For many modern real estate professionals, recording a video is often the easiest part of the content creation cycle. The real bottleneck frequently lies in deciding what to do with the footage afterward.
The post-production marathon is daunting: editing the raw video, generating accurate captions, crafting compelling titles, slicing the long-form recording into punchy short clips, and writing accompanying email newsletters to explain to clients and subscribers why they should take the time to watch. For a busy real estate agent who already struggles to carve out time for marketing, this heavy administrative and creative second job can easily derail consistency, keeping their insights and expertise from ever reaching an audience.
In a recent episode of the industry podcast The Leads Are Sht*, hosts Andrew Fogliato and Taylor Hack tackled this exact operational friction point. Together, they explored a practical, highly relevant question for today’s market: How much of the cumbersome content production process can an agent safely hand over to artificial intelligence, and what parts still fundamentally require human judgment?
Fogliato proposed using the very episode they were recording as a live test case. His vision was straightforward: take the raw recording, trim it down, generate a clean transcript, draft a companion article and newsletter copy for REM, and extract engaging short clips. The ultimate target was a complete, ready-to-review package waiting for him the moment he returned to his desk.
Crucially, the final words of his operational brief set a strict boundary: “I want it all delivered to me to review.”
Rather than presenting a finished, fully autonomous publishing system that operates entirely without human oversight, Fogliato was describing a controlled experiment. That distinction is what makes the conversation particularly valuable for real estate professionals navigating the crowded market of AI tools. It reframes a vague, often overwhelming promise about artificial intelligence into a concrete job with identifiable steps, a tangible output, and a designated person responsible for checking the quality before anything goes live.
Start with the task you keep putting off
Earlier in the episode, Hack posed a question to the live audience, asking them to identify the lingering items sitting stubbornly on their to-do lists. The responses spanned a wide spectrum of operational tasks, including website maintenance, property management, and coordinating the intricate details surrounding a real estate closing.
This line of questioning points toward a far more effective starting point for any artificial intelligence experiment than simply browsing a list of newly released applications. Instead, agents should look inward at their own workflows to determine which recurring task consistently consumes valuable time, creates unnecessary friction, or gets repeatedly postponed despite its importance.
While video production serves as a primary example, preparing comprehensive market information for a client, reviewing website performance metrics, or organizing an existing library of marketing materials could just as easily fit the bill. The most useful question an agent can ask is what a successfully completed task should actually look like within the unique ecosystem of their business.
During the discussion, Hack expressed a healthy skepticism toward technology pitches that promise to radically transform the real estate industry without demonstrating a genuine understanding of the day-to-day work agents perform. Fogliato echoed a similar sentiment, noting that an individual can go through a complex real estate transaction, find a specific part of it deeply frustrating, and still completely miss why the surrounding administrative systems exist in the first place.
By starting with a smaller, well-understood job, an agent gives themselves a much clearer framework for testing emerging technology. They can readily answer critical operational questions: Did the final result actually save time? Was the output accurate? Could the agent utilize the finished product without having to completely redo the work?
Define what happens after the file arrives
As the conversation progressed, Hack and Fogliato walked through a typical video workflow in chronological order to identify where automation can genuinely assist.
The process typically begins with an idea, which evolves into a script or a set of talking points. From there, the agent records the video. Once recorded, the raw file arrives in a designated digital storage location, triggering an editing process that follows established written instructions, applies captions, and prepares the material for its intended distribution channels.
For the podcast experiment, Fogliato’s proposed outputs extended well beyond a single video file. He outlined a multi-format package including a full transcript, a written article, targeted newsletter content, and individual clips. Each of these outputs serves a distinctly different purpose and audience. An article, for instance, must make logical sense to a reader who has never watched the original episode. A short clip needs to contain enough immediate context to stand entirely on its own on social media. Captions must accurately reflect what was spoken without awkward errors.
This operational reality means that an agent’s instructions—or prompt—must be far more specific than a vague directive to “make content from this.” The brief needs to explicitly explain what elements should be removed, what information must be retained, how the final aesthetic should look, and what specific format a reviewer should receive upon completion.
Furthermore, the discussion drew a firm line between the production phase and the actual publication phase. Fogliato emphasized that he wants all materials prepared exclusively for human review first, with a deliberate, secondary approval step required before anything is ever posted publicly. At REM, he described a meticulous process that clearly identifies AI-assisted recaps and mandates rigorous human editorial review to maintain journalistic standards and brand integrity.
The instructions need your judgment
Even within a highly repeatable, automated workflow, human editorial decisions remain non-negotiable.
For instance, a technically complete video clip might inadvertently leave out the crucial middle sentence that fully explains the speaker’s primary point. An automated headline might lean into dramatic phrasing that the underlying discussion simply does not support. Similarly, a generated draft might sound polished and professional while simultaneously getting a proper name or a foundational fact entirely wrong.
The exact same pitfall emerges when artificial intelligence is asked to interpret complex performance data. Fogliato shared an example of giving a system REM’s published articles and analytics data in an attempt to identify the optimal time of day to publish new content. The system quickly pinpointed a specific time window that correlated with stronger overall readership results.
However, the missing piece of the puzzle was the critical context of what REM had actually published during that specific window. Those particular articles happened to be breaking news pieces, frequently supported by dedicated promotional emails and significantly higher levels of sharing across platforms. The apparent timing advantage was not a magic hour for publishing; rather, it was directly tied to the specific type of content being produced and how aggressively it was distributed.
Following the algorithm’s recommendation without understanding that operational context could easily have led to a flawed publishing strategy. The vital lesson for real estate agents is to clearly explain the business reality behind the numbers. A technology report cannot fill in operational context that was never supplied by the user.
Measure the next useful action
When the conversation shifted toward platforms like Instagram, Fogliato deliberately framed his strategic objectives around the actual needs of clients rather than chasing the broad vanity metrics typical of a traditional influencer’s audience goals.
Depending on an agent’s specific marketing strategy, useful signals might include direct profile visits, saves, shares, or direct responses to a targeted invitation. A video that asks viewers to leave a comment to receive a valuable local market resource needs to be evaluated directly against whether viewers actually took that action. Similarly, a campaign intended to spark genuine conversations with prospective buyers or sellers requires a reliable system to follow those conversations far beyond a superficial view count.
This approach gives the content production workflow a much more meaningful and accountable destination. The job is not successfully finished simply because a video file has been exported or a social media post has officially appeared online. Instead, the agent must actively evaluate whether the content is helping people take the intended next step in their real estate journey.
That ongoing feedback loop can then inform the next recording session: refining the core question to answer, tightening the real estate explanation, or making the call-to-action significantly clearer for the consumer.
Give production support a try before handing over relationships
Throughout the discussion, both hosts maintained a notable level of caution regarding the implementation of artificial intelligence on the deeply personal, relationship-driven side of the real estate business.
Fogliato noted that relationship-building would be the absolute last area he would ever want to hand over to automation. While he acknowledges a clear and valuable role for administrative support and filling operational gaps, he strongly favors complete transparency whenever an automated assistant interacts directly with a consumer. Hack connected this critical issue directly to trust, emphasizing that an interaction that feels inauthentic or robotic can instantly undermine the genuine human connection an agent is working so hard to build.
For any real estate professional trying to decide where to begin their own technology experiments, this distinction provides a useful and safe boundary. There is already an immense amount of necessary work that sits squarely between having something worthwhile to say about the housing market and successfully getting that message in front of the right audience.
By choosing a single recording session, carefully defining the desired outputs, supplying clear examples, and reviewing what comes back to refine the instructions, agents can establish a practical method for evaluating new tools. This measured approach ensures that artificial intelligence reduces the exhausting burden of content production while leaving the agent firmly in control of their own professional message.