Newsletter #40: "AI Agent Washing" + 3 Things Enterprise AI Leaders Should Go All In On

If you feel like everywhere you turn, you see messages about “AI agent this” and “AI agent that” and are becoming increasingly concerned about who has the goods vs not so much…

According to Gartner , you’re spot on.

“Over 40% of agentic AI projects will be canceled by the end of 2027, due to escalating costs, unclear business value or inadequate risk controls” - gets the headlines but it’s the excerpt below that encapsulates the challenge / opportunity space.

“Many vendors are contributing to the hype by engaging in ‘agent washing’ – the rebranding of existing products, such as AI assistants, robotic process automation (RPA) and chatbots, without substantial agentic capabilities."

"Gartner estimates only about 130 of the thousands of agentic AI vendors are real.”

“‘Most agentic AI propositions lack significant value or return on investment (ROI), as current models don’t have the maturity and agency to autonomously achieve complex business goals or follow nuanced instructions over time,’ said Anushree Verma . ‘Many use cases positioned as agentic today don’t require agentic implementations.’”

“Gartner predicts at least 15% of day-to-day work decisions will be made autonomously through agentic AI by 2028, up from 0% in 2024."

"In addition, 33% of enterprise software applications will include agentic AI by 2028, up from less than 1% in 2024.”

Needless to say but there are a gazillion reasons the market is moving in the direction of AI agents, a lot of which ladder up to the clear and massive business value intelligent systems aka the next generation, LLM enabled Operating Systems that enable AI Agents will generate.

For example, within the GTM space, check out this quick LinkedIn carousel I published on “Becoming an AI-First Company” which brings to life what an ontology aware intelligent system looks like.

A system I believe all enterprises will ultimately embrace.

The goal for newsletter #40 is to cut through the “agent washing” and describe the 3 critical drivers of AI agent business value generation and impact:

1: Data Architecture is the AI agent foundation. Make sure you build on bedrock, not sand.

2: AI Agents reflect a fundamental architectural shift, not an add-on or tool. Think system, intelligent system.

3: Competitive moats are being built by those building multi-agent orchestration over a knowledge-centric architecture.

As always, I’ll do my best to distinguish signal vs noise.

Here we go.

What is an AI Agent?

AI agents are autonomous systems designed to operate with autonomy to pursue predefined goals and complete complex, multi-step tasks on behalf of users.

They proactively initiate and perform actions within predefined operational boundaries without requiring continuous human input.

They can interpret high-level goals, create action plans and adjust their approach and create new workflows based on environmental changes, initial results etc as they read and react.

They maintain context and improve performance through their memory, recalling past interactions, driving pattern recognition, application of best practices etc.

They effectively operate in the “real world” of enterprise operations by taking actions within various business systems. 

Operating as “digital labor” within an organization, AI agents signify a fundamental transformation in human + machine interaction.

Shifting from command-and-control to a model of goal-oriented delegation and supervision. Redefining the nature of work, moving the human role from task execution to strategic orchestration, oversight and direction.

1: Data Architecture is the AI agent foundation. Make sure you build on bedrock, not sand.

“Poor data sabotages AI and strategy despite significant investments…The organizations making real progress treat data quality not as a one-off fix but as an operational discipline that cuts across systems, teams and decisions.” - Data’s Dark Secret: Why Poor Quality Cripples AI and Growth by Vipin Jain , CIO Online 04.08.25

There is no AI strategy without a Data strategy.

AI agents represent a structural shift in how applications operate, decide and adapt.

AI agents are fundamentally reshaping the enterprise technology landscape as they enable direct interaction with data, dynamically handling business logic and integrating historically siloed data sets and operations. 

Across numerous use cases, instead of manually logging into various systems of record where one data set doesn’t talk to the other data sets, evolving from an application-centric architecture to a data-centric architecture enables AI agents to unlock productivity and efficiency gains that were previously impossible and unimaginable.

It’s mission critical, especially for advanced use cases involving cross-functional insights, for AI agents to be able to access and synthesize data from disparate, often siloed, enterprise systems.

Without this, an AI agent is severely limited in its ability to drive holistic problem-solving, dynamic decision-making and take actions to generate outcomes that realize their outstanding potential.

Those that make the investments in modernizing their Data Architecture today will be building a 24/7, infinitely scalable digital workforce on bedrock.

Those that don’t, won’t. The consequences will be severe.

2: AI Agents reflect a fundamental architectural shift, not an add-on or tool. Think system, intelligent system.

“Agentic AI marks a transformative leap in artificial intelligence. Moving well beyond GenAI, these autonomous software agents can make decisions, adapt, and execute complex tasks with minimal human input, enabling AI task automation at scale."

"These systems are reshaping sectors like software development, legal services, marketing, and customer support. Leading tech firms are rapidly investing in the next-gen AI wave."

"But true impact requires robust governance, infrastructure, and reskilling."

"Agentic AI is not just innovation—it is a redefinition of modern work and a critical step toward a more autonomous, AI-powered future.” - How Agentic AI is transforming the future of intelligent systems by Hari Balaji EY 06.30.25

The “magic” of AI agents requires substantial and often unglamorous investment in data quality, resilient and well designed infrastructure to move from POC to reliable, production-grade solutions.

A lot of immediate value is found in automating tedious, repetitive work that augment human capabilities by automating tasks or research that were previously too expensive or time-consuming for humans or rule-based systems to undertake.

Creating immediate value, providing an enterprise with tangible performance data to increase confidence and ultimately, laying the foundation for the architecture that competitive moats can be built upon.

The fundamental architecture shift requires robust engineering, comprehensive data governance and modernizing data and orchestration infrastructure.

Keeping in mind that AI agents, based on LLMs, exhibit probabilistic behavior, meaning their outputs are not always consistent for the same input which is quite different than the deterministic nature of traditional enterprise software. 

Requiring new approaches to testing, monitoring and governance to manage their inherent unpredictability of the intelligent systems that power AI agents.

Think of it this way…imagine if a CFO provided a CEO with different numbers each time the request was made.

But imagine that was a drawback inherent in an upgraded operating model with exponential upside and one that could be mitigated with the appropriate foresight and system design.

Instead of going back to the way things were done, you would mitigate that risk so that you could participate in the upside?

This architectural foundation enables the real deal competitive differentiator…

3: Competitive moats are being built by those building multi-agent orchestration over a knowledge-centric architecture.

“Finally, to truly scale agent deployment across the enterprise, the enterprise systems themselves must also evolve.”

“In the short term, APIs - protocols that allow different software applications to communicate and exchange data - will remain the primary interface for agents to interact with enterprise systems."

"But in the long term, APIs alone will not suffice."

"Organizations must begin reimagining their IT architectures around an agent-first model—one in which user interfaces, logic, and data access layers are natively designed for machine interaction rather than human navigation."

"In such a model, systems are no longer organized around screens and forms but around machine-readable interfaces, autonomous workflows, and agent-led decision flows.” - Seizing the Agentic AI Advantage by Alexander Sukharevsky , Dave Kerr , Klemens Hjartar , Lari Hamalainen , Stéphane Bout , Vito Di Leo , Guillaume Dagorret QuantumBlack, AI by McKinsey 06.13.25

Magical things happen when you move the definitions, relationships and business context necessary for AI agents to “understand the business” closer to the data.

A knowledge graph which operationalizes a company’s ontology, encodes both domain knowledge and business logic, defining the concepts and rules that an AI agent uses to understand the world within which it operates.

Enabling the AI agent to request data using business terms rather than grappling with the intricacies of database schemas or other complexity that slows everything down.

Increasing execution velocity as well as multiple agents, increasingly with more specialized functions, to collaborate in part with the help of a semantic layer that makes the knowledge-centric architecture that much more actionable.

Investments in data platforms, robust data modeling and semantic layers that provide the “business truth” to AI systems are propelling businesses forward, creating competitive moats along the way.

“Industry trends underscore this convergence, Gartner projects 75% of large enterprises will adopt multi-agent systems by 2026 and Boston Consulting Group (BCG) estimates these systems could generate $53 billion in business revenue by 2030 (up from $5.7 billion in 2024). [1] [2] Forward-thinking firms are taking note, as in Deloitte’s latest survey, many executives see deeply embedding AI into business processes as the #1 way to drive value from the technology.” - Why Multi-Agent Systems Needs Modern Data Architecture by Deloitte 07.22.25

The promise is real. The implementation challenges are real too.

But as Bill Gurley said in one of my favorite AI quotes - “Run at it.” 

You’ve got this.