Use cases / AI consultants

Hand off AI work that keeps running

Your client keeps the work. You keep delivering.

Your AI builds the client's skills, apps and automations in their own workspace. It runs in the cloud, works with any AI tool, and stays open for your next delivery.

What happens after the handoff

  • The agent you built runs on your laptop, or on a server only you know about.
  • It works with the AI tool the client uses today. Next quarter they move half the team to another one.
  • At the client it lives with one person. When that person leaves, nobody can run it.
  • After the handoff the client cannot see what you still do for them, so the retainer is hard to explain.
  • Every client's tools, instructions and keys sit somewhere different.

1Package it

Build in the client's workspace

Your AI (Claude, ChatGPT, Cursor, Codex and 60+ AI tools) builds directly into a workspace the client owns. Skills, apps, automations, data and files sit together in Spaces, one for each piece of work.

  • Skills, saved once. The way the client's work is done is written down as skills in the workspace. Any AI tool connected to it reads the same ones.
  • Apps and records. Dashboards, forms and the records behind them, at a real address the client's team opens.
  • Connected tools. HubSpot, Google, Slack and 3,600+ integrations, connected in the client's workspace, not in your accounts.

2Hand it off

It runs without you

Nothing to move on the last day of the project. The work already lives where the client works.

  • Runs in the cloud. Apps, scheduled jobs and workflows run on Runwork's cloud. Your laptop can be closed, or in another client's office.
  • Any AI tool. The client's team uses the work from Claude, ChatGPT, Copilot, Gemini or whatever they choose next month. The skills are stored in the workspace, so each tool reads the same ones.
  • The client owns it. The workspace has its own plan and the client pays for it. They can export their data any time, free on every plan. If they remove you, the apps, skills and automations stay.

3Keep delivering

One place for your next delivery

The handoff is not the end of the engagement. It is where the retainer starts to make sense.

  • New work lands in the same place. Next month's automation goes into the workspace the team already uses. No new tool to roll out.
  • Each client's setup in one place. Approved AI tools, team instructions, models and permission rules, set once in the client's Team Configuration and sent to each person's AI tool.
  • The client sees each change. Every change is in the workspace audit log, so your work stays visible after the handoff.
  • One login for every client. You are a member of each client's workspace, with the role they give you. Switch between clients from the dashboard, or from the command line.

An engagement on Runwork

One client, one pipeline report, and the three months after it.

  1. Week 1

    The workspace. The client creates their Runwork workspace and invites you as an admin. You connect HubSpot and Google Drive in their workspace.

  2. Week 3

    The build. Your AI builds a weekly pipeline report: a skill for how the client scores deals, an app for the report, and a schedule that runs it every Monday at 8.

  3. Week 6

    The handoff. There is nothing to package. The sales lead opens the report from their own Claude. The schedule keeps running when you close your laptop.

  4. Month 2

    A tool change. Part of the team moves to ChatGPT. The skills are stored in the workspace, so ChatGPT reads the same ones.

  5. Month 3

    The next delivery. You add an automation for renewal reminders. The client sees it in the audit log. That is what the retainer pays for.

Questions

Who owns the workspace, and who pays for it?

The client. Each workspace has its own plan and its own billing. You join as a member, and you use one of their seats. See pricing for the plans.

What happens when the engagement ends?

The client removes you from the workspace. The apps, skills, automations and data stay, and the owner can move the apps you owned to someone on their team. The client can export their data any time, free on every plan.

Can I work with several clients at once?

Yes. One login, a separate workspace for each client, and a different role in each. Switch from the dashboard, or pass --workspace in the CLI. The desktop app connects a computer to one workspace at a time and switches when you do.

Which AI tools does it work with?

Claude, ChatGPT, Cursor, Codex, Gemini, GitHub Copilot, Microsoft Copilot and 60+ AI tools in total, through the workspace MCP server. The client's team can use different tools at the same time.

Can I reuse my own methods across clients?

Keep your methods as skills in your own workspace, and copy a skill into a client's workspace with the CLI. Client data stays inside each client's workspace.

Does the client see what I do in their workspace?

Yes. Changes are recorded in the workspace audit log, which every member can open.

Start with one client

Pick the client you are building for now. We set up their workspace together and walk through it with you.

Building AI for your own team instead? Runwork for consultants keeps what you work out on one engagement for the next.