Many teams already use artificial intelligence to edit messages and summarize documents. The answers got better, but the way the work is organized looks much the same. The next opportunity for AI in a business usually sits in an activity that does not yet show up among the requests.
It can also go the other way: one task gets faster, but the workflow gains a step. To know whether AI changed the business, you have to look at where it can help and at the full path between the need and the result.
Where AI can help beyond the obvious
Three situations help widen the repertoire: information that arrives too late, a comparison no one manages to prepare and an experience the company would like to offer. Each one opens a different conversation. The technology can support preparing the work, analyzing alternatives or exploring something new.
Think of a retail chain that only gathers customer feedback when someone complains straight to management. Before automating replies, leadership could ask which signals go unnoticed between one complaint and the next.
A first experiment could sort an approved sample of comments by topic, keeping the original excerpts. The goal is to find questions worth investigating: does a buying difficulty keep coming up? Does the product description set the wrong expectation? Is there a need the company does not yet meet? The AI's classification is a starting point; an apparent pattern has to go back to the messages before it guides a change.
Give the exploration a purpose
Gather situations from the past week where information, clarity or room to experiment was missing. Pick one and describe what you would like to notice or offer better.
In the retail case, the result might be a question for the next conversation with customers. That is already different from asking for a nicer summary: the deliverable now supports an investigation that did not happen before.
Did AI change the process or just add a step?
A team uses AI to prepare a summary and now reviews one more document before the meeting. The task got faster, but the workflow grew. An automation can be useful even if it changes little. The problem starts when the speed of one part hides waiting, duplication and reviews in the rest.
Map today's path. In a weekly report, teams send updates, someone consolidates them, someone else reviews and leadership receives the result. Now add AI to the map. Which handoff changed? Did any information stop being copied? Was a review replaced or only duplicated? Does the report arrive in time for the decision?
If consolidation got faster but teams still send incomplete data, the limit is in how the inputs are prepared, not in the tool.
Test a redesign with a clear consequence
One option would be to collect updates throughout the week and prepare only the relevant changes for the meeting. That would require accessible sources, relevance criteria and someone responsible for checking the summary. It is a hypothesis to test, not a ready-made recommendation.
Ask the team to show what would stop happening, what would start happening and which conditions would be needed. At the end, compare total time, corrections and the quality of the decision it supported. The number of steps using AI, on its own, says little.
What this means for whoever runs the business
For founders and senior leaders, the useful question is not "where can we use AI?" but "which part of the work stays unattended because it always seemed too hard to keep track of?". The answer shows where AI for business creates a business gain, not just a speed gain.
That examination is at the heart of AI Thinking, Dani Hayoshi's approach: question the handoffs in the work, redesign with AI and rebuild what still makes sense. For the role of whoever leads the team, go to AI for leaders. To think about where the freed-up time should go, see AI, productivity and time.
Dani Hayoshi is the author of AI Thinking. Her work connects artificial intelligence, education and entrepreneurship. About Dani.