How Many AI Initiatives Should a Company Run at the Same Time?
There is no magic number. The limit is how many projects your team can name an owner for and follow without chasing. Here is how to set it, and what to do when you are over it.
By Michael Urness · Saturday 21 November 2026
There is no right number of AI initiatives for every company. The useful limit is not budget or tooling. It is how many projects your team can name an owner for, and follow from idea to result without chasing people.
That is the short answer. The rest of this article explains why, how to set your own limit, and what to do if you are already over it.
Why is there no fixed number?
AI and automation are getting easier to build. That is good news. It also moves the constraint. The bottleneck is moving from development capacity to management capacity: deciding what deserves the next build, who owns it and whether the result was worth it.
Building one automation is mostly a development question. Running several at once asks something different. Who decides which idea gets backed? Who owns each one? How do you see where each stands?
How do you set your own limit?
Count the projects that currently have all four of the following. Anything missing one of them is a candidate to pause or finish before you add another.
- A named owner. One person who answers for the outcome, not a team or a vendor.
- A clear purpose. One sentence on the business result it is meant to produce.
- A visible status. You can see where it stands without calling a meeting.
- A review point. A date when someone asks whether it is working.
The number you can honestly count is your working limit today. It will rise as your management capacity rises, and not before.
A working rule
This is a view, not a measured finding. Start with one or two initiatives, finish them properly, and add the next only when the owners of the current ones are not the constraint. A bounded portfolio of projects that are owned and followed usually beats a long list of pilots that nobody is steering.
What if you are already over the limit?
- List every AI project and automation running in the business, including the small ones.
- Write the owner next to each. Where the owner is blank, that project is the first one to decide on.
- Choose which to back, which to finish and which to stop. Write down the ones you chose not to back, and why.
- Set one review date for each project you keep.
An example from a real engagement
Incra, an international accounting firm, wanted to move beyond a few custom applications and automate more of its business. Mike helped the team establish the development process and environment to support ongoing, increasingly complex projects. The team is now developing internal automations at scale, with a foundation for expanding its use of AI across the business.
No outcome has been measured yet, so none is claimed here. What can be said is that the team now has a way of working that carries more than one project at a time.
Where Better Execute fits
Better Execute helps leadership teams get business performance from AI and automation by deciding what deserves the next build, who owns it and how the result is proven. If you want to see where your own count sits, book a free 30 minute Competitive Advantage Review.
Building more is getting easier. Owning the result is what sets the limit.
Want to talk through whether DCE is a fit for your leadership team?