But inside real companies, a different story is playing out: teams are adopting AI faster, spending more, and reporting ROI today—especially in coding, automation, and early agent deployments.
This matters for revenue teams because AI isn’t just changing productivity.
It’s changing how work gets done, how teams scale, and how companies build new capability without adding headcount.
Below is what the data is signaling—and how to respond if you want to be on the “leader” side of the gap.
What the data says right now
1) Agents are moving into production faster than most people think
KPMG’s AI Quarterly Pulse survey shows production agent deployment in large enterprises jumping from 11% to 42% in two quarters.
This doesn’t mean agents replaced whole teams in 2025.
It means the “pilot phase” is ending for a real chunk of the market—and companies are now dealing with the hard part: scaling safely, training teams, and operationalizing the work.
2) AI budgets are still rising
KPMG reports average planned AI investment rising to $130M (up from earlier in the year), and the direction is clear: spend is not pulling back.
The exact number matters less than the signal: leaders are doubling down.
3) CEOs believe ROI timelines are pulling forward
KPMG’s CEO research shows a sharp shift toward expecting returns within 1–3 years, versus the longer timelines many expected a year earlier.
That expectation alone changes behavior.
It means more pressure to prove impact, faster.
4) Most companies still aren’t at scale
McKinsey’s research reinforces the split: only 7% report AI as fully scaled across their organization.
So yes, adoption is real.
But most companies are still stuck between experiments and enterprise-wide execution.
The most common ROI is simpler than people admit
A large, self-reported ROI dataset (thousands of use cases) shows the most frequent impact is still time savings—often landing around 5 hours per week per person.
That sounds small until you do the math.
Five hours a week is roughly 7–10 work weeks per year regained per employee.
That’s not hype.
That’s capacity.
And capacity compounds.
(Important caveat: self-reported ROI always includes selection bias. But the pattern is still useful because it shows what teams *actually try first* and where they feel wins.)
The part most teams miss: “time saved” isn’t the goal
Time savings is the entry point.
But leaders don’t stop there.
They shift from:
“Save time” → to “Increase output”
“Increase output” → to “Improve quality”
“Improve quality” → to “Create new capability”
“New capability” → to “Revenue growth”
That’s the real divide.
Leaders treat AI like a portfolio, not a tool.
They run multiple initiatives in parallel and connect them into a coherent operating model—rather than launching isolated experiments that never stack together.
Where the next layer of ROI is coming from
1) Automation and agent workflows outperform
When teams build use cases explicitly around automation or agents, the ROI trend is stronger than generic “assist” use cases (drafting, summarizing, basic Q&A).
That makes sense.
Automation is where “time saved” becomes “work removed.”
2) Risk reduction is underused—and high upside
Most teams don’t choose “risk reduction” as their primary ROI category.
But when they do, it’s disproportionately likely to be transformational.
This is the hidden opportunity revenue leaders often underweight:
fewer bad fits entering the pipeline
fewer compliance mistakes
fewer forecasting surprises
fewer pricing and packaging errors
fewer handoff failures between Sales → CS
If you only measure ROI in hours saved, you miss this entirely.
What revenue leaders should do next
Step 1: Start with one measurable workflow
Pick a workflow with real volume and real friction:
lead routing
enrichment + hygiene
meeting-to-CRM updates
proposal + pricing workflows
enablement content creation + reuse
renewal risk triage
If you can’t measure it, you can’t improve it.
Step 2: Treat AI as a team capability, not an individual tool
If AI lives inside a few power users, you get scattered wins.
If AI becomes a shared workflow and shared language, you get scale.
That requires enablement: training, playbooks, guardrails, examples, and coaching—especially as agents move from novelty to real work.
Step 3: Build an ROI portfolio (not one big bet)
Most teams fail by trying to “pick the perfect use case.”
Leaders win by running multiple small bets, tracking what works, and expanding what proves out.
The ROI tends to increase as teams submit more use cases and connect them into a system.
Step 4: Don’t stop at time saved—convert capacity into output
The point isn’t to give people more free time.
The point is to redeploy that time into:
more pipeline
better deal execution
higher conversion
tighter forecasting
better customer experience
faster iteration on messaging and positioning
Time saved only matters if you turn it into outcomes.
The Takeaway
AI ROI isn’t a future story anymore.
The evidence points to three realities:
Adoption is accelerating, especially in technical teams and agent deployments.
Most companies still aren’t scaled, which creates a widening gap between leaders and laggards.
The winners will be the teams who move from isolated experiments to a systematic AI operating model tied to revenue outcomes.
If you want help identifying the highest-ROI revenue workflows, building an enablement plan, or optimizing your tech stack for agentic workflows, we can help.
Want to learn more about how we can help you transform your revenue efficiency - Schedule a consultation**.**