Every executive team believes it can see its own gaps. Most can’t, not entirely, and the ones who lead best are usually the first to admit it. That admission is the actual starting point for building an AI team, not the org chart, not the job postings, not the tool licenses. It’s the recognition that the people currently running the business, however experienced, are operating with a limited view of what AI and data capability actually requires.
Donna Medeiros, VP of AI and Data Advisory at Data Society, has built her advisory practice around that recognition. Her argument isn’t that leaders are bad at their jobs. It’s that even good leaders have structural blind spots, and building a real AI team starts with naming those blind spots honestly before trying to hire or train your way past them.
Even Seasoned Leaders Have Blind Spots
There’s a certain executive instinct to treat any admission of a gap as a weakness. Donna pushes back on that instinct directly.
“There’s always blind spots. I think seasoned leaders know this, that there’s blind spots they have.”
This isn’t a criticism of leadership. It’s closer to a description of how expertise works. The deeper someone goes in one domain, running operations, owning a P&L, scaling a sales org, the more likely it is that AI and data capability sits outside their direct experience. That’s not a character flaw. It’s a natural consequence of where their career actually took them. The mistake isn’t having the blind spot. The mistake is building a team, or worse, a strategy, without first acknowledging where it is.
This is why the 2025 AI Readiness Report consistently finds a gap between how confident leadership feels about AI readiness and how prepared the organization actually is on the ground. The confidence and the readiness are measuring two different things, and closing that gap requires an honest look in the mirror before it requires a new hire: https://datasociety.com/the-2025-ai-readiness-report-insights-to-build-your-2026-strategy/
Why Outside Advisors Speed Things Up
Once a leadership team accepts that a blind spot exists, the next question is how to close it fastest. Donna’s answer is direct: don’t try to do it alone, and don’t assume internal trial and error is the cheapest path.
“Advisors accelerate the time to value for executives, give them, build their knowledge, give them more arsenal in terms of AI and data expertise when they may not have that bench currently.”
The phrase “bench currently” is doing real work here. Most organizations, even sophisticated ones, don’t have deep internal AI and data expertise sitting on the bench ready to be deployed. They have pockets of enthusiasm, a few technically fluent employees, and a leadership team trying to make strategic calls without a peer group to stress-test those calls against. An advisor doesn’t replace the need to build that bench. An advisor compresses the time it takes to build it, by bringing pattern recognition from other organizations, honest pushback on flawed assumptions, and a faster route to the decisions that matter.
This is the thinking behind Data Society’s own move to stand up dedicated AI advisory services built specifically around outcomes rather than generic strategy decks. The value isn’t the advice in the abstract. It’s the speed at which that advice turns into a decision the organization can actually execute: https://datasociety.com/data-society-launches-ai-advisory-services-to-support-responsible-outcomes-driven-ai-adoption/
A Tool Rollout Is Not a Team
Handing employees a login is often mistaken for building a capability. Donna is blunt about why that mistake is so costly.
“You can’t sit there and release a tool and expect folks to just adopt it unless they’re convinced it’s going to help them in their job, probably help them in their career.”
This is the piece that separates a genuine AI team from a headcount of AI tool users. Adoption isn’t a technical event, it’s a psychological one. Employees make a quiet calculation before they change how they work: does this help me do my job better, and does it help me where my career is going? If the answer to either is unclear, the tool sits unused no matter how good the license is.
Building a team from scratch means designing for that calculation from day one, not discovering it after adoption numbers come in flat. That means role-specific framing of why a capability matters, not a company-wide email announcing a new platform. It means showing someone in finance, in HR, in operations, exactly how the tool changes their actual workflow and their actual career trajectory, not a generic use case that could apply to anyone and therefore convinces no one.
Training Has to Be Built, Not Bought Off a Shelf
This is where Donna’s advisory work and Data Society’s training philosophy converge directly. Generic, self-paced course content doesn’t build a team. It builds a completion certificate.
“That’s why Data Society has the capstone project based training, person led training.”
The logic here is straightforward once you sit with it. A team isn’t built by exposure to material. It’s built by doing real work, under real constraints, with a person guiding the process who can correct course in real time. A capstone project forces the learning to attach itself to an actual deliverable tied to the business, which is the same mechanism that makes a hire valuable in the first place: not what they know, but what they can produce with what they know. Organizations that map training to specific roles rather than generic tool access see very different outcomes, a pattern covered in depth in Data Society’s enterprise AI upskilling guide and echoed in the broader case for treating upskilling and implementation as two connected efforts rather than sequential, disconnected ones: https://datasociety.com/enterprise-ai-upskilling-guide-2026/ and https://datasociety.com/ai-upskilling-vs-ai-implementation/
Governance and Fluency Are Built Together, Not in Sequence
Perhaps the most important structural point Donna makes is about sequencing. Leaders often treat governance as the thing you add once the team is already fluent. She argues the opposite.
“Fluency in AI can’t happen without the governance. It can’t happen without the skills enablement.”
Fluency isn’t a byproduct of comfort with a tool. It’s a byproduct of knowing the guardrails well enough to move confidently inside them. A team that has skills but no governance framework is a liability waiting to surface. A team that has governance but no skills enablement is paralyzed, following rules it doesn’t understand the purpose of. Real fluency, the kind that lets a team move fast and stay accountable, requires both built in from the start, which is a big part of why organizations that separate governance conversations from training conversations tend to stall out on both. The same discipline shows up in workforce initiatives like the Data Leadership Collaborative, where impact comes from training that’s connected to real accountability structures, not standalone content: https://datasociety.com/data-leadership-collaborative-five-ways-to-create-impact-from-workforce-data-training/
What Building the Team Actually Looks Like
Put together, Donna’s guidance suggests a build sequence that looks less like traditional hiring and more like a deliberate capability build:
Name the blind spots honestly before writing a single job description.
Bring in outside advisory expertise early to compress the learning curve, not after the strategy is already locked.
Design tool adoption around what convinces an employee it helps their actual job and career, not a company-wide mandate.
Replace generic course access with capstone, project-based training tied to real deliverables.
Build governance and skills enablement together from day one, not governance as an afterthought once the team feels fluent.
A team built this way doesn’t just use AI. It can defend its decisions, explain its process, and improve on both over time, which is the actual definition of bench strength.
Frequently Asked Questions
Start by identifying leadership blind spots honestly, bring in outside advisory expertise to accelerate time to value, design tool adoption around real job and career benefits for employees, use project-based training tied to actual deliverables, and build governance alongside skills enablement rather than after it.
Deep expertise in one domain, like operations or sales, doesn’t automatically translate into AI and data fluency. Blind spots form naturally because a leader’s career path didn’t require that specific expertise, not because of a leadership failure.
An AI advisor accelerates time to value by bringing outside pattern recognition, honest pushback, and faster access to sound decisions, helping executives build knowledge and capability they may not currently have on their internal bench.
Employees adopt tools when they’re convinced the tool will help them do their job better and support their career, not simply because the tool was released. Adoption is a psychological decision, not a technical one.
Capstone project-based training is person-led training where employees apply AI skills to a real deliverable tied to their actual role, rather than completing generic, self-paced course content disconnected from their day-to-day work.
Build Your Bench With the Right Guidance
If your leadership team suspects it has blind spots but hasn’t named them yet, that’s the exact starting point for building a capable AI team instead of a collection of tool users. Book time with Donna to map out where your organization’s bench strength needs to grow first: https://meetings.hubspot.com/donna-medeiros/meet-with-data-societys-ai-and-data-advisor
