Staff get instant answers with links to the actual source document, so the owner stops being the company's search engine.
This one is a guide, not a fixed install yet: the method is written and the open questions are listed below. If your situation fits, the call is where we find out.
Ops-heavy firms where staff keep asking the same questions of the same busy people: franchises, construction and logistics back offices, healthcare admin teams, professional services. Typically 20 to 200 staff with documents living in Notion, SharePoint, or shared drives, and no internal AI platform.
SOPs, policies, and how-to knowledge sit scattered across drives and chats. New hires and ops staff ping the owner or a senior manager for answers that already exist somewhere, onboarding drags, and when someone leaves, what they knew walks out the door.
01 Kick-off call
A 30-minute call to inventory your document sources and pick Slack or a web widget. You leave with a fixed quote.
02 Days 1 to 8: ingest and permission
SOPs, policies, and contracts ingested with Docling, permissions mirrored from your drive, and the surface wired up.
03 Days 9 to 15: quality check
We run 50 real questions from your team and review every answer with a human before go-live.
04 Ongoing
Weekly re-sync of changed documents and monthly tuning from the questions people actually asked.
Questions
Setup runs from $4,000 depending on how many sources and connectors we ingest. After that, from $500 per month covering the re-sync and monthly tuning. Infrastructure runs on your accounts at your cost.
About 15 working days: ingest and permission the documents, wire up Slack or the widget, then run the 50-question answer-quality check with your team before go-live.
Yes. Every answer links to the actual source document, so staff can check the context themselves and trust is not a question of faith in the bot.
Permissions mirror what people can already access. If a document is restricted to a team in your drive, answers from it are restricted the same way.
On your own infrastructure. The system is self-hosted, documents stay inside your environment, and nothing is sent to a third-party SaaS store.
Who you talk to
No account managers, no handoff to a junior team. The 30-minute call is with us, the quote comes from us, and if you go ahead, we build it.
Behind us is a wider NodeStar engineering team, including a PhD in AI, and tech partners like Kolibri Labs.

Piroune Balachandran
Founder & CPO
Co-founded Enso with Alejandro Fenn. Six years at Cornerstone Research on multibillion-dollar class action litigation and M&A.

Alejandro Fenn
Founder & COO
Three-time founder; go-to-market and partnerships. Also co-founder of Enso.

Konrad Łykowski
Head of ML & CloudOps
Designed a 60 PB AI vision system and built Australia's largest Kubernetes clusters.