find, and time-consuming to create.
Engineers spend 1–2 days a week on writing tasks. While clear communication is becoming increasingly critical, companies are facing a massive knowledge problem:
- Engineering organizations lose millions in lost productivity due to poor knowledge management.
- R&D-led companies are heavily disadvantaged when engineers can’t clearly communicate ideas and knowledge.
- ChatGPT and Claude are generic writing tools without tightly integrated access to internal data.
- Fast-growing companies are always changing, making knowledge management a neglected moving target.
- Language barriers and timezones introduce even more friction for large organizations.
The only thing my engineers hate more than reading docs is writing docs.
— From Falconer UXR interview
Tools like Confluence, SharePoint, Notion, and Glean contribute to the problem of garbage in/garbage out: incomplete, inaccurate, outdated, and poorly constructed knowledge.
Opportunity
Falconer will automatically create, update, and organize institutional knowledge.
Falconer's unique approach:
- Falconer will focus on AI-powered writing to become an indispensable tool for engineers.
- Fully tested, accurate information is a core principle of Falconer's approach. This includes written content that is factually and grammatically linted, and code samples and diagrams that are automatically tested against the code base.
- After winning over engineers, Falconer will take on knowledge management and search across the entire company.
- Falconer will provide knowledge agents to replace outdated wikis and painful peer-to-peer Slack support.
- By owning the knowledge layer, Falconer will build the most intelligent coding assistant for organizations.
How it works
Here are some of Falconer's key capabilities:
Knowledge editor
- A Cursor-for-docs experience, where engineers have a writing copilot within their IDE.
- A writing canvas UI, with editable draft, LLM output, and prompt bar working in harmony for maximum flow state.
- Code samples, complex diagrams, and data visualizations all created within the same tool.
- AI-powered review for instant high-quality feedback.
- Ability to learn individual and organizational styles and preferences.
Self-updating knowledge
- Keep knowledge up-to-date. Automatically keep content, diagrams, and code samples updated with the codebase.
- Automatically identify and close knowledge gaps. For example, auto-document what engineers are asking about in Slack.
- Metrics and insights for your knowledge and productivity.
Search and discovery
- “Ask anything” experience to search content and code.
- See what’s new and relevant to you.
Knowledge agents
- First line of defense for internal questions, minimizing the need for costly and painful peer-to-peer Slack support.
- Agents for internal support and education. For example, guided agents will provide custom onboarding checklists and conduct check-ins to review milestones and goals.
Privacy and security
- Enterprise-grade privacy and security, allowing any organization to put even their most sensitive and proprietary knowledge to work.
Alternatives
Knowledge management
- Confluence, SharePoint: Enterprise grade, but focused on project management and widely despised by engineers.
- Notion: Catered toward startups and focused on project management. Poor experience for leveraging AI to write specialized content.
- Mintlify: Focuses on tools to manage external content, not internal knowledge, and doesn’t help engineers write better.
At Microsoft, even the SharePoint engineers didn't use SharePoint.
– From Falconer UXR interview
Search
- Glean: Helps find what exists already. But it doesn't improve the content itself.
AI tools
- GitHub Copilot, Cursor: Don’t address knowledge management.
- ChatGPT/Claude: General purpose, poor UX for structured writing.
Why us
- We have extremely high conviction and strong evidence that perfecting the writing layer is the key—and no one else is solving this.
- We’ve turned engineering knowledge systems from glaring weaknesses into strengths at Uber and Stripe.
- We’ve solved this problem from every angle, building homegrown knowledge systems, content authoring languages, search tools, diagramming solutions, automated testing tools, and entire writing cultures from scratch.
A system like this is really hard to build. It’s a living read/write product requiring our unique expertise in AI, developer productivity, and knowledge management.
The winning product needs to be built with AI from the ground up; not just bolted on.
This is rad. Sign us uuuuupppppp...No really though GIVE ME EARLY ACCESS CMON.
— From Falconer UXR interview