The 15‑Run Limit: Why Scheduled Prompts Still Feel Half‑Baked in Copilot

Why only 15 runs?

For anyone exploring the fantastic new Frontier scheduled prompt features in Microsoft 365 Copilot and agents, this is a common question!

We can now schedule more prompts in more places and in more flexible ways, and these new capabilities will be moving to General Availability by next month according to MC1189911. But we are still limited to only setting up 15 distinct runs on that schedule.

Now, I can understand why Microsoft might not want these to be set-and-forget functions. Particularly if you're using an agent like Researcher, you're consuming a lot of tokens, costing them a lot of inference compute, and the safest way to stop unnecessary use is to simply limit how many times you can make your prompt run.

However, if the parallel we want between humans and AI automations is that they will form part of our teams, it would be a pretty poor collegial experience if every 15th time a coworker carried out an important process, they just stopped and stared blankly into space until we told them to start again.

Of course, if you decide to go a different route and set up your process as an Autonomous Copilot Studio Agent, Microsoft will happily run that for you whenever you like and for as long as you choose, but then that billing is on you with Copilot Credit consumption rather than included in your all-you-can-eat Microsoft 365 Copilot license.

Perhaps what would be great is an easy way to "promote" important scheduled prompts into autonomous agents? This would add a whole set of extra complexity though, as many of the capabilities available in Copilot Chat are not directly addressable from Copilot Studio.

Is this a problem you've run into? What solution would you propose?

First posted on Linkedin on 01/15/2026 -> View post on Linkedin Here