Your documentation,
finally answering questions by itself
Most knowledge bases are written once and then searched badly, because a search box returns ten articles and asks the reader to find the answer themselves. This agent reads the actual documents and gives a direct answer with the source cited, so the knowledge base finally earns its keep.
The role today
Writing good documentation and having it actually get read are two different problems, and most businesses solve the first without solving the second. A customer with a question searches the help center, gets a list of ten articles, and often gives up and messages support anyway, asking the exact question the documentation already answered three paragraphs into article seven. The knowledge base becomes something written once for compliance’s sake rather than a tool customers use.
What the agent takes over
The agent reads your actual documentation, policies, FAQs, product guides, and answers questions directly rather than returning a search results list, citing the specific document and section the answer came from so a customer, or your own team checking its work, can verify it instantly. It runs embedded on your site, inside Telegram and WhatsApp, and in an internal version for staff questions about policy or process that would otherwise interrupt a manager.
When a question has no good answer in the existing documents, instead of guessing, it says so and logs the gap. That gap report becomes a genuinely useful to-do list: it shows your team exactly which questions keep coming up with no documented answer, which is a better way to prioritize what to write next than guessing at what might be missing.
The internal staff version runs under the same citation discipline as the customer-facing one, which matters because an employee asking whether a discount is allowed needs an answer traceable to an actual policy document, not a confident-sounding guess that later turns out to be outdated. When a document is updated, the agent’s answers update immediately rather than lagging behind a cached version, one of the more common failure points in static FAQ pages that quietly go stale for months. Usage data also shows which documents get cited constantly and which never get touched, a more honest signal of what customers actually read than any page-view count on the help center itself.
What stays with humans
Writing and maintaining the actual documentation stays with your team; the agent answers from what exists, it does not invent policy. Reviewing the weekly gap report and deciding what to document next is a team decision.
Guards
Every answer cites its source document, so nothing is asserted without a traceable basis. The agent is built to decline rather than guess when the documents do not cover a question, with every such gap logged rather than silently papered over. New or updated documents are tested with a batch of real past questions before going live, to confirm the agent actually finds and uses them correctly. Outdated answers are caught two ways: automatically, when a cited document is updated or removed, and manually, when a customer or staff member flags an answer as wrong, both feeding into the same correction queue your team reviews. For businesses with seasonal policies, holiday shipping cutoffs, limited-time terms, documents can be scheduled to activate and expire on specific dates so the agent never cites an expired policy as current.
Price and timeline
| Option | Price | Timeline |
|---|---|---|
| Agency runs it | from $2,000 + support plan | 2 to 4 weeks |
| Full control, handover-ready | from $3,400 | 4 to 6 weeks |
“Agency runs it” keeps the agent on our infrastructure with a monthly support plan covering tuning and monitoring. “Full control, handover-ready” delivers the agent on your own servers and accounts with documentation, source and credentials, so your team can run and change it without us; it costs more up front because the handover package, your-infra deployment and internal documentation are built in from day one.
Related
See the AI agents overview for how we build and guard these systems, and Routine takeover for the services that often pair with this one. Inside the agents catalogue: Customer support agent for messengers, Self-service portal agent, Technical support tier-1 agent. For a narrower, single-process version of this work, see FAQ answering, RAG knowledge base with citations in the automations catalogue. For how the same guard patterns held up on real systems: Analytics hub with AI analyst in Telegram.
Want this running for your team? Get in touch and tell us where the messages pile up.
FAQ
How much does a FAQ and knowledge base agent cost?
From $2,000 to build the knowledge base from your existing documents and connect it to your site and messengers, live in 2 to 4 weeks.
How long before it is live?
2 to 4 weeks, depending mostly on how much of your documentation already exists versus needs writing first.
Which channels and tools does it work with?
Your website as a chat widget, Telegram, WhatsApp, and an internal version in Slack or Telegram for staff; it reads from your actual help center, PDFs, policy documents or wiki, whatever already holds the answers.
What if the agent cannot find a real answer?
It is built to say it does not know rather than guess when the documents do not cover a question, and every such gap is logged in a weekly report so your team can see exactly what to write next.
Is our documentation and customer data safe?
Documents stay in your own storage or help center; the agent reads them to answer, does not alter the source, and every question and the source it cited are logged.