AI Is Shipping Faster Than Customers Can Adopt. Here's How Google Closes the Gap.
Inside the Agentic SOC Workshop that trained 50 practitioners in a single session, and what it reveals about hands-on learning as the new AI growth engine.
Walk the floor of any developer conference in 2026 and you'll spot the same word on every booth: agentic.
Keith Manville, a Pre-Sales Engineer at Google Cloud Security, put it bluntly when we asked him about it:
"You can't pass a booth without seeing Agentic or AI. If you're not using it, you're definitely gonna be left behind."
He's right. But there's a quieter problem underneath all the hype, one that doesn't show up in launch posts or product demos. It's the gap between AI features shipped and AI features adopted. And it's where most of the revenue in this category is being left on the table.
The new operational challenge
AI is the greatest accelerant the software industry has ever seen. It's compressing development cycles, expanding product capabilities, and raising customer expectations all at once.
But every release widens the gap between what gets shipped and what customers can actually adopt. According to SlashData's 2026 State of Developer Adoption Report, 92% of practitioners face at least one significant adoption challenge. The top three causes point to an operational problem, not a product one:
- Misalignment across teams (27%)
- Technology complexity (26%)
- Keeping content accurate and up to date (25%)
The companies that win the AI era won't just ship the best products. They'll be the ones who help their customers actually use them.
Why hands-on learning is the new growth engine
The companies winning with developer-led growth aren't gatekeeping their technology. They're giving it away through open, hands-on education and building massive acquisition flywheels in the process.
This isn't a marketing tactic. It's an operational strategy:
- MongoDB Atlas revenue grew 32% YoY in Q1 FY2025, with Atlas now representing 70% of total revenue. Their flywheel starts with free, hands-on MongoDB University courses.
- Elastic's cloud revenue grew 25% to 30% YoY through FY2025, driven by Elastic Training, their AI Playground, and free self-paced labs.
- Google itself has invested heavily in hands-on labs and sandbox environments because developers who learn by doing become production customers, not just free-tier users.
Hands-on education isn't a cost center. It's the most scalable pipeline engine in the business.
Inside Google Cloud Security's Agentic SOC Workshop
At Google Cloud Next, Manville's team ran what they call the Agentic SOC Experience workshop. The goal: bring customers into Google's platform, immerse them in the company's agentic capabilities, and have them walk away knowing how to operate the tech.
The result?
"Yesterday, we had 50 people in the room here doing an Instruqt workshop where they build their own agent," Manville said. "Those participants walked away with understanding what an agent is, what are those primitive features that we build into an agent, and how to use them."
That's 50 practitioners, in a single sitting, going from "I've heard of agents" to "I've built one." No installs, no infrastructure setup, no friction.
For Manville, that frictionless path is the whole point. He defines success this way: "Ease of use for the customer. They can use the product without any friction. Provisioning is automated. User access is automated. All that stuff. Demo data. And so really, that's how I get our customers in the product just as quick as possible."
What makes hands-on work for AI specifically
Manville's clearest articulation of why Instruqt works for AI workloads:
"On the AI front, Instruqt is fantastic because it allows us to spin up a dedicated Vertex environment per participant. And that gives them the ability to experience the agentic capabilities, build their own agents, leverage those AI capabilities, but in their own secure little sandbox."
Translate that into operational terms:
- Per-participant environments. Every learner gets their own isolated Vertex instance, not a shared sandbox where one person's mistake breaks everyone else's session.
- Automated provisioning. No IT tickets, no installer wizards, no "your account doesn't have permissions." Just a browser tab and a real environment, ready to go.
- Real tools, real APIs, real data. Participants don't simulate building an agent. They build one. Then they break it. Then they fix it. That's how skills compound.
- Global scale. Manville's team runs Capture the Flag events and workshops worldwide. "Instruqt helps us provision the Capture the Flag environment, it helps us provision our software, the demo data, ultimately helps us scale these out across the globe."
Same content, same experience, same outcome, without flying engineers to every region.
AI amplifies expertise. Hands-on learning builds it.
Here's the insight Manville keeps coming back to: AI is complex, and "if you've never seen it before, you really don't know where to click or what to do."
Passive content like slide decks, videos, and even great documentation doesn't get learners over that hump. You have to do it.
That's the part of the adoption story most companies miss. AI is a force multiplier, but only for people who already know what they're doing. Practitioners who treat AI as a shortcut to expertise are building on a fragile foundation. The ones who use AI on top of real, practiced skills are the ones who compound fastest.
"In order for those practitioners to up-level and accelerate their own careers and journeys and really just to do their job better," Manville said, "they need to embrace AI and leverage all of the goodness that it brings to our world. And honestly, I think building these workshops and using the Instruqt environment helps us achieve that."
That's the operational adoption strategy in a sentence: build the workshop, run it on Instruqt, send participants home with real skills.
How the world's leading AI companies are closing the gap
MongoDB. Elastic. Google. SUSE. The companies building the backbone of the AI economy aren't using slide decks or passive content to turn developers into confident users.
They're using Instruqt to deliver hands-on, guided experiences that accelerate adoption, reduce churn, and generate qualified pipeline at scale. Learners arrive at sales conversations already knowing the platform, and already confident in it.
That's not a content strategy. It's an operational adoption motion that aligns marketing, sales, and customer education around a single, scalable experience.
Where to start
Manville's parting advice was, fittingly, the simplest possible: "Just start using AI. With Instruqt, I use AI to help me build labs. I use AI in my labs. That's my journey."
If you're a director of developer relations, sales engineering, technical marketing, or customer education watching MongoDB and Elastic build massive communities through open, hands-on learning, and wondering why your team isn't doing the same, that's the right place to start.
The companies that will lead the AI economy aren't just the ones with the most capable models. They're the ones who get those capabilities into the hands of their customers fastest, and make sure those customers can actually use them.
See how leading AI companies use Instruqt to close the adoption gap.
