The business world is awash with artificial intelligence (AI) talk and AI-enabled products. But how much of this is helpful, and how much is hyperbole that could cause your organization to expect too much from what’s still a fledgling business technology?
I spoke with Alex Valette of MPS Limited, a learning company, about the business reality of AI. Here are some of the key points he made as an AI consumer rather than an “AI expert.”
Understand the AI hype vs. reality
Many AI claims are exaggerated for market appeal, and some investments might not yield the significant gains promised. For example, if an AI application delivers 80% correct answers 80% of the time, it makes it useful but not foolproof.
Another way to look at this is to consider your delivery of an AI-generated report to the CEO of your company. Are you okay that it’s 64% correct? Are you alright with the 36% chance that you’re sending inaccurate data or insights to your CEO? No, probably not. So, question whether your AI applications are ready for high-stakes decision-making. AI is already a great assistant, but it’s risky to trust it blindly in critical scenarios (or even sending automatically generated reports to your CEO).
AI adoption will win out, though. But I believe there’s a parallel to the internet boom and the dot-com bubble of two decades ago – the AI “bubble” will likely burst before becoming deeply integrated into our everyday operations.
Realize the initial cost of successful AI adoption
The budget you need to spend on AI is not money; it’s time. If you consider the common advice given on AI adoption, there’s a heavy emphasis on people, such as:
- Experiment first, don’t just buy in – treat AI as an experiment, not an off-the-shelf product
- Train people before implementation – let employees learn AI’s capabilities and limitations before large-scale rollouts
- Recognize that small, controlled AI experiments work better – start with low-risk trials, measure effectiveness, and scale AI usage based on real results
- Smaller companies should initially avoid consultancy costs; instead, they should train an internal team to experiment with AI’s best use cases across the organization.
So, find the right internal people resources, get your training, and then build a small team to undertake the experimentation. Then start a cycle of small innovations.
This team will come from marketing, sales, operations, etc. They will have the goal based on the organization wanting to use and benefit from AI, and rather than investing money, tell them that they each have five hours every week to invest in AI experimentation.
Leverage the initial AI experimentation
Say there are five people, each with five hours a week to experiment. This small pilot team should build a small AI road map after their initial experimentation – these are the things that will be tried across your organization.
The cost or investment might be considered zero (as the staff are already being paid), but your organization now has a team demonstrating to the rest of the company that something can be done with AI. They also become your organization’s AI evangelists because they have been convinced of AI’s value and business use cases.
And if you have chosen this pilot team well, by picking someone from almost every department, your entire organization can start to grow with the mindset of “Let’s use AI.”
Is AI a job killer?
Right now, I think your job will not be replaced by AI, but it will more likely be replaced by someone who knows AI better or a company that knows AI better. It looks scary, though – with AI, for the first time, I think, in history, you have a machine that looks like a colleague and a colleague with way more knowledge than you, way more memory than you, that can process information millions of times faster than you.
However, there’s a strength you have that’s an advantage over AI – it’s that you know your company. You know your company, and you know your customers. If AI isn’t tailored to your specific company and its products, it will respond with “an average answer.” For example, if a question is asked about a product, your product, the AI will likely give an answer that’s the average of all the products that look like yours based on everything written on the internet for the past X years. It’s very important to think about this before you do anything to resist AI adoption in your organization. You can probably imagine a scenario where your CEO gets little traction internally and spends money on consultancy advice instead.
While many employees will fear that AI will replace their jobs, the reality is that AI will more likely replace people who don’t know how to use AI effectively. AI can assist your organization’s employees but cannot replace human expertise, judgment, and company-specific knowledge.
To hear more about what Alex has to say about AI adoption through experimentation (and other things), you can listen to the podcast here.
