By Barry O’Reilly, Author of Artificial Organizations
Published with IIL, Corporate Research Partner on the AI Outcomes Benchmark.
November 18, 2026
Your status report used to take two hours. Now it takes ten minutes.
But did AI help you spot the project going off track any earlier? Did you challenge a dependency before it became a delay? Did anyone make a different decision while there was still time to change the outcome?
Or did you just report the same bad news faster?
That is the challenge for project leaders. If you are following the same routines, asking the same questions and holding the same meetings, sprinkling AI onto the process may leave your biggest opportunity untouched.
Why doesn't faster AI reporting improve project outcomes?
Short answer: Faster reporting tells you what already went wrong. Better outcomes come from using AI to change how decisions get made, early enough to act.
One of the reasons I wrote Artificial Organizations was a study I ran twelve months ago to understand how leaders were actually using AI at work.
Much of what we saw was familiar: drafting communications, summarizing information and getting through administrative work quicker. Only 13%, roughly one in eight senior leaders, identified better, faster decisions as their biggest return from AI.
That finding stayed with me. We had tools capable of helping people challenge assumptions, examine alternatives and recognize patterns. Yet so much attention went toward producing the same outputs faster.
For project professionals, the opportunity is to redesign how you lead the work.
How can project managers use AI for better decisions?
Short answer: Use AI to surface early warning signals across weeks of project data, then verify those signals with the people closest to the work and act while options remain.
A status report tells you a milestone slipped. Better decision support helps you recognize the conditions that could cause it to slip, examine your options and act before those options disappear.
That requires a change in behavior. Instead of asking AI to summarize this week’s updates, use it to examine changes across several weeks:
- Which dependencies remain unresolved?
- Where are estimates repeatedly changing?
- Which assumptions behind the delivery plan deserve another look?
Then take those signals to the people closest to the work. Check what is real, explore the consequences and decide what to do.
The value comes from what you notice and change while you still have room to act.
How do you measure whether AI is improving your work?
Short answer: Define the outcome you want AI to improve, establish a baseline from recent examples, then test a new routine and compare results.
Our current research is exposing how difficult that shift remains. In early responses, 94% said AI helps them complete more relevant work in the same amount of time. Yet only 52% said AI experiments are evaluated against a clear baseline, prior method or expected outcome.
These are early, self-reported findings. But they raise a question every project leader should answer: have you defined what you want AI to improve?
Example: using AI to manage project dependency risk
Consider a project where dependencies repeatedly cause late surprises. Your first instinct might be to automate the dependency report. Instead, follow these steps:
- Examine how dependencies become decisions. When do teams raise concerns? What information gets lost? Who can act, and how long do they wait?
- Define the outcome. More time between recognizing a dependency risk and the point when it disrupts delivery.
- Set a baseline. Look at recent examples to establish your starting point.
- Test a new routine. Before the weekly meeting, use AI to surface unresolved dependencies, changing assumptions and conflicting commitments from information your team already maintains. Ask the relevant people to verify the signals. Organize the conversation around the decisions those signals require.
- Compare results over several cases. Are risks recognized earlier? Do owners act sooner? Are fewer commitments missed because a dependency was overlooked?
- Count the costs. Include false alarms and review effort. An intervention that creates more noise needs adapting.
Now you are learning whether a different way of working improves the project.
What is the Artificial Organizations AI Outcomes Benchmark?
Short answer: A free, anonymous, 10-minute assessment that shows how your AI practices compare with an emerging market benchmark, plus guidance on what to strengthen next. Submissions close November 30, 2026.
You probably recognize some of these habits in your organization. The harder question is which ones deserve your attention first.
Are individuals getting faster while project bottlenecks persist? Are teams changing how decisions get made? How do your practices compare with others trying to make the same shift?
That is why we are building the Artificial Organizations AI Outcomes Benchmark. In about ten minutes, you can establish a starting point based on your experience of how AI is used at work. You will receive insights into how your reported practices compare with the emerging benchmark, alongside guidance on behaviors to strengthen and interventions to try.
Perhaps your next step is defining a meaningful outcome. Perhaps it’s testing AI against a recurring decision, or helping an effective practice spread across teams.
As the year ends, understand where you stand before deciding where to invest more effort.
Take the free, 10-minute AI Outcomes Benchmark and invite colleagues across your network to contribute. It is anonymous, and the deadline is November 30, 2026.
Your AI is saving you time. Make the year ahead the year you can show how changing the way you work led to better project outcomes.
Key Takeaways:
- Only 13% of senior leaders (about 1 in 8) say better, faster decisions are their biggest return from AI.
- In early benchmark responses, 94% say AI helps them complete more relevant work, but only 52% evaluate AI experiments against a clear baseline.
- Project leaders get more value from AI when they redesign how decisions get made, not just how reports get written.
- Start by defining the outcome you want AI to improve, set a baseline, then test a new routine.
- The free, anonymous, 10-minute AI Outcomes Benchmark closes November 30, 2026.
Frequently Asked Questions
What percentage of leaders say AI improves decision-making?
In Barry O'Reilly's study of senior leaders, only 13% (about 1 in 8) named better, faster decisions as their biggest return from AI.
Do organizations measure the impact of AI experiments?
Often not. In early AI Outcomes Benchmark responses, only 52% said AI experiments are evaluated against a clear baseline, prior method or expected outcome.
How should project managers use AI beyond status reports?
Use AI to spot patterns across several weeks of project data, such as unresolved dependencies, shifting estimates and shaky assumptions. Then verify those signals with the team and decide what to do while there's still time to act.
What is the AI Outcomes Benchmark?
A free, anonymous, 10-minute assessment from Barry O'Reilly's Artificial Organizations research, with IIL as a Corporate Research Partner. Participants see how their AI practices compare with an emerging benchmark and get guidance on what to improve.
When is the AI Outcomes Benchmark deadline?
Who is Barry O'Reilly?
Barry O'Reilly is the author of Artificial Organizations and leads research into how organizations turn AI use into measurable outcomes.
Barry O’Reilly
Barry O’Reilly is a bestselling author, entrepreneur, and globally recognized expert on innovation, leadership, and organizational transformation.
He is the author of Artificial Organizations: Build Better Judgment, Speed and Results with Human and Machine Intelligence, and Unlearn: Let Go of Past Successes to Achieve Extraordinary Results, and co-author of Lean Enterprise, part of the Eric Ries Lean series.
He is co-founder and Chief Innovation Officer of Nobody Studios, an AI-driven venture studio, and a faculty member at Singularity University.
Barry’s work helps executives rethink how they lead, make decisions, and build high-performance organizations in the age of AI. His latest work explores how organizations can combine human judgment and machine intelligence to increase decision velocity, create decision advantage, and achieve better results.
To learn more about Barry O’Reilly, go to barryoreilly.com.
Follow Barry on LinkedIn: https://www.linkedin.com/in/barryoreilly/