Climbing Out of the AI Trough of Disillusionment

Are we diving head-first into the AI “trough of disillusionment” already? Are we feeling like AI is not living up to its magical promise to fix all business woes? Or is it costing you more time, money, and energy than originally anticipated?

I’ve seen these comments crop up in an increasing number of conversations. While we need to dampen some of the over-hyped marketing around AI, I believe that some of this disillusionment is due to how we are approaching AI projects. It is our actions that are causing some of these disappointments, not the technology.

Let’s examine our missteps and how we can get back on track. There are three root issues to address when we tackle AI initiatives.

1 – Involve The Business

Too often, the “we need to implement AI” mandate comes down from an executive team or via a strategic initiative and lands squarely in the lap of the IT department, accompanied by the assumption that the organization’s technologists will have all the answers.

True, IT can provide insight into AI, summarize features and functions, and even procure, implement, and secure the technology. But, as we all know, just “turning it on” is not the best approach when it comes to deploying new tools to users.

Why This Approach Falls Short

So why are we seeing so many organizations do just that, deploy a generative AI chat tool, and assume that everyone will know how and why to use it? It helps that tools like ChatGPT are prevalent in our personal lives, and we know how to use a relatively simple chat interface.

But we still need to know the “why”, and IT alone cannot answer that question.

Business Context Matters

Business teams need to be given context and training on how to integrate new tools into their day-to-day activities. If you are an IT leader who has received the mandate to implement AI, don’t go at it alone.

Reach out to your counterparts in the business and engage. Which takes us to the second point.

2 – Find Opportunities For Improvement

AI needs to have a reason to exist, a problem to solve, and an opportunity to satisfy. These situations to apply AI abound in any business.

There are opportunities for improving efficiency. There are new customer experiences to build. There are data insights to be had.

Why Asking Users Often Fails

We need to find these “use cases” around us, and only then can we investigate applying technology to improve them. Uncovering use cases requires work and is a bit of an artform.

I hear a lot of organizations ask users, “How are you going to use AI in your work?”, and then they are disappointed to not get many ideas coming back at them.

The reason is this: lack of time.

Ask Different Questions

It takes an employee time to understand what AI has to offer. It takes time to attend an AI workshop to learn about its new capabilities. In the meantime, work is piling up, and any excitement raised in a workshop about a new technology’s capabilities is quickly erased when that individual returns to their desk and needs to dig out of the work that accumulated while they were learning about the great new AI tools.

Even if they had the time, it might be hard for some to apply what they have learned about AI capabilities to their daily work.

So, change the question.

• Ask what employees wish they could improve about their work.
• Ask about the single most frustrating or time-consuming part of their role.
• Ask where teams are being asked to do more work with the same or fewer resources.
• Look for signals such as excessive spreadsheet use or long approval and review email chains.

Take inventory of what you hear and then consider the third point.

3 – Don’t Assume That AI Will Fix Everything

AI technology firms and their pervasive advertising lead us to believe that AI is a magical solution to all business problems. While the technology is remarkable, it is far from being the Swiss army knife that we can apply to all the use cases we just uncovered.

AI Comes In Many Forms

First, we must remember that AI comes in many different flavors, of which generative chatbots are just one.

There are entire suites of tools like Azure AI Services, found in the Foundry these days. Here, AI can apply vision to analyze images, speech to communicate with customers, and document intelligence to process through mountains of files.

Throw in some computer use to modernize robotic process automation, and you suddenly have a whole toolbelt of AI capabilities at your disposal.

Do Not Forget Automation And Apps

Let’s remember AI’s close cousins, automation and applications. While these technologies may be feeling increasingly passé, they still play an important part in the IT solutions stack.

Sometimes, a well-written Power Automate workflow paired with a pragmatic Power Apps screen or two solves the problem at hand. Now that AI is integrated into the Power Platform AI Builder tool, creating these artifacts is faster and easier than ever before.

Yes, you read that right. If we don’t use AI to solve the use case, we can use it to create the tools that will.

Climbing Out Of The Trough

None of these three points are very novel, are they? They are tried and true approaches that have served us well over the years.

Work with others, define the need, and apply the right technology.

For whatever reason, recent AI initiatives have left these best practices behind. Well, it’s time we worked together to correct that oversight and climb out of this trough of disillusionment together.

Still struggling to find use cases? Unsure of which technologies to use to address a need?

Reach out if you want to explore these concepts more, or if you want to obtain a free set of tools that will help you uncover your AI use cases. I’m here to help.

And, as always, thanks for reading.

Turn AI Momentum Into Measurable Outcomes

If you are struggling to identify where AI can actually improve work, our AI Use Case Discovery & Roadmap helps uncover high-impact opportunities and match them to the right technologies.