The biggest AI mistake real estate companies are making has nothing to do with choosing the wrong technology.
It happens earlier, when leaders decide they need AI before deciding what they actually need AI to do.
That mistake becomes far more amplified with agentic AI: technology that can take actions on a user’s behalf.
Give a person the wrong permission, and the damage may be limited by how quickly that person can act. Give an AI system the same permission, and it can move at lightning speed, touch thousands of records and multiply the risk before anyone realizes what happened.
As Ashley Fidler, chief product officer at MoxiWorks, warns, AI does not simply inherit risk. It scales it.
A road of risks
The next phase of AI will connect more deeply with company data, recommend what people should do next and, increasingly, take action on their behalf. That creates enormous opportunity, but only for organizations that know what they are trying to accomplish and build the right guardrails before turning AI loose.
That’s one of the key themes explored in Episode 3 of Inside AI, hosted by me, Kevin Hawkins, The REAL AI Guy. Inside AI is a video interview series that gives real estate professionals an insider’s look at the people shaping artificial intelligence across the industry.
Through candid conversations with technology leaders, innovators and executives, this one-on-one video series cuts through the hype to explore what is working today, what is changing tomorrow and what agents, brokers, MLSs and associations need to understand as AI reshapes real estate.
In this episode, now available on RE Technology’s YouTube channel, I sit down with Ashley Fidler, chief product officer at MoxiWorks.
Ashley has spent more than 20 years developing AI and machine-learning products across industries and companies, including Microsoft. Her experience spans cybersecurity, legal technology, property management and residential real estate.
Or should I say Dr. Fidler, as she holds a Ph.D. in linguistics from Georgetown University. She began working on natural language processing and automated meeting transcription long before ChatGPT became a household name.
Her experience offers a rare perspective on both the promise of today’s AI and the mistakes companies repeatedly make when adopting new technology.
The biggest mistake starts before the technology
When I asked Ashley about the biggest mistake organizations make when incorporating AI, her answer had nothing to do with selecting the wrong model, moving too slowly or failing to hire enough technical talent.
“The single biggest one across all organizations is not knowing what you want to do,” she said.
That may sound simple, but it is one of the most important lessons in our entire conversation.
Too many organizations begin with the declaration that they need an AI strategy or must start using AI. But “using AI” is not a business objective.
Ashley notes that before choosing a platform or launching a pilot program, an organization needs to identify the business problem it is trying to solve. What outcome does it want? What process needs to improve? How will success be measured?
Without those answers, a company can easily select the wrong technology, implement it in the wrong part of the business or fail to generate the expected return.
As Ashley explains, the technology itself may work perfectly. What failed was the connection between the technology and the organization’s actual business goal.
Sometimes the smartest choice is not using AI
One of Ashley’s most refreshing observations is that AI should not be used for every task simply because it is available.
“The only time you need AI is when you need flexible intelligence,” she said.
She explains that if a process always follows the same predictable rule, traditional workflow automation may be faster, less expensive and more dependable.
For example, moving a new website lead into a CRM does not require artificial intelligence. The rule is straightforward: When a lead arrives, send the information to the CRM.
But reviewing that lead, interpreting what the person may want and drafting a personalized response could require the flexibility that AI provides.
Ashley’s advice is simple: If you can write a consistent “if this happens, then do that” rule, you may not need AI at all.
AI needs connected data to deliver its full value
Real estate has spent years assembling individual technology products for specific tasks. A brokerage may have separate systems for its website, lead generation, CRM, email marketing, presentations and transaction management.
Each product may work well on its own. The problem begins when those systems do not communicate.
Ashley believes the next major stage of AI adoption will involve connecting AI to the information companies already have and allowing data to move more freely across the organization.
“There’s a massive advantage with AI to having all your data talk to each other,” she said.
Connected data gives AI a more complete view of the client, the agent and the business. It can help identify which consumers may be preparing to buy or sell, surface opportunities that deserve attention and recommend the next best action.
It also can help solve one of the most persistent challenges in real estate: making sure leads, tasks and relationships do not fall through the cracks.
Brokerages and MLSs need to evaluate their entire technology ecosystem, not just individual products. Even when tools come from different providers, the information needs to flow together underneath them.
Otherwise, AI may have access to only a small piece of the story.
Do you really want to become an IT department?
The accessibility of today’s AI tools has made it easier for agents and brokerages to build custom workflows, connect applications and even create their own software.
That can be exciting, but it creates a new set of responsibilities that many businesses underestimate.
Building the first version with AI is often the easy part. Someone still needs to maintain it, troubleshoot failures, update integrations, manage permissions and support the people who depend on it.
Ashley recalls speaking with a real estate professional who was excited about automating her brokerage. Ashley’s response was supportive but practical: Who is going to support everything after it is built?
Before long, a real estate professional can become a full-time technology manager.
It reminds me of the real estate app era, when many brokerages believed they needed to build their own mobile app. They later discovered that building an app was only the beginning. It also had to be updated, maintained and supported.
Ashley’s point is not that companies should never build their own technology. It is that they need to understand the long-term commitment.
AI can save time and keep agents organized
The benefit agents want from AI is not always more transactions.
Many agents are satisfied with the size of their business. What they want is more control over their time. They want to complete administrative work faster, stay organized and spend more time with clients, family or friends.
Ashley hears something similar from agents. AI can help them keep track of information and surface items that require attention.
“AI helps them not lose things,” she said.
That may not sound as dramatic as doubling production, but it could be one of AI’s most practical benefits. Agents spend much of their time in cars, at properties and in the community. They are not always sitting at a computer carefully updating every system.
AI can help capture more of that information and make sure important details are not forgotten.
Shadow AI is lurking
One of the biggest challenges for MLSs and brokerages today is that they can’t control what they can’t see. Shadow AI, unauthorized AI tools being used within a business, exponentially increases risk, especially with free AI tools.
Ashley has significant experience with shadow AI.
At a previous property management company, she conducted an anonymous survey after hearing assurances that employees were not placing company or client information into free AI accounts. The survey showed that roughly 65% of employees were already using free AI tools for work.
That experience convinced the company that it needed approved corporate accounts, stronger controls and greater visibility into how AI was being used.
Brokerages and MLSs should assume that some level of AI experimentation is already happening. A written policy alone is not enough.
Organizations need approved tools, secure business accounts, appropriate permissions, audit records, and clear restrictions governing which data and systems AI may access.
Ashley also offers this warning: AI does not merely inherit risk. It scales it.
Experiment, but limit the blast radius
Ashley is not discouraging experimentation. In fact, she encourages real estate professionals to explore AI and become more familiar with what it can do.
But experimentation needs boundaries. Her advice is to “limit your blast radius.”
Instead of immediately connecting an experimental AI agent to an entire technology system, test it in a controlled environment, often called a sandbox.
Provide access only to the limited data and tools needed for the experiment. Observe how the agent behaves. Confirm that it follows instructions and handles unexpected situations correctly.
This becomes especially important as real estate moves toward agentic AI.
Closing the AI action loop
Ashley sees enormous potential in agentic AI because it can connect insight with action.
She describes the ideal process as an action loop.
Data enters the system. AI analyzes that information and makes a useful suggestion. The agent or the technology takes an action. The system then captures the result and uses that feedback to improve what happens next. Real estate is only beginning to close that loop.
Today, many AI tools provide an insight or draft a recommendation. The next generation will help agents act on those insights, track what happened and use the result to create better recommendations in the future.
That could mean identifying a promising lead, preparing a personalized email, helping create a CMA or recommending the next step in a relationship. But human oversight remains essential.
As I taught this week for Monument MLS in Bullhead City and Kingman, Arizona: An agent, broker or MLS cannot blame the AI for a mistake. The real estate professional owns the license. AI does not and cannot have a real estate license.
Being the human in the middle remains key for Gen AI and even more true for agentic AI.
Watch Episode 3 of Inside AI featuring Ashley Fidler, chief product officer at MoxiWorks, on RE Technology’s YouTube channel. The full interview explores connected data, agentic AI, shadow AI, cybersecurity and why the smartest AI strategy sometimes begins by deciding not to use AI at all.
Learn more about AI in real estate and get your free REAL AI newsletter subscription at wavgroup.com/ai.
About Kevin Hawkins
Kevin Hawkins, known throughout the industry as The REAL AI Guy, is a WAV Group partner and the author of the Amazon bestselling book The REAL AI Guide for Real Estate Agents. He is also co-creator, with WAV Group’s Korey Hawkins, of the REAL AI newsletter, real estate’s most popular weekly newsletter dedicated exclusively to artificial intelligence for agents, brokers, MLSs and associations. Through his writing, training and interviews with industry leaders, Kevin helps real estate professionals cut through the AI hype and focus on what is practical, useful and safe. Hire Kevin to speak, train or host your webinar by contacting him here.
The post Inside AI – Episode 3 with Ashley Fidler: The biggest AI mistake real estate companies are making appeared first on WAV Group Consulting.
Leave a Reply