Knowledge Assistants
AI Knowledge Agent for Microsoft 365
Answers staff questions from your documents with citations, connected natively to Microsoft 365 so it reads and cites SharePoint, OneDrive and Teams…
Knowledge Assistants
AI Document Q&A
One document set turned into a chatbot that answers questions and hands off anything it cannot answer.
Customers and staff keep asking the same questions about one manual, spec sheet or policy pack, and someone has to stop and look it up. This chatbot answers those questions from the document set you approve, quotes the passage it used, and hands anything else to a person. It runs on your website, Slack or Microsoft Teams, in your own accounts.
- Answers from one approved document set, with the passage cited
- Hands off anything outside the documents to your team
- Runs on your website, Slack or Microsoft Teams
$4,400
20–30 working days
Optional maintenance $249/mo
What it is
AI Document Q&A is a chatbot that answers questions from one defined set of documents, such as a product manual, spec sheet or policy pack, and cites the source passage behind every answer. Anything outside that set is handed to a person instead of guessed.
Why businesses buy it
Repeat questions about a single manual or reference pack quietly consume hours of skilled time each week, and rushed answers from memory cause mistakes. A cited chatbot answers instantly, day or night, and shows the reader exactly where the answer came from, so trust stays high and your team only handles the questions that need a person. It is a focused starting point for teams new to AI. The Scope Card fixes the channel, document limit, intents and testing before you order, so the build is small, clear and predictable.
Problems this solves
Customers ask questions the product manual already answers
The chatbot answers from the one document set you choose, such as a manual, spec sheet or policy pack. Every reply quotes the source passage, so customers can see where the answer came from.
A bot that guesses can damage customer trust quickly
Questions outside the agreed documents are not guessed at. They are handed to a person through one handoff route, such as your email or helpdesk, so a real answer follows.
Staff lose time paging through long PDFs for one detail
Up to 100 documents or 200 web pages are loaded behind a single chat. Placed on your website, or in Slack or Teams instead, it finds the exact passage and cites it.
A good fit if
You have one clear document set in mind, such as a single product manual, spec sheet or contract template, not a company-wide library.
You have up to 100 documents or 200 web pages, and customers or staff keep asking the same questions about them.
You want a chatbot on your website, Slack or Microsoft Teams that answers only from documents you approve, nothing improvised.
You are fine handing off anything outside that document set to a person, by email or your existing helpdesk.
Not the right fit if
Your questions span more than one department's documents, not a single set. See AI Company Knowledge Hub, built for that breadth.
Your document set is specifically HR or compliance policy. See AI Policy Q&A, which is written for exactly that content.
You need a search bar across your existing tools rather than a chatbot answering from one set. See AI Internal Search.
Your documents are not written yet and need drafting, not indexing. See AI FAQ Generator, which drafts FAQ content from source material.
How it works
- You tell us which document set the bot should learn, such as a 40-page product manual or a folder of spec sheets, up to 100 documents or 200 web pages.
- We index every document and your top 20 approved FAQs, so the bot only draws from material you have signed off, never the open internet.
- A visitor or staff member asks a question in plain language, on your website widget, Slack or Microsoft Teams, and the bot matches it to the right passage.
- The answer is drafted from that passage and shown with its source, so the reader can check it rather than take it on faith.
- Anything outside the approved document set, or a question the bot cannot answer confidently, hands off to your escalation email or helpdesk instead of guessing.
- We test the bot against 40 documented questions with pass/fail results before it goes live, and set up the handoff path you chose.
What you provide within 72 hours
Your portal lists these the moment you order. Delivery starts the working day after everything is in and your access is checked.
- The documents or website URLs this bot should learn from, as one defined set rather than everything you own.
- Your top 20 FAQs, with the answers you want the bot to give.
- An escalation email address or helpdesk access for the questions it should hand off.
- Website admin access, or your developer's time, to install the chat widget.
- Your logo and brand colors, so the widget matches your site.
- Admin access and API keys for every connected system, shared by adding our team as a user or creating a key you control, never by email.
- One decision-maker who can approve the FAQ answers before launch.
- An API key for your own AI model account (for example OpenAI or Google Gemini), created for this build with a spending limit.
The 72-hour rule. Send everything on your checklist, including access, within 72 hours of ordering so your build starts on time.
Timeline
Start
Intake confirmed, with the document set and top 20 FAQs agreed.
Weeks 1–2
Documents indexed and the FAQ set built into the bot's approved answers.
Week 3
Widget and handoff path connected to your website, Slack or Microsoft Teams.
Weeks 3–4
Tested against 40 documented questions, with pass/fail results delivered.
Week 4
Your review and two rounds of changes.
Go-live
Live, with a recorded handoff session, and 30 days of support starts.
| Option | Delivery time | Price |
|---|---|---|
| Standard | 20–30 working days | $4,400 |
| Express | 15–25 working days | $5,150 |
| Rush | 10–20 working days | $5,950 |
What you receive at handover
- The Q&A chatbot live on one channel: website widget, Slack or Microsoft Teams
- Your document set loaded, up to 100 documents or 200 web pages
- Workflow, prompt and guardrail configuration covering up to 5 intents plus human handoff
- Connection to one handoff route, such as email or your helpdesk
- Test report with 40 documented questions and pass/fail results
- Admin guide, recorded handoff session and 30 days of post-launch support
Expected benefits and how to measure them
- Can answer routine product questions at any hour, without waiting for staff.
- Helps your team spend its time on questions that need real judgment.
- Share of chats resolved without handoff
- Compare the handoffs arriving in your inbox or helpdesk with the total number of chat conversations each month.
- Document questions reaching support
- Tag manual and spec questions in your helpdesk for four weeks before launch and four weeks after, then compare counts.
- Answer accuracy
- Each month, re-run a sample of the 40 test questions and compare pass or fail with the original test report.
Before you start: Count how many manual or spec questions reach support in a typical month. Keep a list of the twenty asked most often.
These are things to track, not promised results. Outcomes depend on your data, your team and your market.
Running costs
You pay for the build once. Running the bot costs roughly $30 to $150 a month in your own accounts, covering AI model usage and hosting, and it moves with how many questions your document set answers each month.
AI usage during the build and testing is billed to your own accounts too, so the AI model relationship stays in your name from day one.
Ask about running costsCustom requirements
If you need more than 100 documents, a second channel, or more than 5 intents, add a note at checkout for a written Change Quote. Your order still runs on its normal timeline. Larger or bespoke projects go through the Custom AI request form.
Request a custom versionKeep it running
Monthly maintenance covers monitoring, fixes, updates, a health report and change hours for things such as new FAQs or an updated document set. It starts after the 30 days of support and can be cancelled any month.
$249/mo (the minimum fee, includes 2 change hours)
Optional. Starts after the post-launch support period. Cancel any month.
How maintenance worksHow this compares
| Do it yourself with a SaaS tool | Hire a custom agency | Upmark Labs | |
|---|---|---|---|
| Price | A chatbot builder charges a monthly fee whether anyone asks it a question or not | A custom agency quotes after a discovery call | Ours is one fixed price, shown before you buy |
| Scope | A chatbot builder leaves you to define intents and upload content yourself | An agency agrees scope as the project runs | Ours is written on the Scope Card before purchase: one set, up to 100 documents |
| Delivery | A chatbot builder is instant to sign up but slow to tune well | An agency build usually takes several weeks | Ours has a fixed delivery time shown before you buy, with faster options |
| Handoff | A chatbot builder often leaves unanswered questions to disappear silently | An agency build depends on what was scoped | Ours routes anything outside scope to your email or helpdesk |
Questions about this product
How is AI Document Q&A different from AI Company Knowledge Hub?
AI Company Knowledge Hub spans multiple departments and up to 2,000 documents with permission groups. This bot answers from one defined document set, up to 100 documents or 200 web pages, on one channel. Choose Document Q&A when your questions come from a single manual, spec sheet or reference set rather than company-wide sources.
Can AI Document Q&A learn from a website instead of files?
Yes. It can index up to 200 web pages instead of, or alongside, uploaded documents, so a support site or product catalog works as the knowledge source. You tell us which pages during intake, and we index them the same way we index a PDF, with citations back to the page.
What happens when a visitor asks something outside the document set?
It says so, rather than guessing, and hands off to the escalation email or helpdesk you chose during intake. It never improvises an answer from outside the approved document set, because the model is only allowed to draw on material you signed off during the 40-question test.
How much does AI Document Q&A cost to run each month?
Roughly $30 to $150 a month in your own accounts, covering AI model usage and hosting, and it moves with how many people ask questions. The build itself is paid once and covers indexing, the FAQ set, the widget and the 40-question test before launch.
Can we swap the document set for a different one later?
Yes. Replacing documents or updating the FAQ set fits inside the monthly change hours that come with maintenance. A larger change, such as raising the 100-document limit, is quoted as a Change Quote before the work starts. Maintenance also covers monitoring, fixes, updates and a monthly health report. It starts after post-launch support and can be cancelled any month.
How long until AI Document Q&A is live?
Delivery runs on a fixed schedule from a complete intake, and the delivery panel shows that date plus express and rush options before you buy. Most of the time goes into indexing your document set, building the approved FAQ answers, and running the 40-question test before your review.
Does AI Document Q&A work inside Slack or Microsoft Teams, not just a website?
Yes. You choose one channel during intake: a website widget, Slack or Microsoft Teams. All three use the same indexed document set and the same 40-question test; only the interface changes, so pick whichever one your users already work in.
Stop answering the same document question twice.
Order AI Document Q&A and it goes live on a fixed delivery date, answering only from the document set you approve and handing the rest to your team.
Works well with
Knowledge Assistants
AI Knowledge Agent for Google Workspace
Answers staff questions from your documents with citations, connected natively to Google Workspace so it reads and cites Drive files, shared drives…
Knowledge Assistants
AI Knowledge Agent for Confluence
Answers staff questions from your documents with citations, connected natively to Confluence so it reads and cites spaces and pages your users…