Can Agentic AI Systems Create Value for Companies? Probably Not
Chapter 8: People Want Genius Experiences that Give Them Superpowers
The truth that leaders must face regarding agentic AI systems is not that they will replace employees; it’s that the AI agents will replace business models. Because the customer will have the same tools.
An agentic AI system is like an AI that sits on top of other gen AI and machine learning tools, and which, one hopes, will give people superpowers. According to Google’s AI, an agent AI has at least these seven characteristics.
1. Autonomy: They can perform tasks independently and make decisions without direct human intervention to achieve pre-determined goals.
2. Perception: AI agents can sense their environment through various interfaces (e.g., sensors or software) and collect data.
3. Reasoning and Planning: They demonstrate reasoning and planning capabilities, allowing them to devise strategies and make decisions.
4. Learning and Adaptation: AI agents can improve their performance over time through self-learning and adapt to new situations.
5. Memory: They have memory, enabling them to retain information and learn from past experiences.
6. Tool Utilization: AI agents can interact with external tools and services, such as APIs or databases, to retrieve information and execute actions.
7. Multimodality: Modern AI agents, leveraging generative AI and foundation models, can process and understand diverse types of information, including text, voice, video, audio, and code.
(By the way, Google’s AI was simply summarizing Mark Purdy’s HBR article on agentic AI.)
Joe Pine and I wrote two years ago in HBR about Genius platforms, an article which anticipates and clarifies what AI agents should be doing. My purpose in this article is to discuss value creation in the intelligent era and the role AI agents will play in both creating value and destroying value for companies.
Value creation, to the business strategist, stems from the business model. Value creation, to the marketing strategist, is embedded in the brand. Value creation for the experience strategist is built through meaningful engagement between the customer and the company’s solution. It’s about time well spent.
While both the business strategist and the marketer pay attention to the customer, their forms of strategy only work if the customer finds value from an experience.
Right now, 2025, we are in the beginning stages of the intelligent era. And strategists are completely focused on how to leverage AI to gain more value through their current business models—or how to use AI to get customers to buy. Which is fine, I guess. If you think that the purpose of intelligent technology is to shore up your business model or build your brand.
This is how Google AI describes the uses of AI agents. Let’s see if you see the glaring omission from this list:
AI agents have a wide range of applications across industries, including:
Customer Service: Chatbots and virtual assistants for handling inquiries and support.
Healthcare: Assisting in diagnostics, treatment planning, and drug discovery.
Finance: Fraud detection, personalized financial advice, and automated trading.
Manufacturing: Predictive maintenance, quality control, and robotic automation.
Autonomous Vehicles: Self-driving cars and other autonomous transportation.
E-commerce and Retail: Personalized recommendations, inventory management, and price optimization.
To the experience strategist, the omission is jaw dropping. All of the ‘uses’ of AI agents focus on how businesses are going to deploy these agents. Nowhere does the chat tool describe how the customer is going to use AI agents.
And here’s the real burn. AI agents will destroy business models well before companies can capture gains from AI agents. Why? Well first, because that’s what disruptive technologies do. AI is a disruptive technology—not a business model. Or even a business strategy. Second, AI will be deployed broadly and minimal cost to the customer. You will still not be able to chart for customer service, e-commerce, retail, diagnostics, fraud detection, and quality control. You might get incrementally better at these things but so will every other company.
And the third reason is that consumers/customers/people will have access to AI agents. Why would they even bother with your customer support when they can ask their own personal chatbot? Besides, they will want so much more than customer support from the solutions they hire.
Here’s what we, at Stone Mantel, know. AI dramatically changes customers’ expectations, hopes, and desires—so much so that customers (who have never cared about a business model) will no longer see value in most smart solutions today. This is fact. This is your new reality.
The intelligent era isn’t about business model efficiency. It’s not about brand building. It’s about customers’ expectations. Companies who keep their focus on what customers will want in the near future will survive. But most won’t do either. They won’t keep their focus on customer expectations and they won’t survive.
When a tool becomes so powerful that people ascribe intelligence to it, they expect it to do far more things for them than what most companies are used to doing. Business models and brands are built around business categories: healthcare, banking, travel, etc. Those boundaries started eroding with mobile apps and will become transparent with AI agents. We no longer blink an eye at the fact that the same company that delivers your toothpaste to your door produces movies and has the largest pharmacy service in the world. (Do I even need to mention the company’s name?)
With AI agents the customer will behave in ways that are akin to how rock stars behave. The U2 lead singer, Bono, once stated how grateful he was for his personal assistant. She laid out his clothes for him. She paid his bills. She scheduled his healthcare. And most of the time, she did these things without him requesting them, freeing him up to do other things that to him had more value for time. Bono once said that without her he would start smelling really bad, really quickly because he’d just put on the same clothes he took off and he doesn’t always remember to bathe.
Your company’s business model is going to have to convince a rock star robot assistant that your solution matters. Unless it’s actually worth your customer’s time to interact with you (who on occasion will think that she is indeed a rock star), you probably won’t be talking with her. So, if you want to actually interface with real people, you’d better figure out what they are willing to spend their time doing.
Stupid, Dumb, Smart, Genius
Please don’t misunderstand me. I don’t think AI agents are a panacea for customers. Nor do customers. There’s always a big difference between what is promised and what is reality.
In fact, we are already seeing distinctions being made between the quality of intelligence in solutions. This is something that Joe and I predicted almost ten years ago. We call the phenomenon ‘stupid, dumb, smart, genius.’ And you should be thinking about how your solutions fitting into one of these four categories.
Stupid – a stupid solution that has intelligence but uses it poorly. Every day people get answers from gen AI tools that are wrong. The tech industry likes to call these ‘hallucinations,’ which sounds somewhat mysterious and harmless. The better term is ‘stupid’ and it’s the term that customers use. When Joe and I wrote about this in HBR, they made us change the term to ‘not smart.’ You will see that some of our frameworks use that term because some take offense to calling technology stupid.
Dumb – a dumb solution has no intelligence. And for some things that perfect. Before we had smart tools, everything was dumb. We just didn’t know it. A telephone was a telephone. A key started a car. Headphones allowed you listen to music. But there was no expectation that the tool collected data and used that data to customize the experience.
As we move into the intelligent era, there will be backlash to intelligent things. Some people, sometimes, will want dumb tools and dumb experiences. They will want to disconnect from data sharing and AI. But this will likely be the exception, not the rule.
Smart – A smart solution is a solution that has some level of data collection and can customize the experience for the customer. We are used to smart solutions today. Once a solution becomes smart we expect it to do more, get more jobs done for us. A smart watch does more than tell time (obviously!). A smart car does more than a dumb car. In the intelligent era, smart is common place and almost a base offering.
Genius – A genius solution can do a multitude of things for customers. Unlike a smart solution, a genius solution understands the customer’s situation and can adjust to the situation. It can anticipate needs. It makes the customer feel like they have a superpower.
AI agents are clearly being designed to be genius solutions. That’s the promise of AI. The reality may look more like a mix of stupid, smart, and genius. AI agents will do things that are stupid and the industry will say that the customer must train the agent to be smarter. And maybe that’s the new fatigue: constantly having to train your AI agent to do better. The customer is familiar with this kind of fatigue already, having tried to train Netflix, social media, and even their email software to do a better job of supporting them.
On the other hand, the customer—not the company—will have the most power because of AI. If the company is to stay in business—and I’m talking about any business—then they must find ways to get real people to want to spend time with them. And to do that, they need experience strategists. Not business strategists. Not marketers. Experience strategists.
To be continued …