My first trip to New York City was three companies ago. I was a young executive there for the meeting you prepare for by reading your deck a thousand times.
Top floor of the headquarters of one of the biggest banks in the world. A conference room designed to remind visitors exactly where they ranked in the food chain. At the head of a table far too large for any reasonable conversation was the VP running the meeting.
I knew the pitch cold. I could have given it blindfolded. I pulled up the deck on the big screen, and before I could get a word out, the VP stopped me.
“Why should we give our business to a small company from Oklahoma?”
I froze. Didn’t stumble. Didn’t reach for a mediocre answer. Froze. The question wasn’t complicated, it was just a question I had never thought about. I had prepared for the presentation, not the meeting. I knew exactly what came next as long as everyone else followed the script I had scribbled in my head. The moment he changed the conversation, all that preparation became useless.
I had a plan, not a range.
For years, I thought the lesson from that meeting was that I should have prepared more. I had it backward. Another ten hours rehearsing the same pitch would only have made me better for a meeting that never happened. What I needed were options.
What if they challenge our size? What if they don’t care about the demo? What if they want to talk about risk? What if the first question blows up the agenda?
Rigidity photographs like confidence. Under the load, it behaves like brittleness. The faster the world moves, the more expensive it becomes to have only one move.
In 1956, British cybernetician W. Ross Ashby formulated the Law of Requisite Variety: only variety can destroy variety. Stafford Beer, who carried Ashby’s math into management, restated it the way most people remember: only variety can absorb variety. It’s also the most useful sentence you’ll read about AI strategy this year. The more ways a situation can surprise you, the more responses you need to retain control.
The law has two sides:
- The variety being thrown at you.
- The variety you can throw back.
Control means having enough of the second to absorb the first.
For seventy years, that was a fair fight; both sides were human-scale. Your competitor could only draft so many proposals. Your client could only research so many alternatives. Everyone rationed their options because every new option cost time and money. Rigidity was a budget decision.
Now, AI has given both sides nearly unlimited moves.
More Moves Coming at You
On one side of the equation, everyone you face is now armed with more options. The client walks in with a machine-built benchmark of your pricing against four competitors. The RFP draws thirty responses instead of eight because drafting got cheap for everyone. Your competitor tests twelve versions of an offer while your team aligns on one.
This is a real-world wake up call. At MIT Sloan, negotiation professor Jared Curhan and his colleagues ran a competition in which AI agents bargained against one another more than 180,000 times, executing and refining strategies at a scale and consistency no human negotiator could match. Research findings showed that aggressive agents sometimes won better terms, but they drove deal after deal into deadlock. Warm agents kept talks alive and closed more of them. Hardball lost on moves and countermoves. A deadlock is a negotiation with no options left.
The negotiator across from you doesn’t have to be a machine to benefit from one. He only has to show up having rehearsed his moves with one on the drive over.
In addition, Ashby’s law doesn’t grade on effort. When the variety coming at you is machine-generated, human-scale responses stop absorbing it. You can’t outwork it. Starting earlier or working longer is no match for a counterpart whose preparation has been multiplied by a model.
You are not losing on effort. You are losing on arithmetic. And the math isn’t close.
More Moves to Throw Back
On the other side of the equation, Ashby’s law is an instruction. If the world has more moves, you need more moves. More important than speed, the biggest value of AI is range. Efficiency in doing the same move faster is the shallow reading. The deeper reading is variety: having more options available when the situation changes.
An analyst can generate five deal structures before lunch, buying herself moves. A capture manager can pressure-test a proposal against twenty objections, expanding the range of surprise responses he can absorb. The person using AI well can manufacture variety, and variety is what control is made of.
The caveat here is that variety is not equivalent to volume. Ten near-identical answers are still one move. AI expands your control only when someone can test the options, discard the weak ones, and choose the one that fits the room. Generation is cheap, so judgment is the job.
Having range means holding the standard steady while changing the approach. Your values stay fixed. Your tactics stay flexible. While AI can multiply the ways you pursue an objective, it still requires you to decipher which objectives are worth pursuing.
The machine widens the playbook. You write the rules.
Match. Multiply. Hold.
In practice, Ashby’s law comes down to three disciplines:
- Match the variety coming at you.
- Multiply the responses available to you.
- Hold the values that determine which responses are acceptable.
Match. Read the room at its new scale. The person across from you may arrive with ten scenarios, twenty objections, and a machine-tested response for each. Prepare for the range in front of you, instead of the limits you are used to.
Multiply. Three deal structures are more valuable than a faster first draft. The goal is to have another move in your pocket when the room changes.
Hold. Let your tactics multiply without moving the line. Range without a fixed point becomes drift. Methods flex, while values do not.
Match the room. Multiply your moves. Hold your line.
Now, let’s run the New York meeting back through the framework:
- Match: The VP arrived with the range, and I arrived with a script. His first question showed that I hadn’t read the room or been prepared properly to be in this room.
- Multiply: Ten more hours of rehearsal wouldn’t have helped. It would have been the same move, just run again. The preparation needed to be multiplied. It needed to take into account answers for size, risk, why-Oklahoma, and so much more.
- Hold: The deal I flew there to win never needed to change. Only the routes to it did. More moves only create control when you know what never moves.
I had one move. He only needed one question.
The Company Version
Organizations make the same mistake I made in that conference room. One approved tool, one pilot, one locked workflow deployed with a firm handshake and a fixed agenda.
The rollout is presented as governance. Under the load, it behaves like brittleness. Surprises keep arriving from directions the pilot never anticipated, and the organization has only one sanctioned response.
The pattern is easy to recognize. A company approves an assistant, limits it to a narrow set of uses, and treats everything outside that boundary as failure. Then the work changes, because that’s what work does. A new exception appears. The client asks for something the workflow was never designed to handle. The allowed move no longer fits the situation. But the people closest to the work don’t stop. They open three other tools and keep moving. When MIT researchers surveyed companies in 2025, only 40 percent had purchased an official AI subscription, but workers at more than 90 percent of them were already using personal AI tools on the job. The gap between those two numbers is where your workflow actually lives.
The prohibition didn’t reduce variety. It pushed variety into the shadows, where you can’t see it or govern it. The organization locked the workflow to create control, which became the reason it lost control.
A single approved move may look like discipline. But controlling the workflow is not the same as controlling the situation.
Count Your Moves
Start by asking how many options your team actually has. Pick the meeting that matters most this month and count the moves you will have when the first question blows up the agenda.
If the honest answer is one tool and one plan, you do not have a strategy. You have made a bet that the room will not ask another question.
The person with more good moves is harder to corner. AI is the cheapest good move ever invented.
Options are power. If you’re ready to see what AI can do for your organization, we can help. Doyon Technology Group has AI experts on staff to help you design an AI roadmap that fits your needs and timeline. Connect with us today to learn more.
––––––
About the Author
Greg Starling serves as the Head of Innovation & Growth for Doyon Technology Group. He has been a thought leader for the past twenty years, focusing on technology trends, and has contributed to published articles in Forbes, Wired, Inc., Mashable, and Entrepreneur magazines. He holds multiple patents and has been twice named as Innovator of the Year by the Journal Record. Greg also runs one of the largest AI information communities worldwide.
Doyon Technology Group (DTG), a subsidiary of Doyon, Limited, was established in 2023 in Anchorage, Alaska to manage the Doyon portfolio of technology companies: Arctic Information Technology (Arctic IT®), Arctic IT Government Solutions, and designDATA. DTG companies offer a variety of technology services including managed services, cybersecurity, and professional software implementations and support for cloud business applications.