Knowledge Assistants
AI Company Memory
Captures decisions, context and lessons from meetings and threads into searchable memory.
Knowledge Assistants
AI Company Knowledge Hub for Slack
Ask in Slack and get an answer cited to the document or approved channel thread it came from.
Your team already asks its questions in Slack, and most answers were given once, in a thread nobody can find again. This hub lives inside Slack, answers from the channels you approve plus one more system, and cites the message or file behind each reply. Private and social channels stay out, and your workspace admin controls exactly what it can read.
- Old answers resurface from approved channel threads, with links
- Questions answered where your team already works
- Private and social channels never feed the hub
$11,400
20–30 working days
Optional maintenance $570/mo
What it is
AI Company Knowledge Hub for Slack is a cited Q&A assistant that lives in Slack, reading and updating messages in approved channels, user details and app interactions, and answering from those threads plus one more connected system. Each answer cites its message or file. Private and social channels stay out.
Why businesses buy it
In a busy Slack workspace, the same question gets asked every week and answered by whoever happens to be online, often differently. Experts lose focus, and good answers disappear into scrollback. This hub turns approved channels and one connected system into a cited answer source inside Slack, so people get a consistent reply in seconds and experts are left to do their real work. The Scope Card fixes the channels, the extra system, permission groups and testing before you order, so your admin approves a known, limited setup.
Problems this solves
Staff ask the same questions in Slack channels every week
The assistant answers inside Slack, drawing on your approved channels and up to one more connected system. Each reply cites the message or file it came from, so people can check the source before acting.
Useful answers get buried in old Slack threads nobody can find
It reads threads in the channels you approve and uses them when someone asks a related question. Answers point back to the original message, so something explained once stays useful later.
Private conversations must never surface in answers for everyone
Private and social channels stay out of its reach. Up to three permission groups control who sees what. The documented field mapping records exactly which Slack data it reads and updates.
Answers drift out of date as documents and policies change
A monthly content refresh brings in new and changed material from the connected sources. Before launch, answers are checked against a set of benchmark questions, and you receive the accuracy report.
A good fit if
Answers to recurring questions already exist somewhere in Slack, in an old thread nobody can find again.
Your team asks questions in Slack first, so an assistant there gets used where a separate portal would not.
Only some channels should feed the hub, such as support, operations or announcements, while private and social channels stay out.
A Slack workspace admin can install the app and join a 30-minute setup call.
Not the right fit if
You want meeting decisions and lessons captured into searchable memory, not channel answers. See AI Company Memory.
Your company chats in Microsoft Teams rather than Slack. See AI Company Knowledge Hub for Microsoft 365.
One small team needs a private assistant loaded with its own instructions. See AI Custom Team Assistant.
Most of your knowledge sits in Google Drive and Docs, with Slack only as the place to ask. See AI Company Knowledge Hub for Google Workspace, which can also answer in Slack.
How it works
- A workspace admin installs the Slack app, and on a 30-minute setup call we agree which channels are approved as sources.
- The field mapping records which channel messages, user details and app interactions the hub reads, what it may update, and which permission group applies.
- Messages from approved channels, plus up to 1 more system such as Google Drive or Confluence, are indexed, up to 2,000 documents in all.
- Staff ask the app directly in Slack, and the reply cites the channel thread or document behind it, so anyone can jump to the source.
- Up to 3 permission groups, configured for your Slack setup, stop a private team's channel content from appearing in someone else's answer.
- 60 benchmark questions run against live sandbox or test records before launch, and a monthly refresh adds new approved-channel history.
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 Slack app installed by a workspace admin.
- A Slack admin for a 30-minute setup call to approve source channels and the field mapping.
- The list of channels to index, and any that must stay out.
- Your user list and the permission group each person belongs to.
- 30 sample questions with expected answers, ideally ones already answered in a Slack thread.
- Access to one additional system if needed, and one decision-maker for approvals.
- 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
App installed, setup call done, approved channels and mapping signed off.
Weeks 1–2
Approved channel history and the extra system indexed and grouped.
Week 3
In-Slack question flow, citations and permission groups built.
Weeks 3–4
60 benchmark questions on sandbox or test records, accuracy report shared.
Week 4
Your review and two rounds of changes.
Go-live
Live in Slack, recorded handoff, 30 days of support.
| Option | Delivery time | Price |
|---|---|---|
| Standard | 20–30 working days | $11,400 |
| Express | 15–25 working days | $12,550 |
| Rush | 10–20 working days | $13,700 |
What you receive at handover
- The knowledge hub configured and running in your Slack workspace for up to 50 users
- Slack connection plus one more system, with a documented field mapping
- User access set up through invites or keys you control
- Test report on sandbox or test records, including the benchmark accuracy report
- Admin guide and a recorded 60-minute handoff session
- 30 days of post-launch support
Expected benefits and how to measure them
- Can cut down the repeat questions that interrupt your most experienced people.
- Helps new starters find answers in Slack without waiting for a colleague to reply.
- Makes it easier to trust an answer, because its source is one click away.
- Repeat questions in help channels
- Count questions posted in your main help channels for four weeks before launch, then for four weeks after.
- Answer quality
- Each week, sample twenty assistant answers and record how many were correct and cited the right message or file.
- Time to an answer
- Pick ten common questions and time how long a newer team member takes to answer each, before launch and after.
Before you start: Before work starts, count a month of repeat questions in your busiest Slack help channels. Note roughly how long those questions wait for a reply.
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 it costs roughly $50 to $300 a month in your own accounts for AI model usage, the vector database and hosting; busy channels and frequent questions sit at the higher end. Slack licenses and plan upgrades are not part of this build.
AI usage during build and testing is billed to your accounts as well.
Ask about running costsCustom requirements
Need custom Slack objects beyond the mapping, more channel history than the document limit allows, or a second extra system? Add a note at checkout for a written Change Quote; nothing extra starts until you accept it. Larger or bespoke projects go through the Custom AI request form.
Request a custom versionKeep it running
Monthly maintenance covers monitoring, fixes, model and platform updates, a monthly health report and monthly hours for changes, such as approving a new channel as a source or adjusting a permission group. It starts after the 30 days of support.
$570/mo (5% of the build price a month, includes 5 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 chat-search add-on bills per member monthly | A developer quotes a custom Slack bot by the hour | Ours is one fixed price, shown before you buy |
| Source channels | Access follows each member's channel membership | Depends on the build | Ours reads only the channels you approve |
| Citations | Returns matching messages to scroll through | Depends on what was scoped | Ours answers and cites the thread or file |
| Delivery | Installed in minutes, untuned to your channels | Usually several weeks | Ours has a fixed delivery time shown before you buy, with faster options |
Questions about this product
Which Slack channels does the knowledge hub read?
Only the channels approved on the setup call and recorded in the field mapping. Private and social channels stay out unless you add them. Messages in approved channels are indexed with a permission group, so content from a restricted team channel never reaches someone outside that group.
Will the hub post or change messages in Slack?
It replies when asked, through the installed app, and any other update is limited to what the field mapping allows. The Slack connection is two-way, but its scopes are documented and approved at intake, so it never acts outside them. Its main job is answering with a cited thread or file.
Can the Slack hub answer from documents as well as threads?
Yes. Up to 1 more system is included, such as Google Drive or Confluence, so answers can draw on documents and approved channel history together, each cited. A further source comes through the Knowledge Base Expansion add-on, or a Change Quote agreed before the work starts.
What happens to outdated answers sitting in channel history?
Every reply cites the thread it used, so an outdated answer is easy to spot and trace. During the 60-question benchmark you see which threads answered, and you can drop noisy channels from the mapping before launch. The monthly refresh then indexes new history from approved channels only.
What does our Slack admin need to do?
Install the Slack app in your workspace and join a 30-minute setup call to approve channels and the field mapping. Access details go through the intake process, never by email. All of it is due within 72 hours of ordering, and each day late moves delivery by 3 working days.
When does the Slack hub start answering in our workspace?
Delivery follows a fixed schedule from a verified intake, and the delivery panel shows that date plus express and rush options before you order. The clock starts once the Slack app is installed, the source channels are approved and your sample questions are in, so have your workspace admin ready on day one.
How much does the Slack hub cost to run each month?
Roughly $50 to $300 a month in your own accounts for AI model usage, the vector database and hosting. Your Slack plan is separate and unchanged. Monthly maintenance is optional, for changes such as approving new channels after launch. It also covers monitoring, fixes, a monthly health report and adjusting a permission group.
Find the answer your team already gave in Slack.
Order AI Company Knowledge Hub for Slack and your team gets cited answers inside Slack, from approved channels and one more system, on a fixed delivery date.
Works well with
Knowledge Assistants
AI Internal Search
Searches across your internal tools from one place with AI-ranked results.