Overview

Wilson is the AI coworker in your Slack, built on the Dynamiq platform and delivered as its own product at hirewilson.ai. You talk to it the way you talk to a teammate — a direct message or an @mention in a channel — and it does the actual work: reports, dashboards, decks, code, whole campaigns, delivered back into the same Slack thread. Under the hood it is the same super agent that powers Chat: the same agent loop, the same cloud sandbox, the same connector catalog. Wilson just meets your team where they already are.

What Wilson can do

Wilson runs the same agent as Chat's Dynamiq Agent mode, so it brings the full built-in toolset — web search, browsing, a code-and-terminal sandbox, file generation, image generation — to every Slack conversation. On top of the built-ins it uses the apps your organization connects, discovering each app's tools and calling them as the task needs.

Because Wilson lives in Slack, a few behaviors are tuned for that surface:

• Deliverables come back as native Slack file uploads in the same thread — a report, a spreadsheet, a slide deck — with a short summary alongside rather than a wall of text.
• It uses the surrounding thread as context. Mention Wilson in an existing thread and it reads what came before instead of asking you to restate it.
• It confirms before risky actions — deleting or overwriting data, messaging people outside the thread, spending money, or mutating production systems.

Flagship connectors: AWS and SQL databases

Wilson draws on the same connector catalog as Chat, surfaced for the whole organization through the Integrations page (see Using Wilson). Two of those connectors are what make Wilson useful against real infrastructure:

• AWS. Connect an account with an Access Key ID and Secret Access Key, and Wilson can work against your AWS resources as part of a task — pulling data, driving the CLI in its sandbox, and folding the result into whatever it is producing.
• SQL databases. Wilson queries PostgreSQL, MySQL, ClickHouse, Vertica, and Trino connections through the agent's SQL tooling. Most production databases have no public endpoint, so each of these connection types accepts an SSH tunnel: Wilson connects to a bastion (jump host) you expose and routes the database traffic through it. See SSH Tunnels for the bastion setup and the exact tunnel fields.

This is the core loop for a data team: point Wilson at a read-only database behind your bastion, and ask it — in Slack — for the analysis, the chart, or the file, without anyone opening a SQL client.

Wilson vs. web Chat

Wilson and Chat are the same agent on two different surfaces. Choose by where the work lives:

• Use Wilson when the request starts in Slack, involves your teammates, or should hand its output straight back to a channel or DM — "summarize this quarter's churn," "turn this thread into a brief," "pull yesterday's signups from the warehouse and chart them."
• Use Chat when you want the full web workspace: a model picker per conversation, per-conversation connector toggles, the interactive tool-details and file panels, and a persistent conversation sidebar.

Next steps