Is 2025 the Year of Agent Washing?

If 2023 was the Year of Generative AI and 2024 the Year of Agents, has the promise of Agents already become diluted by hype?

In a discussion yesterday with Rajib Saha , he challenged me, as he often does in our conversations, to think more deeply about the landscape we are pioneering. So, I suggested we bring it to the community for more feedback. He agreed!

In 2023, Generative AI ignited imaginations worldwide, reshaping industries and sparking a wave of innovation. By 2024, as heralded in the 28th Annual ARC Industry Leadership Forum keynote, the baton was passed to Agents—autonomous, decision-making AI systems poised to redefine human-machine collaboration. Yet, as 2025 unfolds, the term "Agent" is everywhere, and everyone seems to claim they have one. From enterprise software vendors to personal productivity tools, the market is saturated with offerings, each branded as the next revolutionary agent.

But is 2025 truly the Year of Agents, or have we stumbled into the murky waters of “Agent Washing”?

Much like "greenwashing" in sustainability, Agent Washing risks obscuring real progress by slapping a shiny label on tools that fail to deliver the transformative promise of true Agentic AI. To navigate this pivotal moment, it’s critical to cut through the hype and clarify what an agent truly is—and what it is not.

What Is an Agent, Really?

An agent is not just software that retrieves information or automates simple tasks. It is much more profound. Should a true agent not emulate human agency—the ability to reason, learn, and act autonomously with purpose?

Consider how humans exercise agency every day. Each of us navigates life with a worldview, a constantly evolving understanding of why and how things happen. We evaluate options, make decisions, and take actions that align with our goals. When things don’t work out as planned, we adjust, iterate, and improve. This iterative process—thinking, acting, learning, and optimizing—is what defines agency.

True agents replicate this cycle on our behalf, orchestrating decisions and actions that align with our goals, preferences, and values. They aren’t just tools; they are digital partners capable of delivering outcomes that traditional systems cannot achieve across the complexity of our lives today both at work and at home.

The Foundation of True Agentic AI

For an AI to qualify as a true agent, it must meet Arthur Kordon 's Axioms:

1. The Ability to Learn An agent must acquire knowledge and improve over time, either through external instruction or self-directed learning.
2. The Ability to Predict It must anticipate outcomes, explore counterfactual scenarios, and leverage its knowledge to assess “what if” possibilities.
3. The Ability to Reason and Decide The agent must apply logical reasoning to make decisions for specific purposes and act autonomously to achieve those objectives.

Without these capabilities, a so-called “agent” is just a glorified assistant—a helper that facilitates tasks but lacks the autonomy, reasoning, and adaptability required to deliver transformative value.

The Human Context of Agency

To understand agency in the digital realm, it helps to first examine what enables human agency. For people to act effectively and achieve their goals, they need an environment that provides:

• Psychological Safety: The freedom to experiment and take risks without fear of undue consequences.
• A Clear Purpose: Direction and meaning that guide actions.
• Social Coordination: The ability to align and collaborate with others.
• Autonomy: The freedom to act independently.
• Access to Resources and Information: Tools and data needed for informed decision-making.
• Recognition and Reward: Incentives that reinforce effective action.

These elements create fertile ground for human agency. But what does this look like for digital agents?

The Digital Agent’s Environment

For Agentic AI to thrive, its software must operate within an enabling environment:

1. Independent Decision-Making Agents must act autonomously, taking initiative without relying on constant user input or pre-programmed rules.
2. Value Beyond Legacy Systems Agents should provide insights and outcomes unattainable by traditional systems, using the same inputs to deliver unique value.
3. Seamless Integration A true agent should complement existing workflows without requiring extensive changes to legacy systems.
4. Collaborative Network Capability Agents must work effectively with other agents, forming dynamic networks that amplify their collective capabilities.

So a Vendor Says to You That They Have an Agent! Ask These Questions

With Agent Washing on the rise, it’s critical to separate genuine agents from imposters. When a vendor claims their product is an agent, ask these questions:

1. Does It Learn Over Time? Can the system acquire knowledge and improve autonomously, or does it rely entirely on pre-programmed rules, and or solely on past evidence?
2. Can It Predict Outcomes and Explore Counterfactuals? Does the agent anticipate outcomes, assess alternative scenarios, and adapt its actions accordingly?
3. Does It Reason and Act Autonomously? Can the agent make context-aware decisions and execute actions aligned with its reasoning?
4. Does It Operate Independently of Legacy Systems? Is the agent interoperable with your existing infrastructure, or does it require a costly overhaul?
5. Can It Deliver Unique Value? Does the agent unlock new capabilities, such as faster innovation cycles or optimized decision-making?
6. Can It Work in Collaborative Networks? Does the agent integrate with others to complete complex tasks, or is it siloed?
7. Is It Transparent and Explainable? Can the system explain its decisions and actions in a way that fosters trust and accountability?

The Risks of Agent Washing

As organizations rush to capitalize on the agent wave, many are mislabeling simple automation tools or enhanced interfaces as agents. This phenomenon creates confusion, dilutes trust, and risks derailing the adoption of true Agentic AI.

For consumers and businesses alike, it’s critical to hold vendors accountable. Without clear standards, the potential of Agentic AI could be buried beneath a flood of hollow promises.

The Stakes of Getting It Right

If Agentic AI fulfills its potential, the benefits will be transformative:

• Accelerated Innovation: Faster decision-making frees humans to focus on creativity and strategy.
• Democratized Expertise: Advanced capabilities become accessible to individuals and organizations worldwide.
• Resilient Systems: Adaptive agents help businesses and societies respond dynamically to change.

But the path forward requires vigilance, discipline, and a shared commitment to building agents that genuinely augment human agency.

From Six Degrees to One: A New Paradigm of Connection

Agentic AI collapses the traditional degrees of separation, creating a world where people, ideas, and resources are seamlessly connected. In this one-degree world:

• Knowledge Becomes Universal: Agents provide instant access to the sum total of human knowledge, tailored to individual needs.
• Connections Become Purposeful: AI prioritizes meaningful relationships, linking people and ideas to achieve specific goals.
• Collaboration Becomes Effortless: Agents facilitate real-time problem-solving across industries, disciplines, and geographies.

This isn’t just a faster network; it’s a smarter, more adaptive one.

A Renaissance in Human Potential

By alleviating cognitive burdens, Agentic AI creates space for individuals to pursue what truly matters:

• Wisdom Through Insight: Agents synthesize knowledge, helping people make better decisions.
• Creativity Unleashed: Humans can focus on innovation while agents handle logistics and execution.
• Stronger Connections: By automating the mundane, agents free us to deepen relationships and build meaningful communities.

The Choice Before Us

As we navigate 2025, we must ask ourselves: Are we building agents that empower, learn, and reason—or settling for tools that merely masquerade as such because someone claims it?

The answer will define not just the future of technology but the future of humanity. Let’s demand more. Let’s ensure that agents fulfill their promise as partners in creating a better, more connected world.

The Year of Agents—or the Year of Agent Washing—rests on this choice. Let’s make it wisely.