AI Co-Pilot (Lucy)

Rechat AI Innovators: Engineer Abbas Mashayekh

Today we have the pleasure of chatting with Abbas Mashayekh, who has worked as an engineer at Rechat for more than six years. He’s Rechat’s Backend/API team leader.

In layman’s terms, he “oversees the development of the core of Rechat’s codes that powers the web and mobile applications,” he explained.

He came to Rechat for a bigger challenge, and he’s certainly tackled that assignment over the years as Rechat’s product and engineering team has grown.

“The fact that we have managed to build this product, this big, big Super App, with such a small team in engineering, that’s really amazing, and I’m really proud of what we have achieved,” he said

Read on for our discussion about Rechat’s latest innovation, Lucy, an AI copilot for agents, as well as what he feels makes Rechat an exciting and fulfilling place to work. 

What has been your role as it relates to the development of Lucy?
Abbas: My role has been more like a consultant, if you will. I was with Emil (Sedgh, CTO) and sometimes with Adam (Schwartz, Senior Backend Engineer) discussing the technical aspects and how to implement various features. But I wasn’t involved in directly implementing those features. I was advising and giving my opinion, and we explored the AI landscape and what tools were available to us. So that’s very high-level.

It seems like it must be an exciting time to be an engineer. How did you feel about being involved in these types of discussions?
Abbas: Oh absolutely! But you know, I was, at the beginning, kind of skeptical about this. Like people were going crazy about AI and about ChatGPT, and the things that it could do. Like you went on social media, and everyone was talking about AI. I was like: What is all this talk about? Why are people getting so crazy? And you know, my background is in AI! Ten years ago, I completed my Masters in Artificial Intelligence and specifically in natural language processing, which is the basis of the language models and the GPT model. So at the beginning, I was kind of skeptical, but as we tackled and took on more challenges, it started to get more and more exciting. I’m really amazed at what Lucy can do now, what it can do today and what’s possible in the near future, all the things that the team is working on. It’s really exciting for me.

So what makes Lucy special is that it is able to understand a real estate agent’s needs, and also access the agent’s listings or deals that they have in Rechat. Is that correct?
Abbas: As you just explained, Lucy has access to everything, like every piece of data, every piece of information that a user might have in Rechat. And this is all private and will remain private: This is all inside the Rechat system. It doesn’t get back to the GPT platform, right? Open AI cannot access everything in Rechat. Lucy, not Open AI, has access to everything. And with that, comes the ability to guess what they need. We can also instruct Lucy, we can embed features inside Lucy so it can improve its suggestions and make tailored suggestions and recommendations based on the specific needs of the agent. 

So the engineering team is essentially creating Lucy’s behavior, the responses, and the kinds of questions and answers that a real estate agent would need to know?
Abbas: Yes, we are giving Lucy instructions so that it can tailor its responses based on the needs of a real estate agent. We are also building tools that it has access to, unlike ChatGPT that does not have access to tools – if you ignore plug-ins, which are not the same exactly as what we have in Rechat. We are giving it tools so that it can interact with Rechat’s data and the user’s data. And then it can give you reports, it can find your contacts and send them an email directly. It can create a deal for you, it can create a marketing piece for you, it can create a CMA for you in under a minute. So these are the tools that we’ve built for Lucy. It’s exclusively for Lucy, and it has access to everything that the user has in Rechat, in order to be able to do all of those things. So just to recap, we are giving Lucy instructions and also tools for all of those features.

In the future, will Lucy be able to predict the agent’s needs?
: So that’s a good question. I’m not sure if we will go that way, but it is possible because the things that an agent does, it’s not different for each agent. They have a common set of tasks and things that they do for their clients. If you are trying to sell a property for your client, you will most likely send mass emails to all potential buyers and buyer agents that might be interested in your property. You will also probably hold open houses on multiple occasions for that property. Those things are predictable. So we can say, hey, according to the data we have, open houses attract more clients, more potential buyers, when you hold the open house at these hours, for example, or on these weekdays. And then it can recommend to the agent that, hey, you’re going to need three open houses for this property, and we suggest that you do them in these time slots, on these days of the week.

So this is possible. It’s not something that Lucy does today, but it’s possible, and I’m sure the team is looking at it and considering such features. But it’s also a way of predicting the user’s needs, like creating an open house, creating marketing pieces, creating something to post on social media or creating ads for the property. These are all the things that agents do and they are predictable and Lucy can suggest and recommend those actions based on the data that we have in Rechat.

Is it impressive for you to look back and think of how Rechat has developed over the years? 
: I will answer this in two different ways. First, we’ve always been strong on product building and building features and making this into the Super App that it is today. That is not to say that I’m not amazed at what we have achieved, but in a way this is all something that we strive for and we try so hard to do. On the other hand, with AI, this came at a big surprise for me, that Rechat will shift the direction into building this AI thing, into building Lucy. I remember back then we used to joke that agents asked for everything and it’s not going to be surprising if they asked for an AI, right? And we used to joke about that, and I never thought that we would be building an AI in such a short time at Rechat. So yeah, that’s really amazing and surprising to me.

After working at Rechat for more than six years, what do you enjoy the most? What keeps you at Rechat?
: For me, there are all sorts of things. The first and most important thing for me is how it is customer-driven. We are putting the needs of our clients at the top priority on our roadmaps, on everything that we do. And our customer success and support team is doing an amazing job of that as well. So everything we do is around what the clients need, the agents, admins, the people who use Rechat in their day-to-day life need. So that’s really something special for me. And the other factor is that Rechat is also growing. We’ve onboarded tons and tons of clients just in the past month, for example. That’s also amazing for me. And also on the technical aspect, I’m an engineer and a lot of technical stuff excites me. So the fact that we have managed to build this product, this big, big Super App, with such a small team in engineering, that’s really amazing, and I’m really proud of what we have achieved. I’ve been working at Rechat long enough to be able to say that I was part of this effort, and I’m proud of what we have done. At this point I can say that Rechat is something that I built, and I’m really proud of it, and I really like what I’ve done here.