Editor’s note: This article was drafted using AI from a transcript of The Leads Are Sh*t. It was reviewed for accuracy and edited for publication by REM editorial staff.

Recording a video can take less time than deciding what to do with it afterward.

There is the edit. The captions. The title. The short clips. The email telling people why they should watch. For an agent who already finds it difficult to make time for content, that second job can keep the first one from ever reaching an audience.

In this episode of The Leads Are Sh*t, Andrew Fogliato and Taylor Hack work through a practical question: how much of that production process could an agent hand over to AI, and what would still need a person?

Fogliato proposes using the show itself as a test. Take the recording, trim it, create a transcript, draft the REM article and newsletter copy, and cut out short clips. The target is a complete package waiting for him when he returns to work.

The final words of his brief matter: “I want it all delivered to me to review.”

He is describing an experiment he wants to build, not presenting a finished, autonomous publishing system. That distinction makes the conversation more useful. It turns a broad promise about AI into a job with identifiable steps, an output and someone responsible for checking it.

Start with the task you keep putting off

Earlier in the conversation, Hack asks the live audience what is sitting on their to-do lists. The answers include website work, property management and arranging help around a closing.

The question points toward a better starting place for an AI experiment than a list of new apps. Which recurring task consumes time, creates friction or gets postponed even though it needs to happen?

Video production is one example. Preparing information for a client, reviewing a website or organizing an existing set of materials might be another. The useful question is what a completed task should look like in the agent’s business.

Hack is sceptical of technology pitches that promise to transform real estate without understanding the work. Fogliato makes a similar point: someone can go through a transaction, find part of it frustrating and still miss why the surrounding systems exist.

A smaller, well-understood job gives an agent a clearer test. Did the result save time? Was it accurate? Could the agent use it without doing the work again?

Define what happens after the file arrives

Hack and Fogliato then walk through the video workflow in order.

An idea becomes a script or talking points. The agent records the video. The file arrives in a designated location. An editing process follows written instructions, adds captions and prepares the material for its intended destination.

For the show, Fogliato’s proposed outputs extend beyond video: a transcript, an article, newsletter content and clips. Each serves a different purpose. An article needs to make sense to someone who has not watched the episode. A clip needs enough context to stand on its own. Captions need to reflect what was actually said.

That means the brief has to be more specific than “make content from this.” It needs to explain what to remove, what to retain, how the result should look and what a reviewer should receive.

The conversation also separates production from publication. Fogliato wants the materials prepared for human review first, with a later approval step before anything is posted. At REM, he describes a process that identifies AI-assisted recaps and includes human editorial review.

The instructions need your judgment

A repeatable workflow still needs editorial decisions.

A technically complete clip may leave out the sentence that explains the speaker’s point. A headline may be more dramatic than the discussion supports. A draft can sound polished while getting a name or a fact wrong.

The same problem appears when AI is asked to interpret performance data. Fogliato describes giving a system REM’s article and analytics data to identify the best publishing time. It found a window associated with stronger results.

The missing context was what REM published during that window. Those articles were breaking news, often supported by dedicated emails and greater sharing. The apparent timing advantage was tied to the kind of content and how it was distributed.

Following the recommendation without that context could have produced the wrong publishing decision.

The lesson for an agent is to explain the business behind the numbers. What was being promoted? Who received it? What was the next step? What else changed? A report cannot fill in operational context that was never supplied.

Measure the next useful action

When the discussion turns to Instagram, Fogliato frames his objective around clients rather than an influencer’s audience goals.

Depending on the strategy, useful signals might include profile visits, saves, shares or responses to a specific invitation. A video asking viewers to comment for a resource needs to be evaluated against that action. A campaign intended to start conversations needs a way to follow those conversations beyond the view count.

This gives the production workflow a more useful destination. The job is not finished simply because a file has been exported or a post has appeared. The agent needs to understand whether it is helping people take the intended next step.

That feedback can inform the next recording: the question to answer, the explanation to tighten or the invitation to make clearer.

Give production support a try before handing over relationships

The hosts are more cautious about using AI on the relationship side of the business.

Fogliato says that would be the last part he would want to hand over. He sees a role for support and for filling gaps, but favours openness when an automated assistant interacts with a consumer. Hack connects the issue to trust: an interaction that feels inauthentic can interfere with the connection an agent is trying to build.

For an agent deciding where to begin, that is a useful boundary. There is already plenty of work between having something worthwhile to say and getting it in front of the right people.

Choose one recording. Define the outputs. Supply examples. Review what comes back and refine the instructions. That creates a practical way to judge whether AI is reducing the production burden while leaving the agent responsible for the message.

Join the conversation live

The Leads Are Sh*t is a weekly strategy lab where Andrew Fogliato and Taylor Hack break down what is working, fix what is not and test ideas inside an actual real estate business.

Join the show live every Thursday at 2 p.m. ET to ask a question, take the hot seat and stay for the off-the-record After Show, which is not included in the replay.

Register once and save your spot for the next live session.

Related Posts