My background is in design and technology. I was actually a touring musician in a rock band before teaching myself design and engineering and opening my own software company. So I’ve had a front row seat to every major tech innovation over the last few decades – from mainframe to client-server systems to cloud and, of course, artificial intelligence.
The next Big Bang moment is quietly unfolding behind the scenes — the traditional SaaS ecosystem is dying fairly quickly because of changes wrought by agentic AI. The entire business model of how software is built and deployed is shifting from the subscription model to consumption.
What’s coming is a new way of interacting with knowledge and being able to take action on that knowledge. Until now, internet searches were the primary avenue for gaining information on anything — baking a cake. Changing a tire. Learning a new language. Then, we figure out what to do with this newfound information.
Agentic AI is turning that process completely on its ear. The world that most businesses work in involves enterprise-level software. The world that they’re moving to is personalized agents.
Agentic AI as a sales rep
A friend has a son in college who wants to follow in his dad’s footsteps and go into sales and marketing. The difference between the two generations is technology. The father didn’t necessarily need it to be great at every facet of the work. He had an analyst who conducted market research, a sales development person to do cold calls and a scheduler to manage his calendar. His strength is building trusted relationships, and those other people in their roles helped him focus on that to bring value to the client.
It’s different for the son. Those three people — the market analyst, the cold caller and the interview scheduler —- will be his agentic AI team. His job, unlike his father’s, is to curate those agents to do those things for him. Only then does the son show up to the actual meeting to close the deal.
As the shift to personalized agents comes into focus, the landscape of how products have to be designed will change too.
Add-ons and options
Think of it this way — Software as a Service (SaaS) will now be Service as Software (SaS).
Under the consumption model of buying, you will pay for a service or product and choose what to turn on (for example, when purchasing a car, you won’t choose additional add-on packages like heated seats and other bells and whistles, but rather will decide what to turn on for a fee).
In Human Resources, an organization typically purchases the base software — the Candidate Relationship Management technology — and it is customized later with add-ons such as integrations, datafeeds and other options they have chosen to be turned on or included in the “base” system.
That is all gone with agentic AI. The talent CRM comes with every agent already installed. Users can now go into the admin panel and click on a screening agent or interview scheduling agent. There are additional integrations. Everything is already there. Agents are paid for as they are used.
All of this gets done through an interface built specifically for AI. It acts as a universal translator so different AI tools and apps can share data and work together seamlessly.
So what’s the long and short of all this techie speak? In the next decade, the internet will not be designed for people. It will be designed for agents.
Jarvis is the future
Technology will no longer be thought of as simply a tool to support the business, but one that innovates the business.
With that shift, the traditional role of an engineer will disappear and will be replaced by engineers who understand human interaction and behaviors with technology, since creation of the technology itself will be done by AI and AI agents. The new generation of engineers will be behavioral experts who develop and deploy tools so people can achieve peak productivity as fast as technologically feasible.
Since humans will be the priority, and not the technology, expect to see fewer typed commands as we move away from interfaces entirely. URLs and keyboards will be a thing of the past. Commands will be all voice-driven, heads-up displays and directed by hand motion. Think of Tony Stark and his natural language agent, Jarvis. We, too, will speak to an agent and have it perform a task with a high level of precision, confidence and trust.
Jarvis is already up and running today, but only in single use cases (making a voice call or doing research, for example). In the future, it’ll be one agent doing multiple tasks. For now, Model Context Protocol is the framework that allows all agents to operate and interact with each other seamlessly.
And so that’s why, in the future, the internet is not going to be a place built for humans. It’s going to be like “The Matrix” with the code floating down the screen. It’s going to be doing its own thing, because we’re not meant to integrate with it.
Who benefits from the SaaS shift?
The end user or consumer will benefit the most, ultimately, because they’re only paying for what they use. It’s like paying for electricity — there’s one standard rate each month.
Enterprise software companies may benefit too since building software in this new framework will be scalable to anyone. The risk lies in the up-front cost to build that framework. Under the traditional subscription model, software companies charged clients for that development. The risk now falls on the software companies under the pay-for-what-you-eat consumption model, but if they get it right, it’s infinitely scalable.
It won’t take 10 years to figure out if the software is accepted by the market. In the agentic world, it’s going to happen in a matter of weeks to months, not years.
The new way to shop for software
It’s important for business leaders to be aware of the technological pivot in the software market.
My advice is, understand how to operationalize agentic technology. That includes how to buy it, how to evaluate it and what processes and skills does their team need to employ in order to be successful.
It’s significant because it’s about two key things: access to knowledge and speed to action.
To wrap up, we have the internet at our disposal to learn anything and everything about all sorts of things, but it still takes weeks to months to act on it. Business leaders will soon have that information at their fingertips. Agents will dispense helpful advice about using the data in an organization’s ecosystem. They’re that smart.
The notion that a business will have to engage with a large consulting firm to evaluate market trends is gone. On the flip side, consultants will provide new value by becoming proficient at helping executives transform their organizations using AI agents.
