The (alleged) promises of Workplace AI

Brian Madden,Workplace AI Consultant & Analyst

This page was last updated on July 2, 2024.

Workplace technology is always evolving, and we’ve seen lots of tech which promised “revolutionary change” come and go over the years. (personal computers, knowledge management, big data, digital transformation, remote/work-from-anywhere...) So here we are in 2024 hearing about how “AI” is the next big thing, and how it’s going to change everything about the workplace, and how this time it’s different. Barf!

As you know by now, much of this strategy guide is about debunking the hype, dialing down the rhetoric, and figuring out what’s actually real and worth paying attention to. So I think we’re to the point where we have to put our BS detectors on and actually take a look at the alleged promises of AI for the workplace, as told by the people who are selling AI for the workplace.

We’ll dissect the marketing claims, analyze common beliefs about AI’s value, and explore specific examples from those selling and promoting AI. My goal isn’t to dismiss AI’s potential outright, but to unpack the sensational claims a bit to understand which parts are real and which are “hallucinations”. ;)

In this chapter, I’m going to dig into the alleged promises of workplace AI. (And, to be clear, since this strategy guide is just about workplace AI, not every type of AI ever invented, I’m just looking at the alleged promises of employee-facing workplace AI.)

Isn’t this guide part of the problem?

It’s understandable to think that the work I’m doing, and in fact this entire workplace AI strategy guide, is part of the AI hype machine. Heck, I’ve already written many times that it’s important for you to create a workplace AI strategy ASAP!

The difference is that I’m not selling AI products, or trying to convince you that you need to implement AI at work. Rather I’m saying that you need to come up with a plan, or strategy, to address the use of AI by the employees of your workplace, since they’re going to start using AI tools (or be tempted to) regardless of whether you create a strategy or not. So the prudent thing for every workplace manager to do is at least spend a couple hours thinking about the impact to your workplace and whether or not you should take action to address it.

In fact this work will be important even if (when?) the AI hype bubble bursts. If/when that happens, the values of AI companies will go down, countless articles will be written about how “AI was a dud”, and people will think, “Well, that was weird, let’s get back to work now.”

However, in a future post-AI bubble, the existing tools like ChatGPT, Microsoft Copilot, and others will continue to exist, and your employees will continue to use them. Even if all AI forward advancement stopped today, there are still enough AI capabilities available to employees that it will still be important for every company to formulate a plan for dealing with them.

(Also, frankly, huge companies like Microsoft, Google, Amazon, Apple, and Meta are not going to give up on their AI ambitions, even if the investor market for AI startup funding dries up completely.)

So you need to have a plan for the employee use of AI regardless of where the market or hype cycle ends up.

Nonsense vendor promises about AI

I originally got the idea for writing this section when I attended the World AI Festival in Cannes in January 2024. (I wrote a longer article on LinkedIn with a full review of that event.) This is where I first noticed vendor presentations that crossed into the “marketing nonsense” category.

For example, check out this slide:

At first glance, you’ll think, “Yeah? So what, this all sounds great!”

And it does! It does all sound great. But also, what? What are they actually selling?

• Use GenAI to implement business processes which were previously impossible / unthinkable!

Which business processes? How does GenAI help? How do I implement this?

• Top 3 impact areas: (1) Providing productivity gains, (2) Streamlining processes, (3) Achieving cost savings

Again, sounds good. But.. what? What gains? What processes? What cost savings? How do I get those? What do I do?

The slides that followed suggested that you should buy some of this vendor’s big servers with lots of GPUs, download some AI open source models, and then “have a play” to understand how they work.

I was stunned! This is total nonsense, in terms of reasons actual companies should adopt AI? (Buy it first, then play with it to see what it can do? What planet is this guy from?!?)

A few months later, in March 2024, I attended the Microsoft AI Summit in Paris. I took the photo below because so many people in the audience were taking photos of this slide. “Wow, there must be a lot of great content here!” But take a closer look, and actually read the slide:

Is anything on the slide true? Refreshingly yes. In fact it’s all true!

• Think Big! Start Small!

• There is no one size fits all

• It will create value, but not everywhere

But think about what you’re reading on this slide. While it’s all totally true, it’s all total nonsense. Nothing more than obvious platitudes! Literally everything on this slide applies to every IT project of the last 30 years. There’s not even anything specific to GenAI on this slide.

I was stunned. I wanted to ask the people taking photos of the slide why they were doing that. This slide is worthless.

While nonsensical vendor marketing has been a staple of IT forever, it’s especially ramped up when technologies are trendy and hot, and AI is about as trendy and hot as it gets! So let’s take a quick tour of the marketing promises of several of the popular workplace AI solutions on the market.

All of these quotes were retrieved on June 19, 2024.

Microsoft

• “Work more productively, boost efficiency, and improve business outcomes with Microsoft Copilot for Microsoft 365.”
• “Copilot in Teams will help you run more effective meetings, get up to speed on the conversation, organize key discussion points”
• “Copilot in PowerPoint will turn existing Word documents into presentations, create new presentations from simple prompts or outlines, or find the best places for impactful imagery—and help you leave audiences dazzled with skillful storytelling”
• “Enhance every role with AI”
• “Transform productivity for everyone—from sales and customer service representatives to IT pros and devs.”
• “Empower sellers to save time, make meaningful connections, and personalize customer engagements.”
• “Helps employees focus on the work that matters most”
• “Use it to attend extra meetings”
• “AI continues to transform work” [so that it’s] “more human, not less”

Google

• “Once-in-a-generation moment where the latest advancements in AI are reinventing what phones can do”
• “Uses generative AI to help you be more creative and productive”

OpenAI ChatGPT / ChatGPT Enterprise

• “Get answers. Find inspiration. Be more productive.”
• “Redefine work in the age of AI”
• “Find new insights. Increase productivity.”

Perplexity

• “Imagine what your team can do with all this power.”

General beliefs and platitudes about the value of AI

Taken together, I repeatedly hear the following general statements about AI at work:

• Increase productivity: The promise that AI will allow employees to accomplish more in less time.
• Unlock hidden insights: The idea that AI can analyze huge amounts of unstructured data in ways humans can’t, revealing valuable patterns and connections.
• Automate boring & repetitive tasks: The claim that AI will free up employees from mundane work, letting them focus on more creative or strategic tasks.
• Remove bias from complicated decisions: The assertion that AI can make more objective decisions than humans, as it’s free from personal bias.
• Enhance creativity: The notion that AI can spark new ideas and approaches, boosting innovation.
• Improve decision-making: The belief that AI-driven analytics will lead to better, data-informed choices at all levels of the organization.
• Personalize experiences: The promise that AI will tailor interactions, whether for employees or customers, to individual preferences and needs.

Specific examples

And I often hear more specific examples where people say AI makes sense in the workplace, including:

• Help with writing / editing / rewriting text
• Personalized learning
• Brainstorming
• Summarize things
• Take notes / transcribe meetings
• Image generation
• Create & debug code
• Analyze data & create charts
• Ask questions about an image
• Create presentations
• Write personalized marketing copy
• Chat with PDFs, videos, or other long documents
• Better customer service chatbots
• Analyze security logs, customer data, or any other huge datasets

Conclusion? Yes, please!

The promises of workplace AI are awesome. What business wouldn’t want this? From boosting productivity to automating boring tasks to unlocking previously unknown potential, workplace AI vendors paint a fantastical picture of a revolutionary workplace transformation. (And again, some of these promises hold real potential as we’ll explore in future chapters).

But it’s crucial to approach these claims with a critical eye. The gap between marketing hype and actual reality can be huge, and what looks awesome in a controlled demo may not translate into a real-world workplace.

Again, my skepticism isn’t to suggest that AI doesn’t have a place in the modern workplace. It’s just that successfully implementing it requires more than buying into grand promises and deploying a couple new apps one weekend. Successful implementation of workplace AI requires a sober assessment of your organization’s needs, a realistic understanding of AI’s current capabilities, and a strategic approach to integration.

In the next chapter, we’ll look at how these promises map to reality, examining where AI truly shines in the workplace and where it falls short of the hype. By understanding both the potential and the limitations of workplace AI, you’ll be positioned to benefit from its use where it makes sense, while avoiding the pitfalls of inflated expectations.

What can AI realistically do today?

AI Guide content update feed