The simpler way to get your team to use AI

There are two ways to use AI to improve your team's effectiveness: one simple and one complex.

The complex approach involves developing or purchasing specialized AI tools that integrate with existing databases and processes. This path typically requires significant investment, extensive compliance reviews, and coordination across multiple stakeholders.

The second approach consists of giving your employees access to general-purpose AI assistants (such as ChatGPT, Copilot, Gemini, or Claude) and training them to use them to improve their personal effectiveness. This method yields substantial productivity gains for knowledge workers with minimal infrastructure changes (we will see examples shortly).

While these two approaches complement each other, many organizations overlook the second one, missing huge effectiveness gains. This hesitation often stems from uncertainty about practical applications and concerns about adoption and potential errors or misuse. This short article explains how to deploy this simpler approach while avoiding most of these concerns.

Practical Applications for Immediate Impact

Let’s begin by seeing a few concrete ways in which AI assistants can help you and your team (we will cover implementation later; for now, let’s focus on the possibilities):

• Reducing stupid mistakes. AI can review communications not just for typos and similar errors but also for potential misunderstandings, likely objections, and missed action items and deadlines.

• Pre-empting failure. AI assistants excel at pre-mortems: ask the AI to imagine scenarios in which a meeting, sales call, or project might go wrong and suggest countermeasures.

• Improving soft skills. Writing clear emails is a core skill for every knowledge worker, yet it is seldom trained. Same for compiling reports, preparing proposals, and drafting presentations. AIs aren't yet good enough to complete these tasks autonomously, but they can provide feedback and opportunities for improvement.

Before we continue with a few more examples, let me highlight a key pattern. In all the examples above, AI is used not to replace a worker or automate a task but rather to help workers be more efficient, effective, or reliable – and to improve their skills. In 2025, I believe this is where the lowest-hanging fruits lie.

Let’s see a few more high-value applications:

• Decision Support: Help analyze options against criteria, identify blind spots, and expand consideration of variables and consequences.

• Process Documentation: Creating and updating standard operating procedures, training materials, and process documentation.

• Personalized Development: Generating customized learning paths, practice scenarios, and feedback – basically functioning as an on-demand professional coach.

The above are just some of the many possible use cases. Soon, more will become viable as AI assistants add capabilities and reduce error rates. As Tyler Cowen puts it, "This is the worst they will ever be."

Fighting resistance to change and getting AI adopted by your team

Here are a few common mistakes I see managers doing:

• Abstract communication. Instead of presenting AI as an all-powerful tool with generic benefits, focus on specific, immediately relevant use cases customized to your team's needs. Explain how AI can be helpful to them, not to their manager or organization.

• Adversarial framing. If you say or imply, “I, the manager, need you to abandon doing things your way and adopt this new method,” you inadvertently frame the conversation as you-vs-me, causing defensiveness. A better approach is to frame the conversation as “you and me against an external enemy,” which could be excessive workload, the competitors getting ahead, etc.

• Trying to change too much at once. If you try to get your people to use AI for too many tasks, you won’t be able to cover one well enough to convince your people it can be used successfully. Much better to focus on them adopting AI for a single use case and ensuring that it works so well that they will be eager to use it more.

Rules and Personal Judgement

Rules and procedures are useful, but, especially when it comes to AI, you also need to cultivate your people’s personal judgment so that they are able to spot eventual errors or navigate unforeseen scenarios. Thankfully, doing so is much easier than commonly thought – if you train your people not (just) by giving them procedures but also by running hypotheticals.

Hypotheticals

The fastest way to build experience and expertise in your team is to schedule a 30- to 60-minute session in which you do the following:

1. List 5-10 common and uncommon scenarios, problems, and dilemmas your team might encounter while performing a given task. (To save time, prepare this before the session.)

2. Mention the first scenario and ask your team what they would do in that situation.

3. Provide feedback, such as “Good idea, because…” or “That doesn’t consider that…” (It’s critical that, both for positive and negative feedback, you explain your reasoning thoroughly.)

4. Move to the next scenario, problem, or dilemma.

This exercise enables you to provide your team with months of experience in just a few minutes. Repeat it as long as necessary.

Hypotheticals are also a great way to prepare your team to identify eventual AI errors or hallucinations.

The importance of the first impression

There’s nothing worse than giving your team access to a powerful tool without explaining how to use it effectively: they will try it once, it won’t go particularly well, and they will lose interest.

Instead, it is paramount that their first experience goes smoothly and yields useful outcomes. Do your homework and make sure you pick a relevant use case for them to use the AI tool. Give them a few examples, show them a few good prompts, and coach them through their first use of AI. Ensuring they have a great first experience is key for effective and engaged adoption.

Conclusions

Of course, there is much more to say and do on how to identify tasks where AI assistants can help, how to get your people to use them, and how to train them to spot eventual AI errors and hallucinations. I plan to write more about it in the future.

For the moment, let me repeat some core principles:

• In addition to looking for complex and structural ways to integrate AI into your organization’s processes, also look for simple ways in which individuals can get AI assistance on their tasks and skills development.

• When introducing AI, do not discuss AI tools generically but lead with specific use cases.

• When training your team, after you give a couple of examples on how to use AI assistance, use hypotheticals to provide them with experience and increase their personal judgment.

Training your team on using AI assistants is important not just for the benefits it can yield today but also because it prepares you and your team with the skills and experience required to leverage the inevitable improvements in AI tools that will become available soon.

Want to discuss AI adoption further?

If you are a leader, manager, or supervisor, and you either already adopted AI tools or are looking to, I’d love to chat with you. You can reply to this email or use ​this link​ to schedule a short call.