Fewer repeated questions
Policy, procedure and how-to questions are answered in Teams, Slack or the intranet, so HR and IT stop answering the same things every day.
Knowledge Guide AI Agent
Knowledge Guide AI Agent answers employees’ questions from your intranet, policies and documents in seconds, always with the source. HR and IT answer far fewer repeated questions, and each person only gets answers from documents they are already allowed to open.
What it means for your business
-
−40%
questions to HR and IT
- Before
- HR and IT answered the same policy and how-to questions every day.
- With Knowledge Guide
- Employees get the answer themselves, with a link to the exact page.
-
< 10 s
per answer, with its source
- Before
- Employees searched folders and intranet pages, or asked a colleague and waited.
- With Knowledge Guide
- The answer arrives in seconds, together with the document it came from.
-
2×
faster onboarding
- Before
- New joiners relied on busy colleagues for the many small questions of their first weeks.
- With Knowledge Guide
- They ask the agent and follow an onboarding path, so colleagues can keep working.
What your HR and IT teams get back
Answers people can check
Every answer links to the exact page it came from, so employees can open the source and check it themselves.
Access rules stay as they are
Each employee only gets answers from documents they are already allowed to open. Confidential files stay with the people who may see them.
Knowledge that stays current
Outdated pages and documents that contradict each other go to their owners, and repeated questions become verified answers.
What it looks like in practice
An employee asks one question in Microsoft Teams. The agent finds the current remote work policy, checks the conditions and offers to start the request. The answer shows its source, so the employee can check it.
EmployeeCan I work from abroad for two weeks in July?
Knowledge GuideYes, for up to 20 days a year inside the EU, with your manager’s approval (remote work policy, updated May 2026). Outside the EU, HR must check first. Shall I start the request?
Looked up
- Intranet
- Company policies
- Confluence
- Microsoft Teams
How six companies use the Knowledge Guide AI Agent
Each one in short: the problem, the results, and how the agent handles the work.
Past projects found in 4 minutes, not three hours
Answers were spread across file shares organised by project number, a drawing register, an internal standards library and team notes in Confluence.
- 1Searches only what the engineer may open
A graduate asks in Teams how the firm checks a bridge bearing replacement.
- 2Answers from the current standard, with sources
The answer follows section 4.2 of the current internal standard and cites it, then links two comparable past projects and their latest drawings.
- 3Flags what is out of date
It notices that a 2016 calculation template still on the file shares conflicts with the current standard.
- 4Names the right person to check the work
From the staff directory and the project archive, it finds the principal engineer who led the last three similar projects, so the graduate knows exactly who to ask for a review.
- 4 minto find a comparable past project
- 92%of answers cite the current standard
- 41outdated templates retired
- 3 weekssooner for new graduates
100% of answers from the current operations manual
The manuals live in Google Drive, station procedures in Confluence and crew bulletins in Microsoft Teams channels.
- 1Takes the question where staff already are
A station agent on a night shift asks in Teams whether a second walkaround is needed after de-icing in freezing rain.
- 2Reads only the current revision
The agent ignores superseded revisions and old copies in shared folders.
- 3Answers with the exact section and page
The reply quotes the rule in plain words and cites the manual, revision, section and page, so staff can open the source and check it in seconds.
- 4Sends conflicts to the manual owner
When a crew bulletin says something different from the manual, the agent flags it to the manual's owner with both texts side by side.
- 100%of answers from the current revision
- 6conflicts caught before the audit
- −58%questions to the operations desk
- 5 sto a cited answer
A nine-week knowledge gap closed in 3 days
Technicians asked each other in Microsoft Teams, and the answers they got depended on who replied.
- 1Spots the repeated question
The agent recognises that 37 differently worded questions are really one, and that none of them got a good answer.
- 2Drafts an answer from what already exists
It drafts six steps from the manufacturer manual and two past work orders where the isolation was done correctly, and checks the asset register for where the new units are installed.
- 3Sends it to the right owner
It finds the safety procedures owner in the staff directory and opens a task in Jira with the draft, the sources and the original questions attached.
- 4Publishes the verified answer
The owner edits one step and approves.
- 37 → 1questions became one verified answer
- 3 daysto close the gap
- 120+ times reusedThe verified answer has been used more than 120 times since it was approved.
- Every safety answer has an owner
Helpdesk tickets down 46% across 180 pharmacies
Pharmacists, technicians and counter staff had questions at the busiest moments: a cold-chain delivery arriving before opening, a damaged carton to return in the evening, a closing checklist on a new counter colleague's second day.
- 1Understands the store question
Staff ask in their own words in Microsoft Teams, at any hour the store is open.
- 2Answers with the exact step and its source
It replies with the relevant step of the current SOP, citing the document and step number so staff can open it and check.
- 3Keeps to procedures
Anything that sounds like a patient's health question is handed back to the pharmacist.
- 4Sends real issues to the area manager
When something is actually wrong, such as a fridge that won't hold its temperature, it opens a Zendesk ticket with the details attached and tells the store what happens next.
- 2in 3 questions answered outside office hours
- −46%helpdesk tickets
- 5 daysto onboard a new counter colleague
- 180stores on the same procedures
64% fewer emails to the policy team
Procedures were spread across the intranet, a policy and regulation library, department file shares and a procedures wiki.
- 1Knows who is asking
When an employee asks in Microsoft Teams, the agent knows their department and what they may open, using the administration's own staff directory.
- 2Reads only what that person may open
It searches only the documents the employee already has access to.
- 3Answers with the step and its source
Every answer cites the procedure, section and version, so the employee can open it and check.
- 4Names the right person when access ends
If the full answer sits in a folder the employee can't open, the agent shares what they may see and names the policy owner to ask, from the staff directory.
- 0answers from closed documents
- −64%emails to the policy team
- 30 sto find a procedure
- 1,200outdated pages archived
Proposals started in 20 minutes instead of two days
Before every proposal, consultants posted "has anyone done this before?" in Slack and waited.
- 1Understands the question
A consultant asks in Slack, such as "have we done a pricing project for a regional retailer before?" or "who knows about supply-chain carbon accounting?".
- 2Finds past work the consultant may open
It returns the most relevant past proposals, reports, method pages and model templates, each with its source, from folders the consultant can already open.
- 3Suggests the people, with the evidence
It names the colleagues who wrote, led or answered the most on the topic, shows why, and adds their availability from the staff directory.
- 4Respects confidential work
When a relevant project sits in a folder the consultant can't open, the agent doesn't share anything from it.
- 20 minto start a proposal
- 3×more reuse of past work
- 1message to find the expert
- 0confidential files exposed
Illustrative scenarios with unnamed companies. The figures show the kind of improvement each setup is designed to deliver; actual results depend on volume, the systems connected and how the agent is configured.
From first conversation to live in three to nine weeks
We set up the Knowledge Guide AI Agent around your documents, your access rules and the questions your employees bring to HR and IT today.
-
1
Week 1
Learn your process
We sit with the people who do the work, map the process step by step and find the systems that hold the information. You get a written plan before anything is built.
-
2
Weeks 2 to 3
Customize your agent
We build and configure the agent around your process: your rules, your approvals, your wording, and connections to your databases and the tools your teams use.
-
3
Weeks 3 to 4
Deploy and test with your team
We install it inside your infrastructure, or in our cloud if you prefer, and test it with your team on real requests until the results meet the standard you set.
-
4
From week 5
Onboard your team
We train your employees and administrators, hand over the controls and stay with you: monthly reviews, updates you approve, and the next process when you are ready.
What your team and managers see
Knowledge Guide AI Agent: questions from business leaders
How quickly will we see results?
Employees ask real questions within the first weeks. Most companies see fewer repeated questions to HR and IT, and faster answers for new joiners, during the first month after going live.
What does it cost?
It depends on the number of employees, the sources it reads and where it runs. After a first conversation you get a written proposal with the price, so you can compare it with the time your teams spend searching and answering today.
Will it replace our HR or IT staff?
No. It takes over the repeated questions, so HR, IT and your experts spend their time on the work that needs them. Decisions and sensitive topics stay with your people.
Does it make things up?
It only answers from your own documents and always shows where the answer came from. If the answer isn’t documented, it says so and sends the question to the team that owns the topic.
Can we control what it answers?
Yes. You choose its sources, decide which answers are verified and which topics always go to a person, and owners approve every new answer before it is used. You can pause it at any time.
Are our documents safe?
It can run on your own servers, together with its AI models, so your documents stay inside your company and no outside AI service is used. We have no access to them, and your data is never used to train AI models.
Will employees see documents they shouldn’t?
No. Each employee only gets answers from documents they are already allowed to open, following the access rules you already have.
Do we have to move our documents?
No. It reads your intranet, Confluence, SharePoint, file shares and policy library where they are today, and employees ask in Microsoft Teams, Slack or the intranet.
What do you need from our team?
The owners of your main sources, access to test versions of them, and a few hours a week from HR, IT and a handful of employees while we test.
Can we start small?
Yes. Most companies start with one department or one set of documents, such as HR policies, see the results, and then add more sources.
See what the Knowledge Guide AI Agent can do for your HR and IT teams.
Tell us where your knowledge lives, which questions reach HR and IT most often and how long new joiners take to settle in. You get a written proposal with the expected results, the setup, a timeline and a price.