We built Knowharbor because we lost too many hours to tribal knowledge
After a decade building operations software for enterprise teams in Atlanta, I kept watching the same pattern: a new engineer would join, get blocked on something routine — how to request staging access, how to reset a user's SSO credentials, how to trigger a production deploy — and spend half a day pinging five people before someone pointed them to the right Confluence page that had always been there.
The documentation existed. The problem was retrieval — and the fact that generic Q&A tools would answer with something plausible but untraceable. A wrong answer with no source citation is worse than no answer.
We started Knowharbor in Atlanta in 2023 to fix this specifically: not just a better search bar, but sourced answers. Every answer comes with a clickable link to the document it came from. That traceability is the whole idea — it's what makes an AI answer trustworthy enough to act on.
— Darnell Hayes, CEO & Co-Founder
"If someone has to ping a colleague for information that's written in a Confluence page, that's a retrieval problem — not a documentation problem. We built the retrieval layer. The answer is already there."
Darnell Hayes, CEO & Co-Founder
Make institutional knowledge queryable — not just stored
Every organization accumulates knowledge: in Confluence spaces, in closed Jira tickets, in Google Drive folders with names like "Final-v3-FINAL." The problem isn't that the knowledge doesn't exist — it's that it's never been retrieval-indexed in a way that produces a trustworthy, source-linked answer. Knowharbor does that. We don't replace your wiki; we make it answerable.
Source citation is non-negotiable
Every answer carries a link to the document, ticket, or page it came from. We don't generate answers we can't trace. That's the line we don't cross.
The question is the interface
Asking a question in natural language shouldn't require IT expertise. The new hire on day 1 and the senior ops lead on year 3 both deserve a direct answer.
ACL inheritance, not an afterthought
Source-system access controls are inherited at index time and enforced at query time. People see only what they're already permitted to see — no configuration required.
Four people, one problem
Based in Atlanta, GA. We do not do general-purpose enterprise search. We do sourced Q&A from your company's own docs, tickets, and wikis — and nothing else.
Darnell Hayes
CEO & Co-Founder
Built operations and workflow software for enterprise teams across financial services and logistics for over a decade. Started Knowharbor in 2023 after one too many "just ask Marcus, he'll know" moments.
Priya Nair
Head of Engineering
Specializes in distributed search infrastructure and incremental indexing pipelines. Previously built connectors and ingestion layers for enterprise document management systems at growing SaaS companies.
Marcus Chen
Head of Customer Success
Spent years guiding IT and knowledge management teams at growing companies through tool rollouts. Has seen every way an internal wiki can fail and built the customer success process around preventing them.
Mei Zhang
Lead ML Engineer
Owns the RAG pipeline, embedding model selection, and the citation engine that traces every answer to a source document or ticket. Has a strong opinion about hallucination as an architecture problem, not just a model problem.
Part of the growing Atlanta tech ecosystem
Atlanta has a strong and growing enterprise software community, and a high concentration of the financial services, healthcare, and logistics companies we serve. We're here by choice — it's where the customer conversations happen in the hallway, not just on a calendar invite.
Knowharbor, Inc.1180 Peachtree Street NE, Suite 1700
Atlanta, GA 30309
[email protected]
+1 (404) 521-7762