Software has an owner the day you buy it; AI adoption has one only if you name it. At a 10-person commercial real estate firm nobody’s title is “AI,” there is no enablement team, and the principal is usually also the top-producing broker with no spare hours. So the question that actually decides whether the tools stick is not which assistant you pay for. It is who, by name, is responsible for making the firm fluent and keeping it that way. The national numbers show what happens when the answer is nobody: the NAR 2025 Technology Survey found 68% of Realtors now use AI, but only 17% report a significant positive business impact. Most firms have the tool. Far fewer have an owner. This is a framework for filling that role at a firm too small to have a department for it.
Before the framework, a word on where this sits. The full 90-day arc of getting a lean team fluent is laid out in our CRE AI training playbook, and the broader case for why small shops can out-operate larger competitors is in the small CRE firm manifesto. This piece answers one question those plans depend on and most rollout advice skips: who, specifically, carries them.
Adoption Has an Owner, Even When No One Named One
Every rollout has an owner. When you do not name one, the default owner is diffusion, and diffusion loses. The tool gets a burst of attention after someone reads an article or sits through a demo, three people try it, one keeps going, and within a month the firm is exactly where it started with a paid subscription nobody cancels. That is not a technology failure. It is an accountability vacuum.
At a large company the vacuum gets filled by a title: a head of enablement, a center of excellence, a network of champions across business units. None of that is available to you, and copying it wholesale is a waste of time you do not have. What a 10-person firm needs is not an org chart. It needs two named responsibilities placed on real people, sized so they fit around the revenue work everyone is already doing. The rest of this framework is those two roles and the rules for filling them.
Owner vs. Champion: Two Roles, Rarely the Same Person
The first and most-skipped distinction is that owning the outcome and owning the momentum are different jobs, and at a small firm they usually belong to two different people.
The owner is the principal. They own the outcome, the budget, and the air cover. They decide that the firm is going to become fluent, they fund whatever training or tools that takes, and they make it visibly safe for people to spend time learning during a busy quarter. The owner does not need to be the most skilled user in the building. They need to be accountable for the result and willing to say, in a meeting, that this matters and that time spent on it is not time stolen from deals.
The champion owns the momentum and the know-how. This is the person who actually knows how to get a clean lease summary out of an assistant, who keeps the shared prompt library current, who the associate turns to when a market write-up comes back wrong. The champion carries the day-to-day, but they cannot carry it without the owner’s air cover, and the owner cannot produce fluency without the champion’s hands-on work.
Conflating the two is the most common way a small-firm rollout stalls. When the principal appoints a champion and then disappears, the champion has responsibility with no authority and the effort quietly dies. When the principal tries to be both and has no time for the hands-on part, nothing gets built. Name both roles, out loud, to two people. If your firm is small enough that the principal genuinely must be the champion too, at least be honest that it is two jobs sharing one calendar, and protect the time accordingly.
Why Your Most Enthusiastic Person Is Often the Wrong Champion
The standard advice is to pick your most enthusiastic, tech-comfortable person. In commercial real estate that instinct backfires more often than it works.
Two reasons. First, the loudest early adopter’s endorsement is already discounted by the room. Everyone knows they will use whatever gadget lands on their desk, so their praise reads as personality, not evidence, which is the same dynamic that makes converting a skeptic so much more powerful than pleasing a fan. Second, technical comfort is not the scarce skill here. The scarce skill is credibility with the people who are hard to move, usually the senior brokers whose habits define how the firm actually works. A tech-savvy junior can drive the software all day and still be unable to change how a 20-year producer writes an LOI, because the producer does not take direction from them.
Pick for credibility and judgment first, fluency second. The right champion is someone the skeptics already trust, who has the standing to sit next to a senior broker and say “try it this way,” and who has the judgment to know when an AI output is confidently wrong. Fluency can be trained in weeks; standing cannot. Sometimes the best champion is a respected ops director or a mid-career broker with good instincts, not the person who signed up for the beta first. The related trap of assuming “trained” means “certified in a tool” rather than genuinely capable is one we take apart in our piece on what “trained staff” actually means.
The Champion Is a Slice of a Week, Not a New Hire
The single biggest reason small firms never appoint a champion is that they imagine it as a headcount they cannot afford. It is not a hire and it is not a full-time job. It is a named, protected slice of one person’s week, on the order of two to four hours, that everyone in the firm knows is real.
The “protected” part is the whole game. If the champion role is unofficial, it is the first thing that vanishes the week a deal heats up, and an unfunded role produces exactly the diffusion failure you were trying to avoid. So the owner’s job is to make the time legitimate: block it, name it in a team meeting, and treat the shared prompt library and the weekly help the champion provides as firm work, not a hobby. A few hours a week, defended, compounds into firm-wide fluency over a quarter. The same hours, treated as optional, produce nothing.
This is also why the champion cannot be someone with zero slack in their schedule. A top producer at 110% capacity is the wrong pick not because they lack the skill but because there is no hour to protect. Choose someone whose week can absorb the slice without the firm pretending it costs nothing.
The Champion’s First Real Job Is the Data Rule
Most champion playbooks open with evangelism. In commercial real estate the first real job is governance, and it is not optional. Your firm handles confidential deal terms, client financials, and material under NDA, and the fastest way to hand a skeptic a permanent veto is to let someone paste a live deal into a consumer chat window before anyone has written down what is allowed.
So the champion’s first deliverable is one page: which tools are approved, what data is fine to use (public listings, generic prompts, hypotheticals), and what never leaves the firm (client names tied to financials, unexecuted terms, anything under NDA). Default the firm to business-tier accounts, whose terms keep your inputs out of model training, rather than personal free logins, and confirm the current settings for whichever assistant you approve, since vendors change them. This is enough of a discipline that it is worth treating as its own small project rather than an afterthought. Getting it wrong early poisons the whole effort; getting it right gives the careful people on your team a reason to relax. The full playbook for rolling out to the doubters, data rule included, is in our 10 rules for rolling out AI to a skeptical team.
What the Champion Actually Does, Week to Week
Strip away the title and the champion role is four recurring jobs, none of them glamorous.
They keep the shared prompt library alive. When a broker lands on a prompt that reliably produces a solid first-draft LOI or a usable market blurb, the champion captures it where everyone can grab it. This is what keeps hard-won know-how from evaporating when the person who figured it out gets busy. A firm without a shared library relearns the same prompt ten times.
They run the help desk. Not a formal one, just the known answer to “who do I ask.” A ten-minute unblock from someone credible is the difference between a colleague who keeps going and one who quietly reverts to the old way.
They rig and stage the wins. The champion picks the safe first tasks for each person, the high-annoyance, low-stakes, low-confidentiality work where the tool obviously helps, and saves the high-stakes documents for after trust is earned. Getting the first task right is most of adoption.
They measure behavior, not attendance. The champion watches weekly active use of the firm’s main tool and the time it takes to produce a fixed set of real jobs before and after, because that is the only honest read on whether anything changed. The gap the industry keeps reporting, most firms using AI but few seeing real impact, closes here or not at all. What non-technical teams actually retain from training, and how to reinforce it, is the subject of our lessons from a dozen AI workshops.
Build Toward Redundancy, Not a Guru
There is a failure mode that looks like success: the champion becomes so good and so central that the firm routes everything through them, and a single person quietly becomes the only one who can get real work out of the tools. That is not adoption. It is a bottleneck wearing a badge.
The goal of the role is to make the role smaller over time. A champion who is doing the job well is spreading fluency, not hoarding it, and the sign of progress is that people stop needing to ask. The shared library, the staged first tasks, the visible human-check habit, these all exist so that capability lives in the firm rather than in one head. Measure the champion partly on how independent everyone else is becoming. A firm that would collapse back to zero if the champion took a two-week vacation has built a dependency, not a capability.
Succession: When the Champion Walks Out the Door
Because a small-firm champion is one person carrying know-how in their head, the role has a specific fragility that enterprises do not: your champion can leave, and take the firm’s fluency with them. This is not hypothetical at a 10-person shop where people change jobs.
The defense is the same discipline that prevents the guru problem. If the prompt library is shared and current, if the data rule is written down, and if a second person has been shadowing the role even loosely, a departure is a setback, not a reset. Name a backup early, even informally. The point of writing things down was never documentation for its own sake; it is so the capability survives the person. A champion who has genuinely worked toward redundancy is, conveniently, also the easiest to replace.
Where a Workshop Hands Off to Your Champion
Outside training and an internal champion are not competing options; they are a handoff. A structured LLM-fluency workshop is the fastest way to get a whole team from zero to a real baseline in days rather than months, and it gives your champion a running start instead of asking them to invent the curriculum. Market rates for that kind of training run from the low thousands to the low tens of thousands, depending on team size and depth.
But a workshop is an event, and adoption is a habit. The day after the training, the champion is the person who keeps the momentum from decaying, reinforces the safe first tasks, maintains the library, and answers the questions that surface once people try the tools on their own real deals. Firms that treat the workshop as the finish line drift back toward the 17%-see-impact average. Firms that treat it as the start, with a named champion to carry it, are the ones where the training actually pays off. If your firm is standing on custom-built automation rather than off-the-shelf assistants, the champion becomes even more important, because someone has to own the workflow the tools now depend on.
Frequently Asked Questions
Who should own AI adoption in a small real estate firm?
Two named people, not one. The principal or owner owns the outcome, the budget, and the air cover: they decide the firm will become fluent and make it safe to spend time learning. A separate champion owns the day-to-day momentum and know-how: the shared prompt library, the informal help desk, the safe first tasks. At a very small firm the principal may have to do both, but it should still be treated as two distinct jobs sharing one calendar. The failure mode is naming a champion and then having the owner disappear, which leaves responsibility with no authority behind it.
What is an AI champion?
The AI champion is the person inside your firm who owns the momentum of adoption: keeping the shared prompt library current, being the known answer to “who do I ask,” staging safe first tasks for colleagues, and tracking whether real work actually changed. In a small commercial real estate firm it is not a new hire or a full-time job; it is a named, protected slice of one person’s week, on the order of two to four hours. The champion carries the hands-on work that a principal’s endorsement alone cannot produce.
Should the AI champion be the owner or principal?
Usually not, and separating the roles is deliberate. The principal should own the outcome, the budget, and the visible signal that this matters, but the hands-on champion work, maintaining prompts, unblocking colleagues, staging tasks, takes time a top-producing principal rarely has. If the principal tries to be both and has no hours for the hands-on part, nothing gets built. Appoint a champion who has the credibility to influence senior brokers and the schedule slack to protect a few hours a week, and keep the principal as the owner who funds and defends the effort.
Is the most tech-savvy person the right AI champion?
Often not. Technical comfort is easy to train; credibility with the people who are hard to move is not. The most enthusiastic early adopter’s endorsement is already discounted by the room, and a tech-savvy junior usually lacks the standing to change how a senior broker works. Pick for credibility and judgment first, fluency second: someone the skeptics already trust, who can sit next to a 20-year producer and be listened to, and who can tell when an AI output is confidently wrong. A respected ops director or mid-career broker frequently beats the person who signed up for the beta first.
How much time does the champion role take?
Roughly two to four hours a week, protected and named, not a full-time seat. The critical part is that the time is legitimate: blocked on the calendar and acknowledged in a team meeting as real firm work, not a side hobby. An unofficial champion role is the first thing to vanish when a deal heats up, which reproduces the exact diffusion failure you were trying to prevent. This is also why the champion should not be your most over-capacity producer; you need someone whose week can actually absorb the slice.
Do we need to hire someone to be the AI champion?
No. The champion is an existing person given a named, protected slice of their week, not a new headcount. Hiring for it is both unnecessary at a 10-person firm and usually counterproductive, because the role depends on internal credibility that an outside hire does not yet have. The tools most firms start with are business-tier accounts of assistants the team can already use, such as ChatGPT, Claude, Gemini, or Microsoft Copilot. Custom software and dedicated roles come much later, only once standardized workflows make hand-work the bottleneck.
What happens if our AI champion leaves?
At a small firm this is a real risk, since the champion carries know-how in their head. The defense is to make the role less person-dependent while it runs: keep the prompt library shared and current, write the confidential-data rule down, and have a backup loosely shadow the role from early on. Done that way, a departure is a setback rather than a full reset. A champion who has genuinely spread fluency instead of hoarding it, so the rest of the firm needs them less over time, is also the easiest to replace.
How does the champion role relate to an AI training workshop?
They are a handoff, not a choice. A structured LLM-fluency workshop gets a whole team to a real baseline in days and hands your champion a running start instead of a blank curriculum, with market rates typically in the low thousands to low tens of thousands. But a workshop is an event and adoption is a habit. The champion is who keeps the momentum from decaying afterward: reinforcing safe first tasks, maintaining the library, and answering the questions that surface once people use the tools on their own deals. Training without a champion tends to fade; a champion without training starts slower than they need to.
Where to Start
The first move is not buying a tool or scheduling a class. It is deciding, on paper, who owns the outcome and who owns the momentum, and whether the person you have in mind for champion has both the credibility and the schedule slack the role needs. That is exactly what a free AI-readiness assessment produces: a working session that identifies your highest-friction workflows, flags the confidential-data rules you need in place first, and gives you an honest read on whether you need a full LLM-fluency workshop, a lighter course, or just a named champion and clearer rules for the tools your team already has. The assessment costs nothing and usually makes the right owner and the right champion obvious. Book a free AI-readiness assessment and you will leave knowing who to name, and what to hand them first.
Arthur Wandzel
