Why agentic AI is the shift businesses can’t ignore

One thing I’ve learned is that it doesn’t take much for expectations to reset. If you’re an iPhone user, you probably disliked the redesign of Photos in iOS 18. You may have even complained about it to friends. But after a week or two, it just became the new normal and you moved forward.

Agentic AI is on the same path. As both consumers and employees, our expectations of how work gets done are about to fundamentally change.

What agentic AI really means

Most consumers’ and businesses’ experience with AI to date has been through generative AI tools such as ChatGPT, Claude and Gemini. These tools are powerful and can create meaningful value on specific, single-threaded tasks. Whether it’s writing code, developing content or creating an image, they perform like an attentive concierge waiting to meet our every need — as long as we initiate the request.

Where agentic AI changes the equation is that these systems possess both agency and action. With agency, the system can plan, adapt and make decisions toward a goal without a user providing step-by-step instructions. With action, it can execute, not just suggest. Thus, agentic moves AI from latent potential that only creates value if it’s explicitly summoned to active potential that’s regularly working toward the achievement of our individual and business objectives.

That’s the leap: the shift from reactive to proactive value creation.

Agentic browsers: The leading edge for consumers

With nothing more than interest and the willingness to pay for a premium subscription, consumers today can experience the future of agentic AI. Early agentic browser offerings like Perplexity’s Comet, Google DeepMind’s Mariner and OpenAI’s ChatGPT Agent (formerly Operator) are driving a paradigm shift from traditional browsers offering access to information to agentic browsers serving as collaborators to complete tasks.

While these nascent offerings are far from perfect, they’ve opened the door to two meaningful shifts. First, consumers will come to expect user experiences that occur seamlessly from beginning to end within a single entry point. Users will lose patience with and actively avoid situations that require them to flip between multiple windows or applications to achieve an outcome.

Second, consumers will seek out technology that not only provides them with information but also takes action and achieves desired outcomes on their behalf. It will become very clear how much of their time is currently wasted on low-value work-in-process tasks, thus decreasing their tolerance for spending time on activities they perceive as unnecessary interim steps to achieving a desired final outcome.

The implications for businesses of this new single-entry-point agentic AI experience are threefold:

1. New user experience expectations. Your employees’ user experience expectations at work will be reset by their positive agentic experiences as consumers.
2. Reorientation of human/AI workflows. The stage will be set to begin offloading lower-value tasks to agents, enabling employees to focus on higher-value work that more directly ties to achieving desired business outcomes.
3. New definition of customers. Businesses will need to expand their view of customers to include individuals, individuals’ personal agents and third-party agents acting on behalf of individuals.

New user experience expectations at work

For many years, most of us have resigned ourselves to a reality full of less-than-ideal user experiences at work. We believed the need to access multiple disparate corporate applications to do our jobs destined us for a disjointed daily experience. This resignation is about to change dramatically.

As usage of agentic browsers proliferates among consumers, these seamless, single-stream interactions will transform individuals’ expectations of what a good user experience looks like. Regardless of whether organizations are ready, employees will rapidly apply these expectations to the applications that drive their experiences at work.

Want someone to work across multiple applications with varied user experiences to reach a specific outcome? How archaic! When employees are empowered to manage their personal lives through a single agentic experience — all while sitting at their desk at work — their patience with the poor experience of today’s disjointed corporate applications will quickly dissipate.

Organizations should begin exploring how agentic capabilities can create a more seamless work experience. In 2026, leading companies will begin transforming some of their key workflows into single-stream experiences that are enriched by voice AI or an agentic browser. This will set a new bar for creating high-impact employee experiences. Google’s October announcement of Gemini Enterprise, which stated “Gemini Enterprise brings the best of Google AI to every employee through an intuitive chat interface that acts as a single front door for AI in the workplace,” reinforces where we’re heading.

Reorientation of human/AI workflows

Let’s be honest — many of the tasks we all perform today as consumers and employees are low-value. They’re a necessary means to an end, but necessary doesn’t equal value-add. With the continuously improving reliability and reach of agentic capabilities, the stage is set to begin offloading lower-value tasks to agents and enabling humans to spend time on higher-value, outcome-focused endeavors.

Two current constraints to more integrated human/AI workflows are existing organizational structures and the cost impact of “always on” agentic solutions. While leading companies have begun the work to envision their new agentic-enabled operating models, the reality is that this is a significant transformation with tentacles that will reach all corners of an organization. Achieving impactful human/AI workflows will necessitate doing the hard work of reinventing leadership philosophies, organizational structures, job architectures, talent strategies and KPIs and measurement systems, just to name a few areas. The lift will be significant, but so is the potential value if executed correctly.

When it comes to the current economic model of scaling agentic AI across an enterprise, the reality is that today’s costs make this very challenging—but language model efficiency is improving at a rapid clip. This is clear in the Q1 2025 state of AI report by Artificial Analysis, where they report on decreasing language model inference prices.

If you’re hesitant to begin the work to enable new integrated human/AI workflows in your organization because the financials don’t quite work yet, remember that you need to be planning for a future cost basis that continues to decrease. The amount of work required to facilitate this truly cross-organization transformation is significant and you won’t incur the full costs of these human/AI workflows for quite some time.

Expansion of the definition of customers

With the expansion of AI capabilities and the single-experience expectation being reinforced by agentic browsers, companies need to extend their definition of customers to include individuals, individuals’ personal agents and third-party agents acting on behalf of individuals. In addition to continuing to improve their traditional customer engagement channels, companies will also need to optimize how they interact with agentic proxies. If your business workflows and technology aren’t designed to provide information to agents in the appropriate structure with necessary expediency, you risk being invisible in the very channels where decisions are increasingly being made.

In late September, OpenAI introduced instant checkout and the Agentic Commerce Protocol, asserting that “this marks the next step in agentic commerce, where ChatGPT doesn’t just help you find what to buy, it also helps you buy it.” With this new capability, purchases occur without leaving a ChatGPT chat. For companies, this means that in addition to maintaining effectiveness in current channels, they now also need to understand and build systems that align with how ChatGPT defines “the most relevant products from across the web” and “ranks (products) purely on relevance to the user.”

This isn’t just retail. Financial services, B2B sales, healthcare and logistics will all see agentic proxies beginning to make decisions, execute tasks and close transactions. Leaders who continue to design only for direct human interaction will miss the growing share of activity that occurs within these new agentic channels.

Managing the transition with clear eyes

None of this means agentic AI is fully mature. The tools can be uneven, but that’s not a reason to hold back. It’s a reason to start experimenting now. Identify where agents already deliver value and offload what you can today, then track how the technology evolves.

Leaders need to face the discomfort that comes with this shift. Many enterprise workflows were designed years ago, sometimes by the very leaders who still defend them. The instinct is to bolt agentic capabilities onto those familiar processes, but that only reinforces outdated assumptions. The harder and more valuable move is to let go of processes you may have personally designed and ask instead: What outcome am I trying to deliver and how should the workflow be redesigned with agents doing the heavy lifting?

The leadership imperative

Agentic AI isn’t just one more application in the stack. It changes who your customers are, how your employees work and where you should place your strategic bets. The smartest leaders I know are already starting to treat AI as a teammate, not a technology, making room for generative AI to accelerate knowledge work and agentic AI to execute on our behalf. Together, they expand the team’s capacity, freeing people to focus on what matters most.

This article is published as part of the Foundry Expert Contributor Network.
Want to join?