Agentic Work, Measured
Agents are moving from demos to the P&L. The companies that pull ahead will design the business around them, price what they deliver, and prove the return before the budget review does.
33%
of enterprise software apps will include agentic AI by 2028, up from under 1% in 2024
40%+
of agentic AI projects will be canceled by end of 2027: rising costs, unclear value, weak risk controls
Inside this pillar
Series · 7 parts · Complete
Agentic SaaS Blueprint
Why agent projects stall, and how to design the architecture, go-to-market, pricing and accountability as one system.
Inside the series →
Series · 4 parts · In progress
Demystifying Enterprise AI Productivity
Licenses aren't leverage. CLEAR and ROUTE turn scattered AI use into a return you can defend to finance.
Inside the series →
Series · Coming soon
The Agentic Org Chart
Function by function: where agents take the work, where people keep the judgment, and how to measure both. In the works.
Get it first: subscribe →
The clock is running
- Decisions are moving to agents. Gartner expects at least 15% of day-to-day work decisions to be made autonomously by agentic AI by 2028, up from zero in 2024. (Gartner, Jun 2025)
- "Agent washing" is muddying the market. Gartner warns that many vendors relabel existing chatbots and RPA as agentic AI, which makes value harder to prove and budgets easier to cut. (Gartner, Jun 2025)
- The gap is measurement, not models. Most pilots that fail do so on attribution and operating model, not capability, which is why the series here focus on architecture, pricing and proof.
What it moves on the board
| Change | Board metric it moves |
|---|---|
| Design business and technical architecture together | Project survival, time to value |
| Go-to-market built for autonomous outcomes, not seats | Win rate, expansion without seat growth |
| Price on what the agent delivers | Gross margin, net revenue retention |
| Measure AI as a balance sheet, not a license count | Operating cost per outcome, defensible ROI |
Forecasts are Gartner's; the MIT figure covers generative AI pilots broadly. Read them as direction, not precision.