Slack AI chatbot for HR — answers from your handbook, not the internet

Type "Slack AI chatbot" into a search box and two different kinds of product come back under one name. The first is a general-purpose assistant in a Slack window: ask anything, get a fluent paragraph assembled from the internet at large. The second is wired into your company's own facts — it reads a specific source before answering, and it can act on what was asked. AI HR Bot is the second kind, scoped to HR: policy questions, leave, and the small requests in between.

The boundary before the argument: the Free tier covers up to 5 Employees, with policy answers, Leave Requests, and up to 25 Knowledge Base entries; Employee Memory and Leave Policies with monthly Accrual sit in the Team tier, at $4 per employee per month. Nothing below depends on reading the pricing page generously.

Here is the second kind, in a channel:

Where the answers come from

A policy question is answered from your own Knowledge Base entries, matched against the question that was asked — the handbook your HR Head uploaded or taught, not the internet's idea of one. The bot has no browsing tool: when it lacks an answer, it cannot reach for the web and present some other company's policy as yours. An answer joins the Knowledge Base only when your HR Head uploads it or teaches it.

That is the part the "AI" in the name tends to hide. The model is a commodity, and every vendor's model writes the same fluent paragraph. What separates one Slack AI chatbot from another is what it is allowed to read — and what it does when the reading comes up empty.

What it does, not just says

A chatbot that can only talk is a search box with manners. Here, "Can I take next Friday off?" files a Leave Request and starts the Approval Chain in the same step — the card lands in the Lead's Slack as a direct message with Approve and Deny buttons, not in a queue nobody checks. The Approval Chains page draws that card.

The Leave Balance moves when the Request is approved, not when it was filed — so the balance the chatbot quotes while the card waits is still the balance actually held.

What happens when it does not know

The failure that matters in an AI chatbot is not the slow answer; it is the confident wrong one. When no Knowledge Base entry clears the similarity bar, the refusal's wording is prescribed for the model — down to "Do NOT guess" — and the miss is recorded as a Knowledge Gap either way. The bot then opens a direct message with the first HR Head listed for the Workspace — one person, not a broadcast — asking to be taught, and the reply becomes the saved answer for the next person who asks.

Here is that loop, in both places it happens:

A prompt is an instruction, not a guarantee, and the case for trusting the structure around the model rather than its manners is made in full in what happens when an HR bot doesn't know the answer.

What it remembers, and what it can read

A chatbot inside your Slack raises two reasonable worries, and they have different answers.

What it remembers is Employee Memory: context kept per Employee between conversations, so nobody re-explains their situation every time they ask.

What it can read is narrower than most people assume: exactly two of its Slack permissions read messages — the direct messages people exchange with the bot itself, and the messages that mention it. Channel history and other people's DMs are outside the request, and every permission is named, one by one, in what a Slack HR bot can read.

Getting it into your Slack

Connecting the bot is one Slack consent screen, and the install guide walks through what follows — including the Approval Chains a fresh install seeds for every Request type. For a team of up to 5 Employees, both halves shown above — the answering and the filing — cost nothing.

If "Slack AI chatbot" was your search, the comparison points that matter are three: what the chatbot reads before it answers, what it files when you ask, and what it says when it has nothing to read. Every claim this page makes about those three links to the page that treats it in full.