The Agent Shift: Why AI Agents Are Reaching Big Companies First

From assistants to agents

The first wave of generative AI gave people assistants: tools that draft, summarise or suggest while a human stays in charge of every step. The second wave is agents: systems that take a goal, plan the steps, use other software and complete the work, checking in with people only when needed.

The difference matters. An assistant makes a marketer faster. An agent can do parts of the marketer's job.

How fast the shift is happening

Spending

Gartner forecasts that worldwide spending on all forms of AI will reach about $2.7 trillion in 2026, up 49.5% on the previous year [1]. That covers everything from chips and data centres to software and services. It isn't a figure for agents alone, and it's often misquoted as one.

Within that total, Gartner now tracks a separate category for AI agents and assistants. Spending there is forecast to grow from about $16.5 billion in 2025 to $29.2 billion in 2026 and $65.5 billion in 2027 [1]. It's a small slice of total AI spend, but one of the fastest-growing.

For context, Gartner's previous forecast, published in September 2025, put all AI spending at nearly $1.5 trillion for 2025 and expected it to top $2 trillion in 2026 [2]. The trajectory has only steepened since.

Adoption in software

Gartner predicts that 40% of enterprise applications will include task-specific AI agents by the end of 2026, up from less than 5% in 2025 [3]. In a single year, agents move from a novelty to a standard feature of the business software companies already use.

Adoption inside companies

McKinsey's State of AI in 2026, based on a survey of 1,719 respondents across 97 countries, reports that [4]:

• Nearly nine in ten respondents say their organisations regularly use AI in at least one function.
• 44% say AI is now scaling across their enterprise, up from 38% a year earlier.
• 40% of large organisations (annual revenues above $1 billion) report scaling AI agents, up from 27% the previous year.
• Smaller organisations were flat at about 22%.

The gap that matters most

The last two numbers are the most important in the data. Large companies are pulling ahead: agent scaling at big organisations jumped 13 points in a year, while smaller organisations didn't move.

That gap has a clear explanation. Deploying agents today still takes resources that large companies have and small ones don't:

• Technical staff to integrate agents with existing systems and data
• Clean, connected data for agents to act on
• Governance teams to manage risk, security and compliance
• Budget for experiments, many of which won't pay off

A company with thousands of employees can fund a dedicated AI programme. A company with one employee can't.

That creates a paradox. Small businesses, which can least afford to hire people, are the ones that would benefit most from agents that do the work of a team. Yet they're adopting agents most slowly.

Why marketing is at the front of the shift

Some functions are better suited to agents than others. Marketing stands out for three reasons.

It's where AI already pays. McKinsey reports that respondents most often attribute AI-driven revenue gains to marketing and sales [4]. It's one of the few areas where AI's impact shows up directly in the top line.

Marketers are already using AI. Salesforce's 2026 State of Marketing report, based on a survey of 4,450 marketing decision makers, found that 75% of marketers have adopted AI [5]. Most of that use is still assistive, such as drafting copy or generating images, which makes the move to agents a natural next step.

The work is continuous and data-rich. Advertising involves constant decisions: which channel, which audience, which ad, how much to bid, when to stop. These decisions are repetitive, measurable and time-sensitive, exactly the kind of work agents are designed for.

The warning in the data

Not every agent project succeeds. Gartner predicts that more than 40% of agentic AI projects will be cancelled by the end of 2027, citing escalating costs, unclear business value and inadequate risk controls [6]. It has also warned about "agent washing", products rebranded as agents without real autonomous capability [6].

This is a useful corrective to the hype. Agents that can't show measurable value will be switched off. For marketing, that sets a clear bar: an agent should be judged on business results, such as revenue, retention and return on ad spend (ROAS), not on how many tasks it automates.

What this means for businesses

For large companies, the agent shift is already under way. The question is how to move from pilots to measurable results, and how to avoid paying for agents that don't deliver.

For small and mid-sized businesses, the opportunity is bigger and the risk is being left behind. They need agents that:

• Work without an integration team. They should connect to the platforms the business already uses.
• Run a whole function, not just tasks. Acquisition, creative and retention, joined up.
• Include guardrails by default, such as budget limits, brand rules and approval points.
• Are priced on outcomes, so the business isn't paying for experiments that fail.

For marketing leaders everywhere, the role is changing. Less time on operating tools, more time on setting goals, judging results and making the calls agents shouldn't make alone.

The takeaway

The agent shift is real, fast and uneven. Agent spending is forecast to more than double in a year [1], and task-specific agents are spreading through enterprise software [3]. But large companies are scaling them almost twice as fast as smaller ones [4], and many projects will fail on cost and unclear value [6].

The winners will be agents that deliver measurable business outcomes, and that reach the businesses that need them most, not just the ones that can afford an AI programme.

---

References

[1] Gartner. Gartner forecasts worldwide AI spending to grow 49.5% in 2026. September 2026. https://www.gartner.com/en/newsroom/press-releases/2026-09-16-gartner-forecasts-worldwide-ai-spending-to-grow-49-point-5-percent-in-2026

[2] Gartner. Worldwide AI spending will total $1.5 trillion in 2025. September 2025. https://www.gartner.com/en/newsroom/press-releases/2025-09-17-gartner-says-worldwide-ai-spending-will-total-1-point-5-trillion-in-2025

[3] Gartner. 40% of enterprise apps will feature task-specific AI agents by 2026. August 2025. https://www.gartner.com/en/newsroom/press-releases/2025-08-26-gartner-predicts-40-percent-of-enterprise-apps-will-feature-task-specific-ai-agents-by-2026-up-from-less-than-5-percent-in-2025

[4] McKinsey & Company. The state of AI in 2026: On the road to ROI. August 2026. https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai

[5] Salesforce. State of Marketing, 10th edition. February 2026. https://www.salesforce.com/news/stories/state-of-marketing-2026/

[6] Gartner. Over 40% of agentic AI projects will be canceled by end of 2027. June 2025. https://www.gartner.com/en/newsroom/press-releases/2025-06-25-gartner-predicts-over-40-percent-of-agentic-ai-projects-will-be-canceled-by-end-of-2027