What is a company brain?

A new hire asks in #eng why the team chose Stripe over Braintree for EU billing. Search returns a 2022 architecture doc, a Notion page someone abandoned halfway through, and forty Slack threads that mention "Stripe." The eng lead who made the call is in a customer call. By afternoon, the question has been asked twice more.

That is the gap a company brain is meant to close. Not by indexing everything your company ever wrote, but by saving the judgment someone confirmed was worth keeping, with a link back to where the decision happened.

The problem: decisions move faster than docs

Every B2B company has a version of this failure mode. A thread in #finance resolves a vendor question. A ticket captures the workaround but not the constraint. Confluence has an ADR template with three entries from 2021.

Then someone new asks the same question. A senior gets pinged. Again.

That cost shows up as the interruption tax: hours spent re-explaining past calls instead of doing new work. Search was supposed to fix it. Wikis were supposed to fix it before that. Neither worked for tacit knowledge because the "why" often never gets written down at all.

Why your wiki doesn't stop Slack pings walks through why documentation mandates fail. Glean searches memory. Ignius creates it. explains why enterprise search and memory creation solve different problems.

What a company brain actually is

Not a chatbot with your Slack export. Not a wiki with AI autocomplete. Not employee surveillance.

A company brain is structured memory built from decisions your team confirms:

• Vendor choices, escalation thresholds, architecture boundaries
• Answers someone is tired of repeating
• The relevant part of a ticket or PR, not the whole repo
• How handoffs actually work, not how the handbook says they should

Each decision has a summary you can scan in results, a source link back to Slack, Jira, GitHub, or wherever it came from, a confidence level, and a folder in your Company Brain it belongs to. Ask queries those decisions and returns citations, not a paraphrase over every message you ever sent.

For storage policy and privacy boundaries, see What a company brain captures (and what it doesn't).

How Ignius builds one

The capture loop is intentionally short:

1. Connect Office, GitHub, and Xero. Integrations identify decisions as they are made and surface candidates: threads that look like calls, repeats, or escalation paths.
2. You confirm what becomes a decision. Nothing saves silently.
3. Decisions land in folders in your Company Brain (Operator, Portfolios, Partner, Employee), not one giant pile.
4. Ask reads saved decisions inside the groups you pick and answers with sources.

Connect integrations are read-only signals. They feed a candidate list. Permanent memory requires a human in the loop. Control access and permission for everyone. Without that trust model, nobody saves anything useful.

How context groups keep answers scoped goes deeper on why boundaries matter for retrieval quality.

Folders in your Company Brain

A company brain without boundaries is just another noisy search index. Ignius uses context groups, folders in your Company Brain, so retrieval stays scoped.

A finance question should not pull unrelated eng debates. Operator should not share a retrieval space with Partner. Most teams already think in these buckets. They just never had software that respected the walls.

When someone joins a team domain, they choose what to publish upward. Private folders stay private until they deliberately share them.

Company brain vs wiki vs enterprise search

Wikis want documentation after the decision, when context is fading. They work for polished reference pages. They fail when the answer never became a page.

Enterprise search indexes what exists in connected apps and retrieves relevant chunks. It works when the answer was written somewhere searchable. It fails on undocumented judgment buried in Slack threads.

A company brain captures what someone confirmed was worth keeping, at decision time, with a source link and confidence score. It fails if capture feels like surveillance and nobody saves anything.

For a direct comparison with Glean, see Ignius vs Glean.

Who this is for

Ignius tends to matter when people still Slack seniors after search fails on "why" questions, when key operator knowledge walks out the door with a departure, or when cross-functional teams need the decision rather than another doc link.

If the problem is finding the PRD from 2022, keep the search stack. If the problem is that nobody wrote down why the PRD says what it says, the team is looking at memory creation.

Deal teams running diligence and integration have a parallel version of this problem. See Company brain for M&A for what to capture before, during, and after a transaction.

Try the Horizon Logistics demo to see capture queue, context groups, and Ask with sources on sample data. For the team product layer, see Company Brain for teams.