Every revenue leader was asked the same question last year: “What’s our AI strategy?”

You answered it. Piloted. Rolled it out across sales, enablement, and CS.

Now we’re all dealing with the reality: nobody’s using the tools. Or worse, they’re using them and the output is too generic to make a revenue impact.

This is the same pattern that caused 42% of companies to abandon most AI initiatives in 2025. Teams skip the work that makes adoption happen and the result is a wasted investment. But the board still expects a working AI strategy from us, and their oversight is intensifying.

Jody Geiger, co-founder of AI Sales Studio, explained the process most enterprise teams skip on our latest GTM Science podcast episode.

This is the same AI implementation process she used at Klue to reduce demo preparation time from 8 hours to 2 minutes. All without losing the personalization that actually won their deals.

In this newsletter, we’re breaking down the key aspects of Jody’s AI implementation process. Listen to the full episode on Spotify here or Apple Podcasts here.

Why AI Implementations Fail

The pattern looks the same every time.

Buy the AI tool. Implement it. It doesn’t “work”. Blame the tool.

But the tool isn’t the problem.

“I think that’s on us as leaders,” Jody says. “We haven’t spent the time to do the mapping, the auditing and the planning for change and adoption and testing, let alone setting up the iterative process of experimentation and that learning loop. We’re missing it.”

AI automates busy work. But it doesn’t automate the secret sauce that wins deals.

“We have to take the time to deeply understand our customers. We have to understand their situations. So we can know ‘why now’ and ‘why us’,” Jody says. “We need to find out what those unscalable things are that we do, and only WE do, and then we have to figure out a way to make those things scalable. So they don’t break the bank as we try to grow.”

Jody’s AI implementation process forces this understanding by:

• Mapping JTBD to see which ones are creating impact
• Identifying what needs human expertise and what’s busywork
• Testing different approaches to find the one that preserves results
• Creating feedback loops that drive adoption

The teams chasing speed are betting that AI will replace thinking. The winning teams use AI to create space for thinking. Here’s what that looks like in practice, according to Jody.

The AI Implementation Process That Works

Jody’s implementation pattern isn’t theoretical. She’s used it throughout her career to transform major bottlenecks, some of which she shared with us on the podcast. For this newsletter, we’ll use her demo preparation story to illustrate what each step of the process can look like.

Step 1: Map the Current Reality

“We have to start in a place where we understand both the people that are doing the jobs and every click required to do that job at a high level. Can we understand and audit that?” Jody explains.

By understanding the actual work in detail, we can make intentional decisions based on where AI will have the most impact.

For demo prep at Klue, the ground-level reality was:

• Teams spent 8 hours creating each demo space
• Customizations included competitive intel for prospects
• Experts who deeply understood competitive positioning were doing manual work
• When demos were customized, deals moved faster and won more often
• But the CAC was terrible because they weren’t winning all those customers

From this amount of detail alone, we can already see two things:

1. There’s great potential here for AI to make a real revenue impact
2. There are obvious AI-capable tasks that don’t take away from the value

Step 2: Break Down the Workflow, Find the Busywork

Once you’ve mapped the current state, break it down.

“We basically started by looking at every box in that old workflow as a step in the process and asked questions. Like, what can we cut? What can we automate? What actually needs AI or is a good prospect for replacement with AI?”

For demo prep, this meant asking:

• What parts of the 8-hour process actually require human expertise?
• What’s just data entry and formatting?
• Where does the personalization actually matter to prospects?
• What could be templated without losing impact?

“That detailed workflow audit led to a completely reimagined flow.”

Step 3: Test Multiple Approaches 

Don’t pick one solution and hope it works. Run experiments.

Instead of choosing one approach and going all-in, Klue tested three for about a month concurrently:

• AI Ops Team (one person and an intern): Build automation using workflow tools
• Enablement Team: Review off-the-shelf demo automation tools and cost them out
• Product Team: Stop customizing at the individual company level, build vertical demo spaces instead

AI Ops won the horse race. They built an AI workflow that auto-generated demo spaces, trading in 8 precious hours of manual work done by experts for 2 minutes and 75 cents per token.

The results:

• Deal velocity improved
• Deal capacity improved
• CAC dropped dramatically with minimum tooling cost
• Their competitive positioning experts were free to do other tasks
• AEs gained instant access to personalized demo spaces when they needed them

“I think that was the experiment that really got me thinking. What does a systemic go-to-market engineering mindset and approach look like?”

Step 4: Get Feedback From the People Who Will Use It

This is where adoption actually happens.

“Our teams have spent so much time one-on-one with people who were doing the work. Getting feedback on: Is this the right output? What would we change? Where do we need to feel more human? Where do we need to add more context? Where do we need to bring more of the buyer’s situational understanding in? Where do we need to improve our value prop? Where can we bring in customer language?”

For the demo automation, this meant working with:

• AEs who would use the demo spaces in actual sales calls
• Enablement to ensure the competitive positioning was accurate
• Prospects (indirectly) to validate the demos still felt personalized
• The experts who had been building demos manually to understand what mattered

When everyone’s involved in designing and testing, adoption isn’t a problem and optimization comes naturally.

“Now the whole team, across the go-to-market motion and customer success, has buy-in to the process.”

Find the Unscalable, Magical Thing. Then Scale It.

The demo prep story illustrates a critical principle: they didn’t start with automation.

They started by understanding why customized demos won more deals. What made them magical.

“You bring in process once something is working. Once a flywheel is in place,” Jody says. “And that’s how you make creativity repeatable and dependable.”

The unscalable thing that made Klue’s demos magical: prospects could see their actual competitors with accurate competitive positioning. They experienced what the tool would actually feel like with their data.

Only after identifying that did they ask: how do we make this scalable?

“We have to take the time to deeply understand our customers. We have to understand their situations. So we can know ‘why now’ and ‘why us’,” Jody says. “We need to find out what those unscalable things are that we do, and only WE do, and then we have to figure out a way to make those things scalable. So they don’t break the bank as we try to grow.”

Jody’s AI implementation process forces this understanding. You can’t map every click without seeing what matters. You can’t audit the workflow without identifying what’s valuable versus what’s waste. You can’t test approaches without knowing what you’re trying to preserve.

“I see so many teams chasing speed before trust. We have to move fast, but really, these winning go-to-market orgs are moving fast because they’re grounding themselves in understanding, yes from data, but also from connection and from curiosity.”

Interested in Doing This For Your Org?

But unsure where to start or how to fit it with the resources you have? We help Enterprise revenue leaders navigate and execute on operational process design and systems architecture, just like the team at Klue did. Start the conversation here.

Ready to Grow Your Revenue?

Bring us on to help with strategy and/or execution of your GTM Ops and RevOps projects.

Book a Strategy Call