Introducing Honen Knowledge Bases & Parallel Training

Today we're launching Honen Knowledge Bases, one place to store everything your company knows, and a way to teach it to both your people and the AI they work with. Drag in your documents, connect your internal sources, or write new ones with AI. Honen turns it into a single source of truth, then trains your team and your AI agents from it at the same time. We call that Parallel Training.

Why we built it

Every company now has two workforces. The people on your team, and the AI agents working beside them: the Slack bots answering questions, the assistants drafting work, the tools everyone opens in Claude and ChatGPT all day.

Here's the problem almost nobody is solving. Those two workforces are rarely trained on the same thing. Your people learn a new process from an onboarding deck. Your AI tools learn it from whatever happened to land in their context window. The result is a quiet, constant knowledge gap, where a human and an agent look at the same task and operate from two different versions of the truth.

As companies hand more real work to AI, that gap stops being an inconvenience and starts being a liability. We built Honen Knowledge Bases to close it.

How it works

Centralize what you know. Link external sources, drag in documents, or create brand-new ones with AI. SOPs, playbooks, product docs, design processes, new protocols, all of it lives in one place.

Let Honen transform it. Everything you add is automatically turned into a clean, website-style document that reads beautifully to people and is fully usable by AI. It's structured, on-brand, and portable across every system you connect.

Distribute it everywhere. From that one source, the same knowledge powers your assistants, your external LLMs, your internal systems, and your courses. Update it once and every surface stays current.

Train both sides in parallel. A single piece of knowledge is delivered to your people as Courses and to your AI as Skills, at the same time, so everyone and everything learns it together.

What teams use it for

Here's how knowledge in Honen gets put to work.

• •

  Courses in seconds. Spin up full courses off your knowledge base content, so the same source that answers questions also trains your people.

• •

  Answers in Slack and iMessage. Your team can text a question or ask in a channel and get a reply grounded in your real, current playbook, not a stale wiki page.

• •

  Your tools in Claude and ChatGPT. Connect the base over MCP so the AI tools your team already uses can read and write against your company knowledge directly.

• •

  Scoped support assistants. Power external systems with API connectors and scoped MCP servers, like a support bot that only sees the support knowledge it should.

• •

  Skills for your agents. Route that same knowledge into your AI agents as Skills, so they execute your processes the way your team would.

Parallel Training

The idea we're most excited about is Parallel Training: teaching your human workforce and your AI agents the exact same organizational knowledge, at the same time.

Picture a single source of centralized knowledge, a company workflow, a design process, a new protocol. From the Knowledge Base, that one source is split and delivered to two audiences in parallel. To your people, it becomes a course they can actually learn and understand. To your AI, it becomes Skills and MCPs that route directly into your agents and systems.

So when you roll out something new, say, a new way to track open issues in Linear, your team and your agents are trained to understand and execute it together. Not the humans first and the bots whenever someone gets around to it. Together, from the same source.

Get started

Honen Knowledge Bases work for teams of 10 or 10,000. Centralize what your company knows, connect it to the tools you already use, and start training both your people and your AI from one source.

Please see honen.com to get started or reach out for a demo.