The planting guide is a part of a series of pieces in collaboration with Fabien Girardin on personal software: an introduction, the makers and the implications for organizations and teams.
These pieces are observations and analysis from the field: a map of the territory and an opening for the conversations that follow. Many of those conversations take place at Pre-Alpha Club: A private club for anyone who uses AI to build their own tools, rethink and reshape their practice, and fundamentally question what all of it changes. If that sounds like your kind of table, you are welcome to pull up a chair.
The organizations we described in Software Gets Personal For Organizations and Teams didn’t stumble into cultures of experimentation. What’s required of leaders here is less like engineering and more like gardening. You cannot mandate curiosity. You cannot impose enthusiasm on anyone. You can, however, prepare conditions where both are likely to take root. You have to cultivate such a culture.
What follows is a set of groundwork for leaders at every level, from a self-starting founder or team member to an executive shaping the future organization. This guide is for all of us stepping up to the challenge of delivering high performance work in perpetually shifting terrain.
Rooted in nature, as we all are, I share this Japanese proverb to begin, the core philosophy of wabi-sabi: “nothing lasts. nothing is perfect. nothing is ever finished.”
1. Promote permission
The most common reason we don’t experiment is that we’re not sure we’re allowed to. We sense that tinkering might be considered as a distraction by those who judge our work. Or, perhaps that building something unofficial might attract the wrong kind of attention. That sharing a half-baked idea might expose us to negative consequences. The silence of not leaping in to explore, in these cases, is most assuredly not laziness. It’s rational caution that sprouts in organizations where polish is rewarded over an off-road quest; and completion wins over curiosity.
The antidote is explicit, repeated public permission. Circulating a policy memo or dropping a one-time all-hands mention isn’t enough to unleash the bravery needed to unveil early green shoots. A pattern of signals—in how meetings are run, in what gets recognized, in what leaders share about their own fumbling attempts—makes it clear: trying things here is safe, even when, and maybe especially when, they don’t work. As we discussed in the Personal Software series, at BBVA in Madrid, Elena Alfaro didn’t hand out licenses with a compliance manual. She told people to dig in, play, learn and use it more. That instruction reads simple at first glance, but it’s precise and ultimately strategic.
Always attract the pollinators
2. Lead by going first
Permission from above means very little if the people granting it aren’t visibly doing the same thing themselves. Leaders who want curious, experimental teams need to take it on themselves to actually be curious, step out of their own comfort zones and explore hard questions, new issues and share the little ruptures these will certainly cause along the way.
This means sharing immature attempts well before they’re ready. It means describing what you tried, what didn’t work, what you learned and what you think you will do next. It means asking questions in public that you don’t already know the answer to. None of this comes naturally to most humans and leaders have been trained to project confidence and authority. Leading was meant to look like you’ve done this many times before. Planting new ideas, tools and seeing how they take shape means each time is new. This may not work the way we expect or want it to. The goal is the learning and building new muscles and the trust to fall down together and grow. Teams take their cues from what they observe, not from what they’re told. A leader who tinkers, shares, falls and learns in plain view gives everyone around them the courage and permission to do the same.
3. Make Yes! the default
Big experiments require approvals, budget, coordination, big promises and a business case. Tiny experiments need almost nothing. They can happen everywhere, constantly, and perhaps most importantly, without asking anyone. The organizations that generate the most learning aren’t the ones with the grandest innovation programs. They are where small experiments are structurally cheap. They tend to feature:
low approval friction
no requirement to formally document failure
zero penalty for trying something that fades, fails or didn’t get completed
Practically, this means protecting small amounts of time for exploration as a cultural norm. It means not demanding ROI projections from someone who wants to spend an afternoon trying something they’re inspired to build. It means recognizing that the cost of a failed two-hour experiment is two hours, and the cost of suppressing creative freedom and innovation, while harder to measure, can lead to lost insights or worse, talent migration. Over time, the cost of both is remarkably high. The richness of the team culture—the soil you’re planting in—acts as a far more effective talent magnet than compensation packages.
Protect the early sprouts
4. Seed sharing
People share when sharing feels safe and ideally, highly valued. Both conditions need to be actively cultivated. Safe sharing means no one gets mocked for an idea that didn’t pan out, no one gets scooped by a peer who takes credit, no one who raises a problem gets handed ownership of it as punishment. These dynamics are common enough that people learn quickly to keep their observations, insights and experiments private. Once that pattern sets in, the organization loses the connective tissue that turns individual learning into a collective capability that generates actionable foresight.
Show your work—an ugly prototype, a half-solved workflow, a gnarly question that hooked you—sharing any of these can make you feel either vulnerable or ballsy. Which will depend on how your team is tuned. Seeding ideas, trying new techniques, thinking together about why something didn't work as expected—all of it takes trust in each other and in leadership. From year to year, season to season, weather, seeds, soil quality change and likewise, so do the plants, their placement and yields. When learning and innovation are the goal, all outcomes offer insights and will get recognized as valuable contributions. Not necessarily with fanfare, but with genuine acknowledgment and, sometimes, with deeper exploration.
That feedback loop is what turns individual provocations and experimentation into a culture of collective learning. It is what creates, what Boundaryless’ Simone Cicero calls an organizational learning engine. For a world in constant shift, crafting and feeding an organizational learning engine is essential for survival. BBVA’s internal GPT Store works because it activates this. It makes sharing the default path, gives the sharer visibility and credit, and lets others build on rather than duplicate the work. The mechanism matters less than the principle: sharing should feel like adding to a communal creation, not personal exposure.
5. Adore bold questions
Curious cultures are built on wild questions. Most organizational systems and compensation are structured to reward answers—to move quickly from problem to resolution, to treat open questions as uncomfortably unfinished business rather than precious exploration. This creates a subtle but powerful bias against the kind of open-ended wondering (and counterpart unsettling feeling) that precedes real discovery.
Leaders can push back against this by explicitly protecting time and space for questions that don’t yet have answers. By treating “I don’t know, but I’d like to find out” as a deeply respected and perfect response. By resisting the pressure to close every inquiry before it has been properly explored. The Latin word quaestio—question, inquiry, investigation—shares its root with quest. The journey is part of the point.
Value arises from surprising places
6. Govern as a gardener, not a warden
Risks, we’re told, are meant to be weeded out. The instinct to protect an organization from risk is legitimate. A common mistake is prematurely thwarting early sprouts from a fear that they may be harmful. A garden that gets weeded before anything has had time to grow yields little or nothing.
Tending your garden, repositioning plants, supporting the soil with fresh input and noting surprises, are all central activities at the core of creating a space that welcomes growth and thrives through change. Governance that arrives before experiments have had a chance to surface their value suffocates the signal before it can be registered. The tiered model BBVA uses—govern lightly at the edges, more carefully at the center, most carefully where stakes are highest—is a gardener’s instinct applied to an organizational system and model. Not everything needs the same attention. Not everything needs attention now. The mix of experience and discipline is in knowing which things to tend closely and which to leave alone long enough to see what they can become.
7. Plant companions
Farmers and hobby gardeners have planted the Three Sisters together for centuries: corn, beans, squash. The beans climb the corn. The squash sprawls low, holding moisture in and keeping weeds out. The beans, meanwhile, are feeding the plant community, fixing nitrogen into the soil. Nobody assigned these roles. This coupling is the design.
Years ago at Barclays, no project got a green light—even a license to play—without three people at the table: a designer, an engineer, and someone channeling the customer. A little triangle. Three views of the same patch of ground.
Most organizations plant rows of the same thing. Ten analysts on a problem will hand you ten analyses. It feels efficient, and the soil gets tired fast. The fertile and resilient pairings are usually the odd ones—the compliance person sitting with the designer, the new hire asking the question the veterans stopped asking years ago. Nobody has to explain in advance what the beans are doing for the corn. Put them in the same bed and watch.
So before you fund the idea, look at who's growing around it.
Trees can't run. When pests arrive, a forest has no exit, so forests built something else: the mycelial network underfoot, the informal web that passes warnings, shares nutrition, and mounts a community immune response. Suzanne Simard calls the elders at its center mother trees. Your organization has an equivalent—the informal structure through which insights and hard-won learning travel between teams. You didn't design it and you can't mandate it. You can feed it.
For the self-starting team member who doesn’t yet have institutional backing: you don’t need permission to notice a problem, to try something small, to share what you learned. The planting guide applies to you too. Your patch of ground is smaller. But the same principles hold: start small, share early and often, treat failure as data and data as learning. Rinse and repeat., Change the frame and keep going and growing.
All of these practices rest on a single belief that can’t be faked: people with their hands in the dirt, those closest to the issues and the work are the ones most worth investing in.
These people are enthusiastic users of tools. They also tend to be wonderful observers, shapers, makers—of ideas, of experiments, fearless creators, of real insights. Organizations that believe this think, build and grow differently. Their culture is the substrate that welcomes regenerative ideas, products, teams and organizational models. They learn differently and together. And when new capabilities arrive—as they are now—they are poised in a way that no procurement strategy, no digital transformation project, and no AI roadmap can rightly anticipate.
The soil is prepared. The seeds are already in the ground. They are waiting for someone to let the energy of the sun and bees do the magic. As leaders, our central job is to welcome the pollinators and stop paving over the tiny fresh buds!
If you have seen your organization in these pages—whether at the crossroads or already experimenting—we want to hear your questions, your experiences, and the examples you are seeing in your own context. Comment or reach out!