Niches · Desktop & mobile app control
Find and start a robot a company already runs, and return its result
No company's docs show a full fit yet. The listings below do the job with a limit, or not at all.
What counts as done
The agent lists the matching published processes, starts one with typed input arguments and files, gets a job id, and reads the job's final state and output arguments from the API or MCP server, with no person signing in.
Measured, for our own tests: start to final state is reported within one poll interval of the job ending. No docs we read state these figures, so they don't count against a listing here.
How agents ask for it: run the invoice bot for these three files
, start the SAP posting process
, kick off the month-end robot and tell me when it's done
Does the job, with a limit 3
It does the job, but something in the docs or a user's account will get in an agent's way. The limit is named in each row.
| Listing | Why |
|---|---|
| Automation Anywhere D Automation Anywhere, Inc. | DiscoverAutomation finds runnable automations, RunAutomation takes matching input variables and returns a job id, and GetAutomationResult returns Running, Completed with output variables, or Failed with errors. Discover and Run need an Enterprise licence with Process Reasoning Engine, US and Europe only, and return at most nine matches. source |
| Robocorp B Sema4.ai (full legal name not confirmed) | RPA.Robocorp.Process lists processes, starts one with input work items, and reads run status, work items and artifacts from the Control Room Process API. Control Room docs and prices on sema4.ai sit behind a bot check, so current availability couldn't be confirmed. source |
| UiPath B UiPath Inc. | Folder-scoped UiPath MCP servers expose published processes as tools, and StartJobs takes InputArguments and returns job ids whose State and OutputArguments GET Jobs reads. acc-01 hit an invalid-scope error from Claude, acc-06 got null file arguments, and acc-08 saw a job stall and fail. source |
Doesn't do this job 7
Listings in the category that do something else, with the reason.
| Listing | Why |
|---|---|
| Aura Autopilot C TribeVR Inc. (named in Aura's terms of service and privacy policy), trading as Ramen | No RPA robots or orchestrator; runs are playtests started per prompt. source |
| AutoCua D Ashish Yadav and contributors (ComputerUseAgent on GitHub) | No robot catalogue, job API or orchestrator; each run is a new model-driven task. source |
| AutoHotkey C AutoHotkey Foundation LLC | A local scripting language with no robot catalogue, job API or orchestrator. source |
| AutoIt D AutoIt Consulting Ltd | A local scripting language with no robot catalogue, job API or orchestrator. source |
| Beam (beam.win) F Noats, Inc. | No robot, job or orchestration concept; it shares windows. source |
| Mage-VL-Cowork C Microsoft | A model, not an orchestrator; no robot catalogue or job API. source |
| Robot Framework BB Robot Framework ry (Robot Framework Foundation), business ID 2754775-1, Helsinki | No hosted orchestrator, robot catalogue or job API; the foundation runs no service. source |
3 listings in the category were added after this reading and aren't marked yet.
What stops agents finishing
- discovery and run tools need an enterprise licence or region
- OAuth scope mismatch between the MCP client and the external app
- robot job outlives the client's tool-call timeout with no job id to poll
Around it
Before: agent-auth.client-credentials. After: email.send, human-in-the-loop.approve-action.
Where this comes from
Weight 14 of 100 in Desktop & mobile app control: 6 of 45 accounts, all UiPath forum threads; Automation Anywhere rests on vendor docs only
The same as data: /api/v1/use-cases/desktop-automation.start-existing-robot.json, or the MCP tool get_use_case with id desktop-automation.start-existing-robot. Other jobs in this category: /api/v1/niches?category=desktop-automation.
For companies
Do agents find, use and choose your tools?
An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.