Before you hire AI: What every business leader should know

Not long ago, I was chatting with the founder of a retail business who said something that stuck with me: “AI feels like an expensive gym membership; everyone talks about having one, but no one knows what to do with it.”

He was spot on. For a lot of small and mid-sized businesses, Artificial Intelligence (AI) still feels distant. There’s interest, yes. But there’s also confusion, cost concerns, and a big question mark around value. Now, there’s a new kind of AI being talked about: Agentic AI--intelligent, goal-oriented digital “agents” that can take on tasks like gig workers. You just bring in the right agent for the right job, at the right time.

This sounds great, but is it actually usable for small businesses? That’s where things get interesting.

What is agentic AI?

Agentic AI is a step up from basic automation. These agents aren’t just reacting to commands, they’re making decisions, solving problems, and figuring out how to get results with minimal supervision. Think of them as mini-AI employees.

For example, you could have a digital marketing agent that runs ad campaigns or a finance agent that can perform bank reconciliation processes which are usually time consuming and cumbersome. It’s a big idea. But also, it’s still in it’s early days.

Gartner predicts that over 40% of agentic AI projects will be cancelled by 2027 because they’re expensive, hard to scale, or just not delivering enough value. Right now, many of these projects are experiments, driven more by hype than strategy. Organisations need to cut through the hype and make careful, strategic decisions, and that’s especially true for smaller businesses, where every dollar counts.

In fact, a Gartner poll of 3,400+ professionals earlier this year found that only 19% have made significant investments in agentic AI, 42% are exploring cautiously, and the rest are waiting or unsure. It’s clear we’re still in “wait and see” mode.

The real challenges

Let’s talk about the main concern: cost and complexity.

Most hosted AI models charge based on how much data (or “tokens”) you send them. The more data they process, the more you pay. That might sound manageable at first, but costs can add up quickly, especially if you don’t have tight control over what the agent is accessing.

And if you’re thinking about training a model from scratch, that’s a whole different ballgame. It requires expensive hardware, and most small businesses don’t have the infrastructure or the team to handle it.

Then there’s data privacy. To make agents useful, you often need to share internal or customer data. That raises some big questions: Where does that data go? How is it used? Who has access later? These concerns aren’t fully answered yet.

Not all “agents” are the real deal

Here’s another thing to keep in mind: not everything being sold as agentic AI actually is. Everything that uses GenAI or conversational AI is not agentic AI. We see that vendors slap the “agent” label on chatbots or simple automation tools that don’t have true decision-making capabilities. Out of thousands of companies claiming to offer agentic AI, only a handful are real.

So, if you’re exploring this space, look closely at what you’re investing in.

It’s not one-size-fits-all

Agentic AI might sound like a great idea, but that doesn’t mean it works for everyone. Let’s say you want to use it for customer support. In theory, that’s one of the easier use cases. But for the agent to give useful answers, it needs access to all your systems, your CRM, support tickets, product data, and more. That kind of integration isn’t always easy. Some of the tech that enables this is still being built.

Also, small businesses often have unique, local needs that off-the-shelf agents don’t understand. What works for a US-based ecommerce company might not work for a local logistics firm in another part of the world. Bottom line: just because you can hire an AI agent doesn’t mean it will work for your specific needs.

So, where’s the value?

This technology genuinely has huge potential. There’s something real here, but you’ll need to be strategic. Don’t just look at individual tasks. Think bigger. Focus on how agentic AI can improve overall business productivity through speed, cost savings, or better decision-making.

What it means for people

Agentic AI isn’t about taking away jobs. It’s about shifting what people focus on. When agents take over repetitive, manual tasks, humans can spend more time on things like strategy, creativity, and relationships. That’s a good thing for individuals and for businesses.

Agentic AI is exciting. It’s powerful. And it’s already here.

But for small businesses, the key is to move with curiosity and caution. Don’t jump in because it’s trending. Take the time to figure out where it can really help, and whether it’s worth the investment today. Because the smartest move isn’t to be first. It’s to be clear about what you’re solving, and whether an agent is the right one for the job.