(This column originally appeared in Forbes)
This past week, and in light of media reports on how larger corporations are pushing back on the higher-than-expected expenses for their AI projects, Google released a number of measures to help their customers better control costs. Among them are pay-as-you-go pricing, monthly spending caps, Flexible Savings Plans and a forthcoming option that can run eligible workloads during off-peak periods at substantially lower prices.
The idea is to get more businesses — particularly smaller and mid-sized companies — to use AI (Gemini in particular) and not be spooked by these recent reports.
Part of Mike Clark’s job is to get more businesses using Gemini and, more specifically, the agents that Google wants them to build on its platform. As the Director of Product Management, Gemini Enterprise Agent Platform — Google Cloud, Clark leads the product strategy and development of Google Cloud’s platform for building, deploying, managing and governing AI agents. In a recent interview I did with him, he described his team as “tip of the spear for agents” who work with Google DeepMind and other Google teams to package agent technology into tools businesses can actually use — covering how agents are built, hosted, run, governed, evaluated and optimized.
It’s good news for Clark that Google is addressing AI cost concerns because, if he’s successful, many companies will be significantly increasing their use of agents to do tasks that were normally being done by humans. The last thing Clark wants is for prospective customers to be scared off from using this technology because they perceive that it will be too expensive. To that end, his biggest task will be educating his business audience.
The Difference Between Agents and Generative AI
And he’ll start with explaining that AI agents are fundamentally different from ordinary generative AI because they can reason through multiple steps and take action. To him, a chatbot responds, but an agent can determine what needs to happen next, gather information, make decisions and potentially execute.
“The capability fully exists today,” he said. “There’s a layer of orchestration that allows a model to think in multiple steps, and that’s what an agent does, it starts to take multiple steps at a time, which unlocks the capabilities of a model to be able to do so much more than when you just talk to it directly. It allows it to take actions and think more.”
Website Preparation Is Critical
According to Cloudflare data cited by Forbes, bots now account for more than half of web traffic. Not all of those bots are AI agents, of course, but AI-driven traffic is growing quickly — which is why it’s critical that, as I’ve written before, businesses prepare for the agentic future now and they need to start with their websites.
Clark says that how businesses prepared their websites for search in 2005 wa very different from how they prepared for mobile in 2012. What drives this is consumer preferences — the same that’s driving the agentic web today.
“The biggest shift in consumer behavior around agents have been people using Gemini, people using ChatGPT, people using Anthropic to go ask those questions, where they might have gone in search traditionally and seen a list of blue links,” he said. Clark also says that this will lead to more simplified sites that “have enough content and context for an agent to know why it would make this decision from a purchase perspective and how to do it.”
Agent Building Is Getting Easier
Building agents is still a developer’s job. And — knowing my clients — it’s going to continue to be that way. It’s not that people running businesses can’t figure out how to do this, but they’ll instead rely on both internal and external resources to accomplish this, so they can focus on running their businesses. However, Clark argues that the technical barrier is rapidly falling.
“We’re watching more and more consumers act like developers,” he said. “I don’t know how to write code, but I can drag boxes together and type in a prompt that I want and build an agent with very simple steps. I don’t have to understand how to write a line of code, but I can now go connect my Salesforce with my Twilio number with other things and start to do things that I would want to do as a small business.”
Agentic Use Cases Are Growing
He’s also seeing agents do the kinds of things that many companies had to outsource. For example, launching an agent to do market research, whether it might be a new product that’s being considered and the markets that would be potential customers.
“I might have taken two weeks as the CEO and gone off and done or the owner of the business and now I can just start it, kick it off 20 minutes later, have the full corpus of information that’s available in a really consumable report with some actionable items and understand where I’m sitting in the market and what are some things I could do,” he said.
Agents can get even more involved in the internal workings of a business.
“I’ve got a friend that runs a machine shop and something that sounds so manual, yet he’s driving everything via agents today on everything from how people are specifying parts and what they’re trying to build all the way through to the CNC machines that are cranking out these physical pieces and devices using simple agentic tools,” Clark said.
Trust Is Still A Significant Factor
Of course, there’s always pushback. And it’s always because of trust. Even Clark says that business shouldn’t just blindly trust AI agents. Instead, trust should be earned incrementally through transparency and human approval. For him “total transparency is everything.” He wants to see every single thing that the model is doing, that the agent is thinking, and that the agent is taking actions on.
He advises business owners to make sure that agents are always asking for permission before taking a significant next step. Ultimately you can get to places that you trust enough because you better understand what it’s doing.
“Think of the consequences if something did go wrong and build your own governance in your own mind and in your own systems around what actions can be taken,” he said. “The models are getting better, the agents are getting better, and that continues to grow.”
Small Businesses Have An Advantage
Agents are a generational shift in business management. They will be enormously disruptive over the next few years. But the good news is that small businesses may have an AI advantage precisely because SMBs generally have fewer legacy systems, bureaucratic layers and entrenched processes. To that end, Clark recommends being experimental, curious and willing to play with AI tools so that you can find the things that can really expand your business.
“As a small business, you don’t have 50 years of legacy business,” he said. “You don’t have regulations in the same way against large businesses. You get to follow that technology curve as close as possible. I’m shocked every day how I see people open up new parts of their small business by things that they’ve unlocked just by using GenAI in new ways.”
