Real estate agents and loan officers need each other, and the platform was built around that relationship. It started as a marketing tool for creating listing flyers and neighborhood infographics, then gradually became a broader networking platform through co-branding, automation, and direct MLS integration.

That evolution changed what we were asking users to trust us with. Someone who originally signed up to make a flyer was now being asked to let the platform reach out to another professional on their behalf.

The experience also had to operate within U.S. consumer-protection requirements, including RESPA. Our business model and direct MLS partnerships were structured around those requirements, but the product still needed to make the relationship feel legitimate and understandable. The LO's job was essentially to give us permission to make the introduction.

That still required a lot of trust. Sending an email to a stranger in someone's voice, or formalizing a relationship they'd built themselves, goes beyond adopting another feature. You're asking them to hand over part of a professional interaction and trust the product to handle it well.

We approached that in two ways: warm relationships and cold ones.

Warm relationships came first because they were the foundation of the business. If an LO already knew an agent, we needed to make it easy to bring that relationship onto the platform without getting in the way. Existing relationships were also where the value of the network was most obvious.

Cold outreach was more complicated. There was no shared history and often no obvious reason for an agent to respond. A generic message could make the LO look unprofessional, while an unexpected email could make the agent question why they were being contacted in the first place.

That meant the product had to do more than facilitate an introduction. It had to give the LO enough context to make the outreach feel appropriate, while giving the agent enough information to understand what was happening and why.

The warm pairing experience started as a manual process. LOs could search for agents they had worked with and send an invite, but that meant they had to remember who they had worked with in the first place. The system could facilitate the connection, but it wasn’t helping them identify the right people.

Once we had transaction data, we were able to change that. We surfaced suggested agents directly below the search field, using past transactions to identify people the LO had an existing relationship with. It was a relatively small change to the interface, but it shifted some of the work from the user to the product. Instead of relying entirely on memory, LOs had relevant relationships surfaced for them.

We carried that same idea into the outreach itself. The email language was intentionally familiar and specific to the existing relationship rather than sounding like a generic invitation to connect. LOs could also preview exactly what the agent would receive before sending it. That was particularly important because the platform wasn't communicating as the LO; it was communicating about them: "[Lenny Lender] has invited you to pair." The preview gave the LO a chance to see how their name and relationship would be represented before putting it in front of someone they knew.

The agent experience had a different problem. From a business perspective, the platform was built around networking, but an agent could only have one active LO pairing. That created a tension in the experience: we wanted agents to discover and connect with more LOs without implying they could maintain multiple active pairings.

We designed the experience around the value of pairing first, then surfaced suggested LOs based on transaction history. We also made the limitation explicit: agents could have one active pairing, but they could still invite other LOs. That gave the constraint some context instead of making the interaction feel like the product was simply preventing them from doing something.

Cold outreach was difficult for a straightforward reason: there usually wasn't enough context. An LO reaching out to an agent they'd never met had little reason to expect a response, and the agent had little reason to respond.

We tackled that in two places: the information available to the LO and the mechanics of the outreach itself.

Agent Insights gave LOs a reason to reach out. It brought together a nationwide database of agents with transaction history, production volume, price range, loan types, and a map of past and current listings. Instead of approaching an agent with no context, an LO could understand their business before deciding whether to make contact. LOs could also follow agents who hadn't accepted an invitation yet, keeping up with their activity without having to initiate a relationship. When they were ready to reach out, they had something specific to talk about.

We gave LOs similar control over how that outreach happened. They could have the platform send the invitation on their behalf, with a full preview of the email and the option to send a copy to themselves first. Or they could copy a unique invite link and send it from their own inbox. Both paths led to the same outcome: once the agent clicked the link, the two accounts were automatically paired. That meant we could remove the friction from connecting without taking control of the relationship away from the LO.

We also introduced a smart email option to make the outreach easier to write. It started as a simple template and became more personalized as the underlying technology improved. The LO remained responsible for sending the message; AI helped with the drafting.

We were equally deliberate about the business model. LOs paid for the platform, while agents could use it for free. We made that clear during the invitation flow rather than leaving the agent to figure it out after receiving an unexpected message. In a relationship-driven product, that kind of transparency mattered.

Over time, we moved suggested agent pairings earlier in the experience, including them in LO onboarding and surfacing them again on the home dashboard. The goal was to make finding relevant agents an ongoing part of using the product rather than something the LO had to remember to do on their own.

The platform was eventually acquired, with growth driven primarily by word of mouth.

Two things helped that growth compound. The first was reputation. An NPS of 81 suggested that users were comfortable trusting us with their professional relationships and giving the platform a role in their outreach. That confidence translated into referrals and brought more users into the network.

The second was distribution. We chose not to white-label the product, so our logo appeared on every piece of co-branded marketing material the platform generated. Every flyer sent to an agent was another opportunity for someone new to encounter the platform. As the network grew to 550,000 users, those existing relationships helped bring more people into it.

The same trust we had to earn to make the product work also helped the network function as a growth channel. That traction ultimately contributed to the acquisition.

The invite flow was our most fragile moment. Data gaps occasionally forced us to ask users for information mid-flow, which was friction at exactly the point we could least afford it. When research later surfaced that users weren't pairing, the fix that shipped was a paragraph of explanatory text. Watching it happen clarified something I already believed: patches at the UI layer don't fix structural problems, they just make them harder to see. The trust you build in the first screen can be undone by one moment of unnecessary friction two steps later.

In trying to eliminate uncertainty, we may have overcorrected in places. The option to send the invite email to yourself first, so you could click through the agent experience before anyone else did, came from the right instinct. But the mechanic was convoluted. A simpler preview would have done the same job with less cognitive overhead.