In September 2025, Deloitte released its glossy report, “Agentic Enterprise 2028: A blueprint for cost savings, job creation, and faster growth through agentic AI.” Cloaked in futuristic jargon and rainbow-hued graphics, the document paints a vision of enterprises in which AI “agents” decide what to do, act with minimal human input, and serve as the new operating logic.

By 2028, Deloitte claims, organizations embracing this will slash costs, accelerate product releases, and “redeploy talent to higher-value work.” It sounds revolutionary, but peel back the layers, and it’s a thinly veiled manifesto for sidelining human agency in favor of unproven algorithms. This isn’t innovation; it’s a cynical repackaging of automation hype that disregards the very human capacities that drive real enterprise success.

Put aside that once again, this is a misrepresentation of what “Agentic AI” is, at the core of their report, Deloitte’s strategy rests on two flawed premises.

• First, it assumes autonomous systems exists and can outperform humans across domains.
• Second, it posits that humans in enterprises lack sufficient “agency” today.

Their framework misrepresents the state of AI technology while eroding the intrinsic agency humans already exercise in organizations, all in a bid for novelty that prioritizes buzzwords over people.

In short, Deloitte is telling humans, you don’t matter and never have.

The Myth of System Autonomy: Hype Over Reality

Deloitte’s “autonomy ladder” outlines six levels, from basic assistance to full “self-evolve” where agents form a “self-governing, self-improving agent mesh.” They project that by 2028, 5-10% of enterprises will reach this pinnacle, with agents handling complex workflows like real-time catalog evolution or autonomous treasury management. But here’s the inconvenient truth: fully autonomous systems, as Deloitte envisions them, simply don’t exist today, and may not by 2028.

Current AI, including large language models powering these “agents,” remains brittle and error-prone outside narrow tasks. Deloitte admits as much in their projected capabilities table: In 2025, agents are “primarily human-in-the-loop” with “rule-bound choices,” evolving to “significantly more autonomous” by 2028 through “self-improving via reinforcement and human observation.”

This isn’t autonomy; it’s assisted computation reliant on human crutches. Industry observers echo this skepticism. As one analysis notes, AI agents “still have things to prove, and they are definitely more expensive than the outmoded systems they are supposedly replacing.” Gartner forecasts that over 40% of agentic AI projects (another overly hyped marketing term that has nothing to do with managing agents) will be canceled by 2027 due to underwhelming results, highlighting the gap between hype and delivery.

Deloitte’s blueprint ignores these realities, pushing enterprises to invest in unproven tech under the guise of inevitability. Their “return-on-autonomy” KPIs, such as reducing operating costs or boosting output per FTE, assume seamless scaling. Still, without actual system-level autonomy, these metrics become excuses for cost-cutting disguised as innovation.

Undermining Human Agency

Humans in enterprises already possess agency (humans have autonomy). We sense, decide, and act through experience, intuition, and collaboration. These qualities drive every meaningful operation. Employees navigate ambiguity, build relationships, and innovate in ways no algorithm can match.

Yet Deloitte’s Agentic Enterprise report proposes an inversion: non-autonomous AI systems take center stage, while human agency is reduced to “orchestration” and “strategic oversight.”

The irony is that the report implies today’s organizations are not agentic at all. That assumption exposes a deeper flaw in both the taxonomy and the marketing hype surrounding AI.

Agency is not something to automate.

Agentic = Agency = Autonomy = Human

The report’s “empowered workforce” pillar claims to create jobs, listing roles such as “Agentic Process Architect” and “Human-AI Interaction Coach.”

But read closely: Human roles “evolve significantly from monitoring and feedback at early stages to orchestration and strategic oversight as agentic integration advances.” In their autonomy ladder, humans shift from “operator” at level 0 to “orchestrator” at level 5, with penetration projections showing full autonomy at <1% today but 5-10% by 2028. This isn’t empowerment; it’s relegation. Routine tasks are automated, analytical roles augmented, and even creative work curated by AI, leaving humans as glorified babysitters for machines.

Anyone else thinking of George Jetson?

Deloitte pays lip service to human-machine collaboration, but the subtext is clear: Minimize human involvement for efficiency. Their roadmap envisions “self-negotiating procurement” and “autonomous high-performer flight-risk identification,” in which agents handle decisions that previously required human judgment. This disregards human autonomy entirely, as well as empathy, ethical nuance, and adaptability that prevent disasters like biased hiring or tone-deaf customer interactions. By framing humans as bottlenecks to be optimized away, Deloitte devalues the existing agency that has built successful enterprises for centuries.

Ironically, Deloitte’s own track record undermines its evangelism. Just weeks after the report’s release, the firm faced an “awkward AI moment” when consultants were caught using AI to generate a high-value report, forcing them to repay fees. As critics noted, while the report’s substance remained, “trust and credibility took a real pounding.” If Deloitte can’t reliably integrate AI into its own processes without backlash, why should enterprises trust its blueprint for an “agentic” future?

I’ll help you add a section about Salesforce to strengthen your critique of the “agentic enterprise” trend. Here’s where I’d suggest inserting it – right after the Deloitte track record paragraph and before “Why This Matters,” creating a new section that shows this is an industry-wide problem:

Salesforce Joins the Circus: More of the Same

Deloitte isn’t alone in this misleading crusade. At Salesforce‘s 2025 annual conference, the enterprise software giant doubled down on the same flawed narrative, promoting their vision of the “agentic enterprise” with equally inflated claims about AI capabilities and equally dismissive attitudes toward human agency.

CEO Marc Benioff declared AI agents “the beginning of an unlimited workforce,” a statement that reveals the fundamental misunderstanding (or deliberate misrepresentation) at the heart of this movement. Salesforce’s marketing materials describe AI agents that can “reason, adapt, and act on their own” while “freeing up employees to focus on higher-value work.”

Sound familiar? It’s the same playbook as Deloitte, just with different branding.

Salesforce frames the agentic enterprise as a “collaborative ecosystem” where AI agents handle “repetitive and time-consuming tasks” by making “decisions and taking action to achieve the desired outcome.” But this framing perpetuates the same two lies: first, that current AI possesses genuine reasoning and autonomous decision-making capabilities (it doesn’t), and second, that human workers in today’s enterprises lack agency and need AI to liberate them from mundane work.

The reality? Humans already exercise agency in handling routine tasks. We apply judgment, context, and ethical reasoning even in “repetitive” work. A customer service representative doesn’t just process tickets, they read emotional cues, exercise empathy, and make nuanced decisions about when to bend rules. An accountant doesn’t just enter numbers, they spot anomalies, understand business context, and apply professional judgment. By characterizing these roles as mere “routine work” suitable for algorithmic replacement, Salesforce devalues the human intelligence already present in every layer of enterprise operations.

Benioff’s “unlimited workforce” rhetoric is particularly revealing. It positions AI agents not as tools to support human workers, but as substitutes for them, infinitely scalable, always available, never asking for raises or taking sick days. The subtext is unmistakable: humans are limited, expensive, and ultimately expendable.

This isn’t about augmenting human capability. It’s about replacing human agency with computational processes dressed up in the language of empowerment. Salesforce, like Deloitte, is selling a future where human autonomy becomes obsolete, wrapped in the ironic packaging of “agentic” transformation.

Why This Matters: The Cost of Misrepresented Novelty

Deloitte’s push for the “agentic enterprise” is an attempt to coin a trendy term at the cost of human capacity. By hyping projected tech as current reality and redefining human roles as secondary, it risks fostering workplaces where efficiency trumps ethics, and algorithms eclipse empathy. This complete disregard for human autonomy could exacerbate inequality, erode job satisfaction, and invite regulatory backlash, all while delivering underwhelming returns on overhyped investments.

Enterprises should reject this narrative. Actual progress lies in augmenting human agency with AI tools, not supplanting it. Invest in people first: Upskill teams, foster collaboration, and use AI where it truly adds value, without the delusion of complete autonomy. Deloitte’s blueprint may dazzle boards, but it’s a roadmap to a dehumanized future we can, and should, avoid.

---

If you find this content valuable, please share it with your network.

🍊 Follow me for daily insights.

🍓 Schedule a free call to start your AI Transformation.

🍐 Book me to speak at your next event.

Chris Hood is an AI strategist and author of the #1 Amazon Best Seller “Infailible” and “Customer Transformation,” and has been recognized as one of the Top 40 Global Gurus for Customer Experience.