Most engineering, product, and design (EPD) leaders can tell you their team’s AI adoption rate: how many licenses are active, which tools the team is using, how many tokens are consumed daily.
But adoption is the wrong metric to measure. It tells you whether your team has access to AI. It doesn’t tell you whether that AI is actually returning value. For most technical organizations right now, adoption is up, and value isn't keeping pace.
Why? AI returns value only when the environment around it is ready. For most technical teams, it isn’t.
The stack your AI is inheriting
Our latest research surveyed nearly 500 technical professionals across engineering, product, design, and IT and found that the environment most technical teams are working in was never built for AI success. Here’s what technical teams are saying:
69% say they have access to more tools than they actually need.
67% say many of their tools overlap in what they can do.
61% say their tools feel fragmented rather than part of one unified system.
61% say deciding how to do their work takes as much effort as actually doing it.
That’s the environment AI is walking into. When AI gets deployed into it, it adds yet another layer to manage: another tool in a stack that's already too big, with context that still doesn’t travel between systems.
This is why adoption and value diverge. Teams are using the tools. The tools just aren’t working together well enough for that usage to compound into anything meaningful.
The shadow AI signal
When the official stack isn’t working, technical teams don’t stop working. They route around it. They reach for tools that feel easier or work better with their other systems. This is shadow AI. It’s not only a signal that your official stack isn't working; it’s a liability. Unsanctioned tools fragment your data, create governance gaps, and leave IT with no visibility into how AI is actually being used.
Among technical workers, 67% report turning to unofficial AI tools because they’re easier or more effective than company-approved options. Over half (59%) say unofficial tools have better features, and 49% say they work better with their other tools.
The question worth asking isn’t whether your team is using AI. It’s whether the AI they’re using is the AI you're investing in.
The gap between deployment and value
Shadow AI reveals something else underneath the adoption problem: Most organizations haven't equipped their workers to use the tools they’ve already deployed. Licenses are active. The tools are there. But there’s an enablement gap between deployment and AI fluency. Organizations are deploying AI, but their workers aren’t adopting it correctly.
There are three concrete places where that gap shows up—and the data supports it:
Awareness: Organizations are deploying tools without telling workers what they do or when to use them.
22% of technical workers say they don’t know why certain features exist.
26% say they don’t know when to use them.
Training: Organizations are rolling out tools faster than they’re training workers to use them.
27% say they weren’t trained on tools that have already been deployed.
29% say they need better training on the capabilities they already have access to.
Capability: Even when workers are aware and trained, the tools themselves aren’t connected enough to support how technical work actually flows.
35% want tools that work better together.
32% want AI that understands their context and workflow.
None of these are individual failures. They’re the predictable result of treating deployment as the finish line.
Where does your organization stand?
The organizations getting compounding value from AI aren’t the ones with the most tools or the highest adoption rates. They're the ones that closed the gap between deployment and enablement.
The AI Value Quadrant is a diagnostic framework built from original survey data that shows exactly where that gap exists in your organization—and what to prioritize to close it.
Deploying more AI doesn’t move you up the quadrant. Closing the enablement gap does. Find out where your team stands.
