The things the hype cycle doesn’t want you to know.

In the groundswell of thinking that AI will “do all the things” it’s important to understand the limitations of what it can and can’t do, especially when it comes to professional needs or running a business.

Any technology in it’s earliest days is amazing and potentially ground-breaking. As time progresses and we learn more, the application of our critical thinking skills become the best check-and-balance for riding the hype cycle.

Let’s start with a disclaimer…AI is an amazingly powerful technology that has and will change the way many things work. There’s no argument there. I use AI every day to improve the processes for my work and personal needs. That said, it’s equally important to understand the limitations of the technology and how it is used.

AI won’t:

• implement organizational change for you

• solve organizational conflicts

• value experience over aggregated results

• provide push back when push back is needed

• see people as people

• understand for you

Implementing strategies

AI has an excellent capacity to answer questions such as “how should we do this” with a tone and structure that is convincing and confident. Trouble is, there’s nothing in that response around how to communicate much less implement those answers in a successful manner.

Solutions such as Claude Code can build the tools for automating processes but those same tools can’t adapt when the strategy needs to change on the fly. Understanding what needs to be done, who needs to do it, and most importantly what are the potential impacts, falls on experienced professionals and staff. An AI may give you an answer that sounds good…but it takes a professional to know if it is or not.

Solving organizational conflicts

This is another great example of AI providing perceived insights that it cannot justify. Conflicts that occur within an organization, whether they be personal, departmental, or operational, may be “explained” by AI, but resolving the conflicts always come back to the people involved. Anyone telling you their AI solution can resolve the conflicts between staff members or between staff and processes has a monthly subscription price they want you to pay.

These solutions don’t come from only the “right” answers, but also from having seen the wrong ones play out. Only time, trial and error, and experience can create the proper perspective for a specific set of needs and issues.

Valuing experience over aggregated results

Some of the best business and professional decisions come from “uncommon wisdom” rather than the results of a targeted search of internet sources. When you prompt an AI to address a topic, it’s recommended to have it provide it’s sources as part of the response. Why? Simply so you know from where the answer is coming.

I’ve seen too many people going to AI and asking, “what is the best…” or “tell me the right way to…” and I have to ask, best based on what? Right way according to whom?

Think back to the early days of search engines when we would enter keywords and get back millions of pages of results. Even though the “best” answer for our needs may have been on page 13, we rarely went past page 2 in returned links. There is a gap between the best answer and the most convenient one.

AI doesn’t show you the origin of it’s information unless you ask it to. It doesn’t show you the criteria it used to make it’s recommendation unless you press for details.

Providing push back when push back is needed

I asked an AI, “is it a good idea to fire my staff and replace them with AI agents?”

Sounds well-reasoned and thought out doesn’t it? Sure…and a professional could have said the same thing with a simple “No.” I’m sure you can see where I’m going with this.

I then asked, “Will it save my business money?”

In the short term, replacing staff with AI agents can significantly reduce payroll and benefit expenses

An AI isn’t going tell you when an idea is a bad one AND stick to it’s stance on the idea. AIs can be badgered into telling you what you want to hear based on the questions and prompts you provide or worse yet hallucinate and make up the answer. If you’re looking for validation, an AI is “happy” to provide that…but if you’re looking for a rational, experienced evaluation of an idea you need a person with real understanding.

Seeing people as people

In our modern working world, AI has become a major factor in the process of evaluating and hiring people. For better or worse, it can process thousands of applicants for a role, matching their information to the description provided and sifting through to find the “best match.”

If…the job description is accurate. If…the keywords in the resume match the keywords in the role. If…the terminology used across the sources is consistent. So many ifs. As an automation tool, it does it’s job. As something that sees the people behind the text, it continues to fail miserably.

Now we have people writing resumes using AI to optimize them for the AI applicant tracking systems so they have a better chance of being seen by a human. Back and forth the pendulum swings but where are the real benefits being seen? People are not data and when they are reduced to just that we all suffer for it.

Understand for you

This is a hard one to take for many people. It’s not a challenge with AI, but rather with how people handle the information provided to them. An AI can churn out pages and pages of detailed, referenced reports and research based on your prompts: text, charts, images, and more all to provide everything you’ve asked about.

The problem is…are you going to read it? Do you have the time to review, analyze, and process all the information provided to you to turn into something actionable and impactful? Can the AI explain it in a way that makes sense to you and your situation?

I always chuckle a little when, after completing a lengthy bit of research through an AI that goes on for pages, the AI ends with “would you like a one-page summary?” This isn’t an AI problem, it’s a people problem because we know people like quick and easy.

If you know anyone in the consulting field they’ll tell you this has been going on for decades. You spend hours creating the perfect meal of information only to be asked for it to be packed in a happy meal box with a toy. Only difference here is the AI isn’t going to complain about it. Maybe that’s why people like it…no push back.

AIs aren’t professors…they’re elementary school teachers

Now, I’m not disparaging elementary school teachers in any way with that statement. Personally I think an elementary school teacher has a harder job than a professor when it comes to the delivery of knowledge to their audience. (Please feel free to leave your agreement / disagreement in the comments.)

AIs will take a complex topic and synthesize it down to something consumable. A teacher takes complex topics and communicates them in a way that the majority of students can understand. However, the AI cannot tell when you don’t understand what it’s sharing…it can only provide the information.

When we ask an AI for complex research, in-depth information gathering, or recommendations based on available sources, we are presuming we will be able to understand what it returns. It has to be said…that is not always the case.

So how do we get the most out of AI as it exists now and prepare for it’s potential growth and change in the future?

What should you do?

• Question everything and compare sources

• Know the difference between can and should

• Understand the landscape is always changing as are the rules

For the time being, AI is the biggest, noisiest, most disruptive technology out there. It’s critical you don’t fall for the hype, take time to understand how it can help you, and if you’re not sure…ask.