Why Your AI Transformation Leaders Need to Be Gardeners

To cultivate the right conditions to realize value from your AI transformation initiatives, you have to be a gardener. I’m not talking about growing tomatoes here—though I’m not going to stop you from doing that, they are delicious—but a nurturing strategic mindset that will move your organization toward realizing value from AI and away from experimentation mode. Let’s dig into what this really means.

How We Started Our AI Transformation, and What We Could Have Done Better

Last year, our marketing team was enthusiastically experimenting with AI. They were testing new tools, exploring use cases, and building all kinds of custom GPTs and agents. Every week, there was a new “you gotta see this!” moment where someone applied AI in a novel way. There was no shortage of cool ideas, and some were seeing solid time savings and productivity improvement from the tools they used or built. 

We had the excitement and drive, but we lacked the cohesion and intentionality needed to take things to the next level.

There are two standard approaches for AI transformation: bottom-up and top-down. Most organizations gravitate toward one or the other, but neither works on its own.  

We started out in the bottom-up camp. We gave people access to tools, time, and space to plant their AI seeds, watch the flowers bloom, and trust that these ideas will continue to self-propagate. It was important work, and it generated a ton of ideas, but it was impossible for anyone to see the bigger picture and fundamentally rearchitect how work gets done within and across teams. 

Sure, one person may have built a brilliant workflow, but they are the only one who has it and knows how it was built. Others end up building similar things and duplicating efforts. And if that person leaves, the workflow withers away and dies. When you have one person tending to each plant who can’t see the entire farm, you end up wasting a lot of resources and underutilizing the produce.

The top-down approach is the more common corporate move. You buy the licenses, mandate their use, put everyone through some basic training module, and hope for the best. This will definitely lead to more AI usage, but the increase in token costs will likely outpace your yield by a significant margin.

Both approaches will ultimately fail because nobody at the top of the org—or anywhere else in it, for that matter—has sufficient visibility into how work is getting done to guide, implement, and improve these new AI processes. You’ve got a bunch of people planting seeds all over the place, with no idea how to keep them growing or profit from the crop.

This is where the gardener comes in. They pull both approaches together to oversee each plot and each crop, and pull it all together to make a functional and profitable farm.

How to Be a Gardener Instead of a Builder

The difference between a gardener and a builder is that a builder works from a specification to build a finished product. They source the parts, assemble them, and make sure the product works as specified. This works well if you’re building a deck, not so much an organization. 

A gardener doesn’t actually build anything; they hold the vision of what they’re growing, and create controlled conditions so it can thrive. They control where things get planted, how much light and water they get, say what gets pruned back and what gets staked so it doesn't collapse under its own weight, and what gets planted next to what so it cross-pollinates. 

That is a much more accurate description of the executive role in AI transformation than "build the AI strategy." Here’s what that looks like in practice. 

1. Plant something real, with a clear objective

Marketing leadership at Invoca pointed the AI Launch Lab at our annual benchmark report. It’s a high-visibility, deadline-bearing, revenue-relevant initiative that the team was going to have to deliver either way. This project had also been done before, so there was a control that enabled us to validate the results.
2. Allow time to grow 

We extended the report’s launch timeline to give everyone space to rethink how the work got done. The instinct with AI is to compress schedules and book the savings immediately, which is exactly how you guarantee nothing changes. Under a compressed timeline, a team will just run the process it already knows, maybe a little faster and definitely more anxiously. Rethinking a workflow,  pulling it apart, questioning every step, and asking whether a step should exist at all takes time. 

3. Intentionally cross-pollinate 

During the launch lab, the teams documented their prompts, workflows, and outcomes in shared repositories, including the attempts that didn't work. The failures turned out to be among the highest-value entries because they saved four other people from hitting the same dead end. Cross-pollination happens because someone built the place where knowledge goes and made contributing to it part of how the work gets done.

What we Grew in the AI Launch Lab

The results of the AI Launch Lab were a significant increase in our crop’s yield and shortened time to harvest. The marketing team would normally produce one report by the deadline and start their promotion campaign. The efforts of the AI Launch Lab enabled them to produce nine additional industry-specific versions, along with UK-localized editions, in the same timeline with the same headcount.
Alongside the content assets, the team also realized meaningful reductions in effort across campaign planning, project setup, and account analysis. The workflows the marketing department developed became the template that other parts of the company are now working from.

This project showed that we could do far more with the resources we already have. Reducing headcount was never a concern. It was all about speed, scale, and increasing impact with the same amount of effort. The promise of AI is that we can accomplish more valuable work than we ever could by ourselves. 

Great Gardens Don’t Grow Overnight

Gardening is slower than building and much less satisfying to announce. There's no launch date, operationalizing AI takes time, and we're still doing it, still refining workflows, still finding steps that shouldn't exist, and still writing down what didn't work. You don’t get the rush of all the instant wins you saw in the experimentation phase, but what you end up growing will feed you all season long.

Get the Forrester Case Study, Invoca Enables Organization-Wide AI Adoption Through Its AI Launch Lab, to learn more.