Most enterprises are duct-taping AI agents onto operating models designed for humans with clipboards — and then wondering why the ROI never shows up. The honest read on agentic AI right now isn’t that the technology is overhyped. It’s that buyers keep treating it like a productivity plugin when it actually demands a redesign of the org chart, the tech stack, and the scoreboard. The numbers in a recent MIT Technology Review Insights report make the gap painfully clear.
The Ambition-Execution Gap Is Wider Than Leaders Admit
According to MIT Technology Review Insights, 85% of organizations say they want to be agentic within the next three years, but 76% say their current operations and infrastructure can’t support that change, citing a lack of readiness across people, processes, and workflows (per a Celonis report cited in the piece). That’s a nine-point gap between intent and capability — and it’s the entire ballgame.
Why this matters: when leadership commits to an agentic roadmap without the underlying scaffolding, projects don’t fail loudly. They fail quietly through disillusionment, stalled pilots, and budget reallocations away from AI in the next planning cycle. Prasun Shah, global CTO for workforce consulting and chief AI officer at PwC UK Consulting, calls this the “sticky tape problem”: embedding AI employees into what is still fundamentally a human operating model. The value gets capped before anyone even measures it.
If you’re a mid-market insurer running a claims-automation pilot, this looks like dropping an AI agent in front of three legacy systems that still expect a human to copy-paste between them. The agent works. The workflow doesn’t. Our take: 2026 will be the year boards stop funding agent pilots that don’t come bundled with operating-model redesign — the patience for sticky-tape demos is almost gone.
Why Agentic Business Transformation Isn’t Just A New Buzzword
Enterprise agentic AI platform Ema, in partnership with HFS Research, coined the term “agentic business transformation” (ABT) last year to describe what they argue is a categorically different shift. “Digital transformation was about moving from paper to software. AI transformation was about adding artificial intelligence to existing processes. Co-pilot is about AI assisting in various human tasks,” says Ema CEO Surojit Chatterjee. “But ABT is something categorically different: It’s the integration of AI agents into the fabric of the organization.”
The vocabulary matters because it forces a different conversation in the boardroom. Co-pilot framing leads to seat-license budgets and individual productivity dashboards. ABT framing forces leaders to redesign operating models, workflows, decision rights, and performance systems — what Shah describes as everything needed to make agents “active participants in value creation, rather than just point tools or productivity aids.” ABT rests on three pillars: technology stack, workforce, and success metrics.
Picture a regional bank that has spent two years rolling out co-pilots across customer support. Volume metrics look great. Loan origination time hasn’t moved. That’s the difference between AI transformation and ABT — and it’s why teams evaluating these tradeoffs should read up on the practical line between AI agents and AI automation before signing another procurement order. Our prediction: within 18 months, “ABT readiness” will start showing up in analyst frameworks the same way “cloud maturity” did a decade ago.
Agents Are Connective Tissue, Not Another Stack Layer
The first pillar — the technology stack — is where most engineering leaders get the architecture wrong. “Your existing tech stack was designed for human-operated, application-centric workflows,” Chatterjee notes. “It needs to be reconsidered when the actor is an AI agent operating at machine speed across multiple systems simultaneously.” Shah pushes the point further: the real value of agents isn’t as another layer in the stack, but as connective tissue moving across layers to coordinate tasks and interpret data from multiple discrete applications.
Why this matters: it inverts how IT typically buys software. Instead of standing up another SaaS tool in a workflow, leaders need to invest in the integration substrate that lets agents reach across CRM, ERP, ticketing, and data warehouses simultaneously. According to BCG research cited in the piece, AI agents could accelerate business processes by 30% to 50% and cut low-value work time by 25% to 40% when deployed at scale — but only when they can actually reach the systems where the work lives. That’s an integration problem before it’s a model problem, which is why enterprise-grade API and data pipeline work tends to be the unglamorous prerequisite nobody budgets for.
Imagine you run operations at a logistics company. Today, rerouting a delayed shipment requires four humans touching three systems. With agents as connective tissue, one configured agent reads the carrier feed, checks customer SLAs in the CRM, updates the WMS, and notifies the account manager — in seconds. Chatterjee claims time from business need to production workflow drops from months to days when this architectural shift lands. Our take: the winners of the next five years won’t be the companies with the best models, they’ll be the ones with the cleanest integration fabric underneath them.
Redesigning The Workforce Around Hybrid Teams
The second pillar is workforce, and this is where the change gets uncomfortable. Workforce structures today still mirror the hierarchical model of early industrialization — standardized processes, clear lines between strategic business units, promotion based on managing the people below you. Agents break that model because they execute, coordinate, and optimize without managerial coordination. McKinsey estimates that by 2030, three-quarters of current jobs will require redesign, upskilling, or redeployment.
Managers don’t disappear; their job description changes. Shah argues managers “will need to be able to manage issues around trust, explainability, psychological safety, and even status dynamics” in hybrid teams. That’s a different competency profile than what most leadership development programs produce today. Recruitment, retention, and remuneration all need to be rebuilt around it.
If you’re an HR leader at a 5,000-person professional services firm, this means your 2027 promotion criteria probably shouldn’t look like your 2024 ones. The senior associate who used to be valued for managing six juniors may now be valued for managing two juniors and twelve agent workflows — and your comp bands have no language for that. Our prediction: the first wave of Chief Agent Officer or “Head of Human-Agent Operations” titles lands in the Fortune 500 within the next 24 months.
From Output Metrics To Outcome Metrics
The third pillar is the scoreboard, and it’s the one most likely to get quietly skipped. “When you add AI employees into the workforce, activity metrics become meaningless or actively misleading,” Chatterjee warns. An agent handling a thousand customer interactions in the time a human handles ten will look heroic on a volume dashboard while telling you nothing about retention or revenue.
The source offers a concrete data point: when one of Ema’s large enterprise customers swapped tool metrics like cost-per-query and AI accuracy for outcome metrics like the percentage of contracts reviewed without human escalation, measured ROI from agentic AI tripled within two quarters. The customer also stopped pouring money into point solutions for high-volume, low-complexity work and redirected investment toward workflows with the highest outcome value. Without it, AI-integrated software programs are just vanity pilots.
For a SaaS company, the practical shift looks like this: stop reporting “tickets resolved by AI” in the board deck. Start reporting “net revenue retention on accounts touched by agents versus human-only accounts.” Shah adds that new metrics force new accountability questions too — who owns the mistake when an agent gets it wrong, and what guardrails protect the customer. Our take: by the end of 2026, expect activity-metric dashboards to be quietly retired in any serious agentic deployment, replaced by outcome dashboards that finally tell the truth.
FAQ
Q: What is agentic business transformation (ABT)? A: ABT is a term coined by Ema and HFS Research to describe the integration of AI agents into the fabric of an organization — not just adding AI to existing processes, but redesigning the operating model, workforce, and success metrics around agents as active participants. It’s positioned as a categorically different shift from digital transformation, AI transformation, or co-pilot adoption.
Q: Why aren’t most enterprises ready for agentic AI? A: According to the MIT Technology Review Insights piece, 85% of organizations want to be agentic within three years but 76% say their current operations and infrastructure can’t support it. The dominant failure pattern, per PwC’s Prasun Shah, is layering agents onto human-designed operating models — the “sticky tape problem” — instead of rewiring how work flows.
Q: What metrics should replace traditional activity-based KPIs? A: Outcome-based metrics. Ema cites a customer that shifted from cost-per-query and AI accuracy to measures like the percentage of contracts reviewed without human escalation, and saw measured ROI from agentic AI triple within two quarters.
Key Takeaways
- Treat operating-model redesign as a prerequisite for agent deployment, not a follow-up project — sticky-tape rollouts will increasingly fail board scrutiny.
- Invest in integration fabric before buying more agent tools; the connective-tissue layer is where competitive differentiation will compound.
- Rebuild manager job descriptions and promotion criteria for hybrid human-agent teams now, before McKinsey’s 2030 redesign pressure hits your retention numbers.
- Retire activity metrics in any workflow touched by agents and replace them with outcome metrics tied to revenue, retention, or risk reduction.
- Resolve accountability and guardrail questions at the senior leadership level early — diffuse operational ownership without clear fiduciary lines is the fastest path to a public agent failure.