As AI becomes further and further integrated in our lives, it’s involvement in decision making is taking shape in interesting ways.
What started as a tool for drafting emails, summarising documents and generating ideas is now being used to shortlist vendors, compare options, prepare recommendations and influence business decisions. In many ways, this is a positive shift. Used well, AI can help teams move faster, consider more information, and reduce time spent on repetitive analysis.
But there is a difference between using AI to support a decision and allowing AI to make the decision for you. That distinction is becoming increasingly important.
Use AI to Expand Your Thinking
AI is particularly useful at the start of a decision-making process. It can help you get unstuck, generate different angles, or identify questions you may not have thought to ask. For example, you might ask AI to list possible risks in a project, compare three implementation approaches, or create decision making criteria.
The key is to use AI for breadth, not authority. Treat the first response as a starting point, not a conclusion. Good prompts might include:
- “What are five different ways we could approach this problem?”
- “What assumptions am I making?”
- “What would a customer, CFO, salesperson and operations manager each care about in this decision?”
- “What are the strongest arguments against this recommendation?”
This is where AI can genuinely improve thinking. In one study discussed by MIT Sloan, highly skilled workers using generative AI within its capability boundary improved performance by nearly 40% compared with workers who did not use it. The important phrase there is “within its capability boundary”. AI is most helpful when it is being used to expand, test and organise thinking, not replace it.
A Clear Answer Isn’t Always a Correct One
AI can be inaccurate, sometimes in hilarious fashion.
Other times the errors are not so obvious.
These errors are often called hallucinations. Responses that sound plausible, but are wrong, incomplete, or unsupported. The issue is not simply that AI can make mistakes. Humans make mistakes too. The issue is that AI often presents mistakes with the same confidence, formatting and tone as accurate information.
This matters most in high-stakes or specialised areas. A Stanford/Yale-linked study on legal AI tools found that even leading legal research products hallucinated between 17% and 33% of the time, despite being designed for a specialised professional context. That does not mean AI is useless. It means AI outputs need review, especially when they are influencing contractual, financial, technical, legal or strategic decisions.
The danger is that an answer can feel correct because it is clear. A well-structured table, confident summary or neatly ranked recommendation can make weak information appear strong.
The Risk of Letting AI Think for You
Picture this. You’ve just given AI a task, it’s responded with a concise, helpful answer, formatted in a way that your can easily understand. What a joy; now it asks “would you like me to ….. for you?, I can …….., just let me know!”
We’ve all felt that feeling. It’s the feeling you get when you feel heard and understood. The comfort you feel when someone provides you with exceptional service. If only every colleague interacted with you that way.
The problem is this feeling can be addictive. Several levels deep you have more answers than you imagined, and not once have you had to inject a single thought.
When you take away the challenge you remove the need to overcome them. You have to be sure you don’t lose critical thinking.
This is one of the least discussed risks of AI adoption. The obvious fear is that AI will replace people. The quieter risk is that people slowly outsource the thinking that makes them valuable.
AI can remove friction, but some friction is useful. Questioning a recommendation, debating a trade-off, checking an assumption, or sitting with uncertainty are all part of good decision making. If AI compresses every messy question into a neat answer, teams may become faster at producing outputs while becoming weaker at forming judgement.
In practical terms, businesses should avoid asking AI only for answers. Instead:
- Ask it for counterarguments
- Ask it to identify missing information
- Ask it what would change its recommendation
Most importantly, be conscious of the bias you may have picked up from AI along the way. It can be very easy to believe our ideas are our own, despite being initially planted by another person or in this case, AI.
Don’t Let AI Turn Decisions Into Loops
Proper usage of AI in decision making scenarios will help you identify and resolve things you may not otherwise think of. This can speed up the process as you are able to summarise and evaluate far quicker.
There is however a point where this process can slow you down. With infinite information, next steps and suggestions, it can be easy to get stuck in an infinite loop. Furthermore, when evaluating vendors, you may be inadvertently putting them through undue process, extending the review process, and ultimately costing more time.
One way to avoid this is to define the decision criteria before using AI. What are you actually trying to decide? What information is essential? What would be “good enough” to move forward? Without these guardrails, AI can encourage endless comparison.
For example, if you are evaluating software vendors, AI can produce a long list of questions, risks and alternatives. Some of these will be useful. Others may create unnecessary process, especially if they do not reflect your actual business requirements. The result is a slower procurement cycle, more meetings, and frustrated vendors, all without necessarily improving the final decision.
Use AI to accelerate the path to a decision, not expand the decision-making process forever.
Your Clients Are Using AI Too
You’re not the only one utilising LLM’s to speed up your workflows. Your clients may be doing the same thing, evaluating and summarising your interactions with the aid of AI. Some things worth noting:
- They may not be reading your proposals, but an AI summarised version of your proposal. You should run any important documents through AI to see how it may interpret your communications.
- They may very well be using AI to sense check your recommendations, and ask for alternatives. You need to be ready for the possible objections and able to justify your recommendations.
- They may not fully understand the responses they provide you. Be careful answering questions for what they are, and seek to get to the root of their concerns. Answering questions supplied by AI that they don’t fully understand may alienate them further.
The giant societal shift that AI will bring is not from the technology itself, but the accessibility of it.
AI in some form has existed for decades. Never before however has this level of technology been so accessible to the general public.
Where Humans Still Matter Most
AI can do a lot of things, but not everything. In order to maintain your value, you need to be aware of the areas in which you are stronger:
- Critical Thinking
- Reasoning
- Empathy
- Context
- Trust building
Think of AI as an assistant (albeit a very competent one). You wouldn’t delegate your decision making to a junior employee, likewise, you shouldn’t pass off your core accountabilities to AI.
AI can draft, summarise and suggest. It cannot fully understand your client relationship, your company’s appetite for risk, internal politics, or the consequences of a poor decision.
That is where human value remains. The role of the person is to decide what matters, apply context, take responsibility and communicate the decision clearly.
Summary
Where are you at in the journey of AI adoption?
AI is already influencing how businesses make decisions. That is not inherently good or bad. The outcome depends on how intentionally it is used.
Used well, AI can improve ideation, speed up research, highlight risks and strengthen recommendations. Used poorly, it can create false confidence, unnecessary process and weaker critical thinking.
The goal is not to keep AI out of decision making. The goal is to keep humans in it.
So the question is not simply “Are you using AI?”, It’s “Are you using AI to think better, or to avoid thinking?”
We’d love to share our insights with you; simply book a meeting or get in touch directly to organise a time to chat.