Gartner: 15% of Work Decisions Will Be Autonomous AI by 2028

The News

Gartner released projections this month stating that by 2028, 15% of day-to-day work decisions will be made autonomously through agentic AI -up from effectively 0% in 2024. That's a 0-to-15% leap in four years, a steep curve for any enterprise technology.

Why This Matters

This isn't about chatbots answering customer questions. Gartner specifically defines "autonomous decisions" as agents that:

• Evaluate options based on business context
• Make choices without human approval
• Execute actions (send emails, approve purchases, update systems)
• Learn from outcomes to improve future decisions

Think: An AI agent that reviews 200 inbound sales leads, scores them, and automatically books meetings with the top 15 -without a human reviewing each one.

What's Driving This

Three catalysts converged in 2024 to make this prediction credible:

1. Function-calling APIs matured

OpenAI, Anthropic, and Google all shipped reliable function-calling capabilities in 2023-2024. Agents can now trigger external tools (databases, CRMs, email) far more dependably than earlier models could.

2. Enterprise trust increased

McKinsey's 2025 AI report shows 88% of organizations now use AI in at least one business function regularly, up from 55% in 2023. Early adopters proved agents work reliably enough for production use.

3. ROI became measurable

As more teams run agents on real workflows, it becomes easier to measure the hours saved against the cost. That evidence weakens the "too risky" objection and shifts the conversation from whether to deploy agents to which workflows to automate first.

Which Decisions Will AI Make?

The decisions most likely to be handed to agents first are:

Notice a pattern: high-volume, bounded decisions where mistakes are low-stakes and reversible.

Strategic decisions (M&A, product direction), personnel decisions (hiring, firing) and high-risk compliance decisions (legal approvals) are far less likely to be automated any time soon. Expect those to stay with humans.

What This Means for Your Business

If you're not piloting agents yet:

Companies piloting now will have years of refinement by 2028, giving them operational advantages that competitors can't catch up to quickly.

Start with one workflow. Common first choices:

• Customer support triage
• Sales lead qualification
• Finance/accounting automation

If you're already piloting:

Focus on expanding from a single-department pilot to multi-department production. The gap between an early pilot and scaled deployment is often many months; companies that navigate it well will hit Gartner's 15% threshold ahead of the market.

If you're at scale:

You're ahead of most enterprises. Double down on:

• Multi-agent orchestration (coordinate multiple specialized agents)
• Human-in-the-loop workflows (blend autonomous + supervised decisions)
• Measuring and communicating ROI internally (become the case study others reference)

The Contrarian Take

15% sounds aggressive, but it might be conservative. Here's why:

UiPath's agentic AI research suggests a meaningful share of enterprises are *already* experimenting with agentic workflows. If adoption follows typical enterprise S-curves (slow start, rapid middle, plateauing end), 15% of *decisions* being autonomous by 2028 could actually undershoot.

Support, sales and finance generate a large share of an organisation's routine decisions. If companies automate a good portion of those, the total climbs quickly.

The real question may not be whether we'll hit 15%, but how far past it we go.

What to Do Now

1. Audit your decision-making workflows

List all recurring decisions your team makes weekly. Which are:

• High-volume (>10/week)?
• Rule-based or pattern-based (not requiring novel judgment)?
• Low-stakes (mistakes are reversible)?

Those are automation candidates.

2. Pilot ONE workflow by Q1 2025

Don't wait for perfect. Pick your highest-pain, best-fit workflow and run a 60-day pilot. Measure:

• Accuracy (% of agent decisions matching what humans would do)
• Coverage (% of decisions agent handles vs escalating to humans)
• Time saved (hours per week reclaimed)

3. Build organizational buy-in

A common barrier to agent adoption isn't technology, it's change management. Teams resist when they don't understand what agents will/won't do or fear job displacement.

Be transparent: "Agents handle repetitive decisions you hate; you focus on complex judgment calls requiring expertise."

Sources & Further Reading

• Gartner's full agentic AI predictions
• McKinsey State of AI 2025
• UiPath Agentic AI Report 2024

---

Bottom line: 15% of work decisions becoming autonomous by 2028 is a massive shift, freeing up real capacity in every team without hiring. The winners will be companies that start piloting systematically now, not those waiting for "more mature" technology in 2026-27.

The technology is ready. The question is whether your organization is.

---

Frequently Asked Questions

Q: How do I measure automation ROI?

Calculate time saved per execution multiplied by execution frequency, reduction in error rates, faster cycle times, and freed-up capacity for higher-value work. Well-scoped automation often pays back within months.

Q: What processes should I automate first?

Start with high-volume, low-complexity tasks that cause friction - data entry, report generation, routine communications. These deliver quick wins that build confidence and budget for more sophisticated automation.

Q: How do I avoid over-automating?

Maintain human touchpoints for decisions requiring judgment, customer interactions where empathy matters, and processes where errors have high consequences. The goal is augmentation, not complete removal of human involvement.