The Chase-and-Catch Framework for AI Adoption in Agentic AI 2026

Executive Summary

New York, 2026. Agentic AI entered 2026 as the enterprise technology with the most aggressive adoption intent on record — and the widest gap between ambition and execution. Gartner reports that only 17% of organizations have deployed AI agents to date, yet more than 60% expect to within two years. McKinsey finds 23% of respondents scaling an agentic system somewhere, but no more than 10% scaling within any single function. Forrester frames it bluntly: enterprises are chasing, few are catching. This article presents the Chase-and-Catch Framework — a four-phase adoption model synthesised from verified research by McKinsey, Gartner and Forrester, with decision criteria and named enterprise examples for each phase. It is designed to help decision-makers move from experimentation to defensible, EBIT-positive production.

Key Takeaways

• Gartner projects 40% of enterprise applications will integrate task-specific AI agents by end of 2026, up from under 5% in 2025.
• McKinsey reports 62% of organisations are at least experimenting with AI agents, but only 39% report enterprise-level EBIT impact.
• Forrester's "trust tax" — the cost of logging and defending every autonomous action — is the central barrier to scaled multi-agent systems.
• Salesforce Agentforce reached US$800 million ARR (up 169% year-on-year), the clearest primary-sourced commercial signal in the sector.
• Governance lags capability: McKinsey finds only ~30% of organisations reach a responsible-AI maturity level of three or higher on agentic controls.
• Gartner predicts that by 2030, 50% of agent deployment failures will stem from insufficient governance runtime enforcement.

Market Analysis: Sizing the Gap Between Chase and Catch

The 2026 data landscape is defined by a consistent split: adoption intent is extraordinary, production reality is modest. Gartner places agentic AI at the Peak of Inflated Expectations on its 2026 Hype Cycle, noting that most deployments remain narrowly scoped and fully autonomous agents are not ready for the majority of enterprise use cases. Its best-case scenario projects agentic AI could drive roughly 30% of enterprise application software revenue by 2035 — surpassing US$450 billion, up from 2% in 2025. McKinsey's State of AI survey supplies the adoption baseline, while Forrester supplies the readiness caution: the technology has arrived, but enterprise readiness has not caught up.

The through-line: enterprises are buying capability faster than they are building governance. That mismatch defines the framework below.

The Chase-and-Catch Framework: Four Phases

The framework organises agentic adoption into four sequential phases — Experiment, Contain, Govern, Scale — each with its own decision criteria and verified enterprise reference point. It maps directly onto the data: most organisations sit in Phases 1 and 2, while the EBIT payoff concentrates in Phases 3 and 4.

Phase 1 — Experiment

Characterised by pilots and "agentish" chatbots. Forrester's State Of Agentic AI, 2026 observes that three-quarters of enterprise leaders say they are adopting agentic AI, but only a small minority have it running in meaningful production beyond chatbots, and true scaled multi-agent systems are rarer still. McKinsey finds that beyond the 23% scaling an agentic system, an additional 39% of organisations have begun experimenting with AI agents — 62% at least experimenting in total.

Per Forrester's Q1 2026 Technology Landscape Assessment, Based on evaluation of 150+ vendor implementations and third-party assessments, Related: Hermes vs OpenClaw: Which is Better, Autonomous AI Agent?

Decision criteria for exiting Phase 1: a documented use case with measurable cost or revenue baseline; a human-in-the-loop design; and a clear owner. The risk is pilot sprawl — Salesforce's own Connectivity Report notes organisations use an average of 12 AI agents, with roughly half operating in isolated silos rather than coordinated systems.

Phase 2 — Contain

Here the enterprise narrows scope deliberately. Gartner's guidance is explicit: most successful deployments remain narrowly scoped, because fully autonomous agents are not ready for the majority of enterprise use cases. Forrester predicts fewer than 15% of firms will turn on the agentic features in intelligent automation suites in 2026, expecting most to keep running deterministic automation given ROI and governance challenges. Containment is not failure — it is the rational response to the "trust tax," Forrester's term for the currently prohibitive cost of logging and making every autonomous action defensible to an auditor.

For deeper context, see our Agentic AI analysis: "OpenClaw vs. NemoClaw: Which One Is Better for Businesses and Enterprise".

Phase 3 — Govern

Governance is the phase most organisations skip and later regret. McKinsey's 2026 AI Trust Maturity Survey — drawing on roughly 500 organisations surveyed between December 2025 and January 2026 — found that responsible-AI maturity is improving, but strategy, governance and agentic controls lag, with only about 30% reaching maturity level three or higher. The survey's key insight: in the agentic era, organisations must contend not only with systems saying the wrong thing but doing the wrong thing — taking unintended actions, misusing tools, or operating beyond guardrails. Gartner reinforces the stakes, predicting that by 2030 half of all agent deployment failures will trace to insufficient governance runtime enforcement for capabilities and multisystem interoperability. In Forrester's Security Survey 2026, 49% of security decision-makers named agentic AI as a concern.

Phase 4 — Scale

The catch. Few organisations reach it, and the ones that do share a common trait: a workforce ready to manage highly autonomous agents inside tightly controlled processes. Forrester singles out Bank of New York as about as far out front as a regulated enterprise gets — noting that even BNY has not captured the full value of agentic promises, but has something most lack: a workforce prepared to manage autonomous agents inside a tightly regulated business. On the commercial side, Salesforce provides the sector's clearest scale signal: Agentforce ARR of US$800 million, up 169% year-on-year, with 29,000 deals closed and 2.4 billion agentic work units delivered across Agentforce and Slack. Combined Agentforce and Data Cloud ARR reached roughly US$1.8 billion, up from US$1.4 billion a quarter earlier. Salesforce reports over US$100 million in annualised customer cost savings and a 34% productivity increase from agentic and generative AI, according to Salesforce's own customer-reported outcomes (not independently verified).

Additional coverage: Agentic AI Faces A Security Stress Test: New Guardrails, Regulatory Heat, and Risk Findings

Competitive Landscape

The vendor field splits between platform incumbents embedding agents into enterprise applications and the analyst houses defining the maturity models buyers use to evaluate them. Forrester predicts that in 2026 enterprise applications will move beyond enabling employees with digital tools toward accommodating a digital workforce of AI agents — a structural shift benefiting incumbents with existing data gravity.

Practical Business Implications

For decision-makers, the framework converts noise into sequence. First, resist the temptation to jump from Phase 1 to Phase 4 — Forrester expects enterprises to delay 25% of AI spend into 2027 precisely because value is failing to land, with only 15% of AI decision-makers reporting an EBITDA lift. Second, treat governance (Phase 3) as a prerequisite, not a compliance afterthought; Gartner's failure forecast makes runtime enforcement a board-level concern. Third, invest in workforce readiness — Forrester predicts 30% of large enterprises will mandate AI training, and BNY's advantage is human, not merely technical. Finally, measure at the use-case level before claiming enterprise EBIT impact; McKinsey's 39% figure is the honest ceiling for 2026. These dynamics echo broader enterprise-software repricing trends, such as those documented in our coverage of how Vercel Signals IPO Readiness as AI Boosts Revenue Growth in 2026.

Related: Microsoft Takes Personal AI to Next Level with Microsoft Scout

Forward Outlook

The 12–24 month trajectory is one of consolidation around governed, narrowly scoped agents. Gartner's longer horizon — 90% of B2B buying AI-agent-intermediated by 2028, pushing over US$15 trillion of B2B spend through agent exchanges — suggests the destination is real even if 2026 is a way-station. The organisations that will capture that value are those completing Phase 3 now. As with adjacent automation waves chronicled in The Rise of Robotics: Transformation Trends in 2026, the winners treat autonomy as an operating-model change, not a feature toggle. Parallel data-integration lessons appear in How Health Tech Is Integrating Data and Care in 2026, According to Gartner and Philips, while infrastructure repricing themes surface in PacBio: Sub-$300 HiFi Genome Resets Long-Read Sequencing Economics and connectivity shifts in Consumers Shift to Satellite-First as Starlink Holiday Deals and FCC D2D Approval Reshape Buying.

Frequently Asked Questions

What is the Chase-and-Catch Framework?

It is a four-phase adoption model — Experiment, Contain, Govern, Scale — synthesised from verified 2026 research by McKinsey, Gartner and Forrester. It sequences agentic AI deployment so organisations build governance before scaling autonomy.

For deeper context, see our Fintech analysis: "Fintech Backbone Rewired: Visa Direct, Plaid, SWIFT Move Faster on Real-Time Rails".

How many enterprises have actually deployed agentic AI in 2026?

Gartner reports only 17% of organisations have deployed AI agents to date, though more than 60% expect to within two years. McKinsey finds 62% at least experimenting but only 23% scaling somewhere in the enterprise.

Why do most agentic AI projects struggle to show ROI?

Forrester attributes this to the "trust tax" — the high cost of logging and defending every autonomous action — and reports only 15% of AI decision-makers see an EBITDA lift. McKinsey finds just 39% report enterprise-level EBIT impact.

Which enterprise deployment has the strongest verified data?

Salesforce Agentforce, which reported US$800 million ARR (up 169% year-on-year) in its FY2026 earnings reporting, alongside 29,000 deals closed — the sector's clearest primary-sourced commercial signal.

What is the single biggest predictor of failure at scale?

Gartner predicts that by 2030, 50% of agent deployment failures will be due to insufficient AI governance platform runtime enforcement for capabilities and multisystem interoperability — making Phase 3 governance the critical gate.

Sources include company disclosures, regulatory filings, analyst reports, and industry briefings.

Related Coverage

• More Agentic AI Analysis
• Latest Technology News

Analysis based on company announcements, investor disclosures, regulatory filings, Reuters, Bloomberg, Financial Times, CNBC, SEC documentation, and publicly available market data as of publication.