Internal knowledge
Ask the company anything. Get the source.
The answer usually exists somewhere, in a slide, a thread or a policy nobody can find. This pipeline connects your knowledge sources, keeps them indexed with their permissions, answers questions in Slack or Teams with citations, and routes what it cannot answer to the expert who can, turning that answer into documentation.
- < 10 s
- Answer with sources
- 100%
- Permission-aware
- −40%
- Repeat questions
How it runs
Step through the pipeline.
Click a step or let it play. The question record on the right fills in as the agents work.
Step 01 of 08
Connect the sources
Wiki, shared drives, help desk, CRM notes and chosen chat channels are connected, each with its existing access permissions.
- Sources
- 7 sources · 48,000 documents · permissions synced
- Question
- —
- Retrieved
- —
- Answer
- —
- Routing
- —
- Docs
- —
- Freshness
- —
- Insights
- —
01 · Connect the sources
Wiki, shared drives, help desk, CRM notes and chosen chat channels are connected, each with its existing access permissions.
Tools: Confluence / Notion, SharePoint / Drive, Slack / Teams. Output: Sources.
02 · Ask where you work
Employees ask in Slack, Teams, the browser or the intranet, in their own language.
Tools: Slack / Teams. Output: Question.
03 · Retrieve what they can see
Only documents the person is permitted to read are searched, ranked by relevance and freshness.
Tools: Velum Runtime. Output: Retrieved.
04 · Answer with citations
The answer is short and specific, with links to the exact sections it came from, so people can check it.
Tools: Velum agent. Output: Answer.
05 · Route the unknowns
When no reliable source exists the question goes to the named owner of that topic, instead of a guess.
Tools: Slack / Teams. Output: Routing.
06 · Turn answers into docs
Expert answers are drafted into documentation for the owner to approve, so the next person gets it instantly.
Tools: Confluence / Notion. Output: Docs. A person approves before this step completes.
07 · Flag stale content
Documents that contradict each other or have not been reviewed in a set period are flagged to their owners.
Tools: Velum agent. Output: Freshness.
08 · Report what people ask
A monthly view of the most asked, least answered and fastest-growing topics shows where onboarding and documentation should improve.
Tools: Sheets / BI. Output: Insights.
Impact
What changes when it runs.
Illustrative targets from comparable engagements. Your baseline is measured in the audit and the targets are agreed before we build.
Hours saved and ROI
What it is worth to your team.
Start from our defaults for this pipeline, then move the sliders to your numbers. The maths is simple and shown.
Knowledge workers commonly report several hours a week searching for or re-asking information; the assistant removes most of the searching and the repeat questions.
150 employees each losing 3.5 hours a week to searching and re-asking at a $60 loaded hourly cost, with 60 percent of it removed: about 315 hours a week back, roughly $980,000 a year.
Gross capacity, not cash savings. Excludes implementation, software and supervision costs, which are quoted before we start.
Your numbers
An assumption to explore, not a promised result.
What that is worth
315 h
reclaimed every week
$982,800
value per year
1,364 h
every month
7.9 FTE
of capacity returned
Hours on repetitive work, per week
Gross capacity estimate, not cash savings. Excludes implementation, software, and supervision costs. We validate assumptions in the audit.