Recruiters turn to AI in their quest to lure top agents

AI

Agents are the lifeblood of real estate brokerages. Recruiting and retaining high-performing agents remains a critical challenge for brokerages, especially in the environment created by the National Association of Realtors’ recent $418 million settlement in a litany of agent commission lawsuits.

While real estate brokerages have no control over interest rates, transaction levels or home sales prices, they realize that proactive recruiting of agents is a key to success in a slow market.

Sales managers often struggle to source enough quality agents to feed the top of their funnel. They might have a hard time crafting a system that produces consistently strong hires, according to Mark Johnson, managing partner at Recruiting Insight. This can also be tied to the fact that managers typically have to juggle many other tasks at the same time, such as coaching or contract signing, which poses a challenge to their organization.

Johnson said that “word of mouth” is the best way to organically recruit new agents. If a brokerage enjoys a good reputation, its agents will speak highly of it and draw new agents to the brokerage. If agents don’t bring in new recruits directly, a third-party source such as an existing client or a business partner might inspire an agent to join a brokerage. Curated events can also be a great way to recruit agents as it allows a brokerage to showcase its culture, Johnson added.

Tech enters the picture

Alongside these possibilities, the integration of an artificial intelligence (AI) tool can also be a solution. When AI started to gain momentum in the real estate industry, most investment capital was poured into consumer-facing products such as Zillow’s Zestimate or predictive marketing solutions.

Launched in 2006, the Zillow Zestimate leverages public documents to provide estimates of value for every house in a given neighborhood. These types of consumer-facing products generated a lot of interest from investors. Simultaneously, predictive marketing also gained a lot of traction. Agents started using predictive modeling to determine which person in a neighborhood, or which person in a list of contacts, was most likely to list their property in the near future.

The focus on aiding brokers in team recruiting and management through AI remained limited until more recently. In 2016, Robert Keefe pioneered this niche with Relitix, paving the way for a new wave of innovative companies in the sector. AI can help streamline the recruitment process in real estate by leveraging data from diverse sources to identify candidates with the desired skill sets and cultural fit.

“The scarcest resource in real estate is brokerage manager time,” Keefe told HousingWire. “I felt that emerging data tools and technology paired with MLS and other agent-level data would allow us to help leverage that scarce time resource and make an hour of manager time 10 times as productive through data superpowers.”

Companies like Relitix, Lone Wolf Technologies, Courted, Brokerkit and MoxiWorks offer AI tools to streamline the recruitment process of agents. These companies leverage different techniques such as predictive analytics and machine learning — including the subsets of active learning and deep learning — to forecast the future performance of an agent.

Deep learning is a more advanced form of machine learning. It uses multiple, sequential layers of machine learning to produce highly refined and nuanced answers, and it forms the basis of the AI large language models such as ChatGPT that have become prevalent in the past year.

When labeled training data is not available (i.e., there is data but no “answers“), data engineers can use active learning to have the AI create a smaller dataset with more important data points for review and labeling by human reviewers. This smaller dataset can stand in for the large, fully labeled datasets without the need to label as many examples.

The goal of these analytics tools is to improve expert decision-making by converting raw data into insights, inferences or predictive models that can aid operational processes. Analytics are best viewed as a repetitive process in which predictive models are continually refined and improved.

“Data certainly abounds in real estate. In fact, it’s the existence of data in this industry that facilitates advanced analytics approaches,” said Sean Soderstrom, co-founder and CEO of Courted, a New York-City based firm founded in 2021.

“The problem, however, is that the vast majority of brokerages use historical data to make recruiting decisions,” he added. “Courted has shown that there is significant volatility in agent production year over year.

“Without AI, it is nearly impossible to discern between an agent whose best selling years are behind or in front of them at scale. You can possibly do this if you have a connection to that agent and you know their business deeply, but it’s impossible to comb through all this information for all agents who might be a good fit for your brokerage, given one and a half million agents in the country.”

Alleviating pain points

Companies have identified three main problems with how brokers have been recruited in the past.

First, there tends to be significant volatility in an individual’s production from year to year. Second, each agent has a varying likelihood to change brokerages at a specific point in time. Lastly, recruiters can struggle to create the right message that will attract a specific agent.

To tackle these issues, tech-based recruiting companies leverage data from multiple listing services (MLSs), which undergo meticulous data engineering to generate comprehensive agent profiles. These profiles are then subjected to comparative analysis against a relevant cohort of agents. Key performance metrics such as sales volumes, the average time that listings spend on the market, and sale-to-list-price ratios are evaluated.

Additionally, machine learning and predictive analysis models forecast an agent’s production volume for the ensuing year and assess the likelihood of turnover for a specific brokerage.

For example, the so-called “switch risk rating” from Relitix is used to assess the relative likelihood that an agent will switch brokerages in the next three months. Recruiters use it to help prioritize recruiting lists and managers use it in conjunction with agent retention efforts.

Meanwhile, the company’s “rookie potential rating” measures the success potential of agents with less than 36 months of MLS experience. It allows recruiters to uncover high-potential agents with limited experience amid the noise of large numbers of new agents.

Lastly, the “listing effectiveness grades” evaluate agents on how effective they are in closing their listings. Every agent in the MLS is assigned a grade (A to F) each month, with 20% of the total agent pool receiving one of the five grades. Agents with “A” grades represent the top 20% of the pool and can be counted on to close the most difficult listings with a high level of reliability.

Conversely, agents with “F” grades have failed to close relatively easy listings and have an outsized number of canceled, expired or withdrawn listings relative to their peers. Relitix’s machine learning algorithm is retrained each month to reflect current market conditions, and to ensure that the grading is adjusted for listing difficulty and local supply-and-demand conditions.

“This is going to be an especially important metric in the next few years,” said Keefe, the founder of Relitix. “We have been producing this grade on the agent pool since 2019.”

Next-level insights

Joshua Paul, vice president of operations at ONE Sotheby’s International Realty, uses AI recruiting tools from Courted. They allow him to prepare effectively before meeting with a job candidate, and they help make the conversations more personalized and more intentional, he said.

At a glance, Paul can access an agent’s yearly transaction details and their results across various key performance metrics. He can also consult predictive analytics to forecast the potential performance of an agent.

“I can understand the agent’s entire business in less than 20 minutes, so that when I’m having an in-person conversation with the agent, it feels like I’ve done a substantial amount of research,” Paul said.

In these cases, AI helps pinpoint some of the high and low points of an agent’s production. Although there is no “crystal ball” when it comes to real estate agents, AI can help benchmark a  selection, according to Nick Weitekamp, executive vice president at West USA Realty, a brokerage that uses Relitix tools.

“Relitix gives us a little more insight into what the potential for an agent could be, based on the metrics that have already been pumped into Relitix,” Weitekamp said. “That’s a nice kind of conversation piece.

“When we are talking to different agents, it gives us a little more of an insight into that crystal ball. It’s not perfect. But it just helps guide the conversation about what these agents could become if they got a little more support, a little more training, a little more focus in the business.”

Ultimately, when it comes to recruiting, the value proposition of a company is key. Many brokerages offer financial incentives, such as upfront cash and stock bonuses or better splits, to lure more agents into joining. But culture, technology, support and other factors can be strong arguments to onboard more agents.

Most large, publicly traded real estate brokerages invest heavily in their recruiting functions. Some even choose to invest in their technology to recruit agents.

According to Soderstrom, AI-assisted recruiting can contribute to leveling the playing field for players across the real estate industry. It allows bigger companies to get a better return on investment for their recruiting efforts while providing smaller companies with powerful tools that enable them to compete with larger, more well-capitalized competitors.

ENB
Sandstone Group