Game Changers: Neil Dholakia


Neil Dholakia, Chief Product Officer at Keller Williams International, talks with editor-in-chief, Tracey Velt about his brokerage innovation, emerging tech and some of the trends he’s seeing in the industry.

Listen or read the full podcast interview below.

Thanks for joining the Meet The Influencers podcast. So we’ll just start a little bit at the beginning. Tell me a little bit about your background. How did you come to Keller?


Well I’ve been in Austin, Texas, for about 25 years. I started off as a developer myself in a little startup and I moved through the startup community here in Austin, group companies and was acquired a few times. And through my career I moved from development to engineering management to product to executive management, and learned a lot of things. I thought I was a great product guy just because I’d seen a lot and grown companies.

And then my last stop before Keller Williams, I was at a private equity company also here in Austin, and we specialized in acquiring enterprise software companies and I learned quite a bit about the end of a product lifecycle. So I filled a gap in my product playbook if you will on how to incept the product, innovate on a product, get it to market, grow the market and now how to put a product in bed properly.

I wanted to get out of PE because we spent a lot of time putting products to bed quite frankly, and I didn’t want to do that anymore, and I wanted to apply everything I’d learned throughout my career in that lifecycle and find some place that would allow me to do that without having to worry about some of the other mundane things like going out for more VC funding like you have to do in a startup or being in a non-interesting space.

So the opportunity at Keller Williams came about. We’re very well-funded, we’re the largest brand in the space and we’re disrupting an industry, so that keeps it fairly interesting. As well as gives me an opportunity to exercise the playbook I’ve developed over my last 20 odd years here in Austin.


You’ve got an interesting background and Keller Williams has a lot of opportunity for you with Kelle being at the core. Tell me about Keller Williams’ investment into technology and what you’re most excited about


Well, the investment really started in earnest about three years ago. Now, I’ve only been at Keller Williams for two years now and so it started a little bit before me, but I like to think it got going once I joined. However, prior to three years ago the technology investment strategy for Keller Williams was more of an IT focus. We had an IT department that would go out, vet third-party technologies, create or find tools, stitch them together and then give them out to our agent force. That was the prevailing strategy not just at Keller Williams but all of our competition as well.

However, Gary Keller our founder three years ago said, “You know what, we need to make a shift in our strategy we need to move away from how we’ve done things in the past and make a complete transformation of the company.” And that’s when he declared that we’re now going to become a technology company, a real estate technology company. And the first step in doing that was tying together all of our data.

So we made a big investment in integrating all of our systems together and making available all the data locked in those systems. And we did that in what we call Keller Cloud, we started building up all these capabilities inside of our own cloud infrastructure. Then in year two, we started building applications on top of all that integrated data and the applications were built using modern technologies, which was really refreshing for me in that we got to start off building and I’m going to use a little jargon here, micro services with react front ends for both web and mobile, without having to carry forward a legacy and that’s awesome and very refreshing.

That led to now in year three the rollout of those applications to our agent base. Basically, the users that we built all this technology for and that is what we’re really excited about because that rollout marks the broad general availability for the majority of our agents to finally get a view and hands on of the first product in our cloud, or the first application in our cloud, which is Command and it is a real estate focused CRM capability.

I like to think it’s more than CRM because this is the way an agent will engage with their clients, their customers, it allows them to not only keep track of who’s in their database, who’s in their sphere of clients but it allows them to manage the engagement from an agent perspective, to be most efficient in that engagement. But from a client perspective, allowing the client to dictate how they want to be engaged with so that the agent is interacting with their clients across the most appropriate channel, their preferred channel, whether that in-person meetings or chat or Skypes or emails. Whatever the channel that preference is what the system allows for an agent to be able to interact with their media database.


So you’re currently rolling it out to the agents. Talk to me about some of the lessons learned through the process of developing KW’s tech platform.


Yeah, that’s pretty interesting and that when I first joined I was thinking I was joining a traditional software organization. And Gary had just prior to that made the declaration that Keller Williams was going through a digital transformation and transforming the company. And so when I got here, I was bringing that playbook of how to build software products, and it was more of a traditional playbook when it comes to let’s record a bunch of requirements and interviewing some users, then build some product related to those requirements, and then rollout of V1, learn from what we did right, learn from what we did wrong and then let’s roll into a V2. And that’s your traditional cadence and the challenge in an enterprise software as well.

How do you squeeze that cadence so that you’ve got many turns to get through your versions in a given time period? And so you know major enterprise software companies maybe do two turns, a major release every year and a minor release every half year, more agile startups may do a quarterly turn on their on their software releases.

What Gary asked us to do was, “Let’s turn this around rather than releasing a piece of software and then figuring out what we did right and wrong. Let’s go straight to our agents. Let’s design a system that involves our users, our primary users, that involves them from the beginning and build the order exactly what they want, and then release it.”

So we created this concept called the labs process or KW Labs, where we start with a key set of users for a given piece of functionality and we’ll bring them into a specially designed room we created here with lots of screens and lots of collaboration space in our building, and we positive question them, “How would you like to see all of your contacts?” It’s a simple question, but it’s a very loaded question. And we hand out pieces of paper that have web templates or mobile templates, and we ask them, “Hey, just draw, how you’d like to interact with your contacts or how you’d like to see your contacts from your database in your CRM?”

And because these are key multi-mega business owners inside of the Keller Williams ecosystem these guys know what’s right and what’s wrong, they know what’s worked for them in their own businesses, they know what they’ve trained their teams on. And so their insights and in their drawings are hugely valuable to us. And so then, we do this a few times and we collect all those user designs and then a product manager on my team will take those designs, and call out the ideas that may not be on topic and maybe push that to another stream of labs in case somebody went off topic.

But they take that, correlate the information and then take that to our design team, and ask our design team, “All right, take these drawings and this additional information the product manager provides and ask our design team to come up with a pixel-perfect mock of what we’ve learned,” and then we take that that mock set of screens back to our labs environment, back to the same users plus a few more and ask them, “Hey, did we hit or not?” And they’re just looking at now the design screens and they say, yes or no, or, “Hey, yeah you hit right but that doesn’t look quite right. Let’s move this button here. Let’s expand these fields over there. Let’s add in this new feature that allows me to sort in this way.”

So we go through that iteration and then once we get to a steady state, we’ll then write the code that implements that screen. And once that code is finished we create a sandbox inside of our cloud and invite those lab participants now to hands-on, touch, feel, play with the functionality that they ostensibly designed. And that gets a lot of engagement because now those users feel ownership in the design and the product that they just finished seeing from start to hands on code ability, hands on usability.

And now what we’re doing is because they’re also influencers in our ecosystem, other agents, other users want to become part of our labs process. And so that we have a waiting list of folks that also want to get their fingerprints on the functionality of the product. And so we create this waiting list, we keep inviting more and more folks to our lab, and so those labs that sandbox keeps getting bigger and bigger and bigger with more and more folks. And we’ve designed our processes and our systems to record their feedback and to incorporate their feedback into subsequent designs.

As we push forward and expand that sandbox we’re adding more and more people to the product functionality, and we’re also combining sandboxes. So if you imagine you know we started this one stream around contacts we have other streams around activity plans, basically what can you do once you have contacts in your database with an automated fashion. So we have another stream and other set of labs going on along those lines and we recently combine those two sandboxes together.

Now the participants in those labs get combined together and now we’ve got a bigger population, now testing the integrated capability and giving us feedback on that. And so what that has in a traditional software sense done is it’s taken that cadence of turns for a major developer from doing six month turns on software, we’re down to bi-weekly turns or fortnightly turns of software. Every two weeks we’re updating software in one of our lab streams or we’re combining streams together to get more and more functionality.

We’ve been doing that for the last two years and now this year we’re rolling out to the masses, what’s been the result of that lab stream creating command our CRM system. Meanwhile, in parallel we’ve launched up another 20 odd lab streams around our consumer strategy which we’ll be launching in Q2 of this year. And so that lesson learned there.

Now, back to your original question was, well how do we engage our users and cut out a lot of the misses from traditional software development. Meaning, the product managers we have they’re not real estate agents, so they’re guessing and they’re educated guesses around what the agent needs to run their business. But if we can get directly from the agent what they want and involve them in the process, then we know we’re going to hit with our software right out the chute and we’re able to very quickly create smaller and smaller turnaround times for our new versions to where if an agent says, “Hey, wouldn’t it be nice?” We can say, “Here it is.”


Kudos to you for developing a product that everyone wants to use. There are obviously a lot of brokerage firms and networks jumping into the tech world. So tell me how do you envision the future of real estate given this trend?


Well, from what I’ve seen, no matter where the entities coming from whether you’re a brokerage like us or maybe a mortgage lender or some other participant in real estate. Everyone in the ecosystem is trending towards the same place and that is creating an end to end transaction platform and we’re doing it, our competitors are doing it, folks that aren’t our traditional competitors and have been partners they’re also moving in that direction.

Everybody wants to enable or even own the customer journey, whether it’s the journey from selling a home or buying a home or even home ownership, the more of that journey that is presented to that end user by any one entity you feel like, “Okay, well I’ve got them locked into my platform and therefore I can monetize that lock.”

Everyone is trying to create that same value in order to get the consumer to stay with them. I call that creating the first screen app, an application that won’t drift from your first screen when you install it back to the fourth screen or fifth screen where you’ve got to scroll to find it. Everyone’s trying to do that and the way you do that is you establish a platform that allows you to integrate the various systems required for a real estate transaction.

So let’s take a buy-side transaction you would need to have the agent tools there for managing contracts for interacting and selecting upstream of that, interacting with an agent to get help in selecting a home. Narrowing down your search and so forth to make an informed decision. And then as you move along you’re going to apply for a mortgage.

Well in traditional systems right now, you’re popping out a whole other application to get through with the mortgage process. Well, we think that should be in the same platform, let’s integrate lenders into our system and allow them to present their process within one experience, within one application. Now, once you get that done you’re going to want inspection.

Again, same pattern, we want home inspectors to be able to plug into our platform and offer their wares to the agent to the consumer at the right time so that it’s easy for a consumer to say, “Yeah, I’d like to use this person as my inspector. Here, I’m going to click this button or check this box, and then that initiates the right thing behind the scenes to make it happen.

And then you know you can keep carrying that forward too. All right. Now, that I’ve had that home inspected I want to get some work done on it rather than popping out to a whole separate app and a whole separate decision making process. One and I stay in context and be able to deliver, “Hey, these are the things that are the must haves out of the home inspection report that need to be repaired. Let’s take those and pass those along to three contractors that have already been vetted and have them provide bids and then select one to get the work done.”

We can do that all through one platform, we have created value for the users so that they don’t have to go anywhere else, and they have a trusted adviser and the agent that they can rely on to say, “Hey, if I ever have a question about any step in the transaction I have one place to go whether it’s picking up the phone and calling them or initiating a chat from the application or any other avenue that they want to be able to interact with that agent to get their questions resolved.” That could be a million problems.


Let’s talk a little bit about some trends you’re seeing related to real estate and technology other than the end-to-end platform. What are some of the trends you’re seeing now?


Well, with enterprise software in general and it’s trending into real estate software when it comes to agent tools, everyone’s also trying to get to the conversational UI. What I mean by that is being able to have a natural language conversation with your back-office system or your front-office system. And for us, that is enabling an agent to talk to their CRM and get the work done so that they don’t have to touch a keyboard.

For us, that is launching our bot named Kelle and enabling Kelle with real estate skills. And so now, an agent can as of a middle of last year talk to Kelle and say, “Hey, get me a market snapshot for the neighborhood I’m standing in,” and then Kelle will generate that for the agent and then the agent can say, “Well, let’s share this with Darryl.” And now I can send Darryl a market snapshot of the neighborhood he’s interested in. I can do all that with voice.

We want to get to a point where an agent gets into their car, their phone Bluetooth connects to the car, and the agent on the way to their office or to a show and can say, ask questions of, “Who do I need to contact today?” Kelle examining their system will come back and say, “All right, well this is the list of folks that you should contact today.” And the agent should say, “Great, start dialing.”

And then as the folks respond they can have that conversation, they can take notes verbally about each conversation and have all that recorded within their CRM system for them to access later on. So that’s one of the major trends that we’re seeing as conversational UIs associated with agent tools.

Now, other companies are taking the same approach in creating conversational UIs for consumers. So a consumer can come in and say, “Hey, give me a list of houses in this neighborhood that fit my criteria or fit this criteria and create a new playlist for me to look at on my phone app or maybe even on my television.” That’s another trend we’re seeing is the same types of convenience factors with a conversational interface here in the consumer market as well, and I think you’ve seen it as well if you look at Google system all those types of things that are happening.


Right, and just for the audience who might not know, UI is user interface, correct?


Correct. It’s a way of interacting with the tools. And so we’re doing the usual UI channel, so we use interface channels. So a web application or a web interface, a mobile interface and now to that we’re adding a conversational interface.


Okay, great. Well you pretty much answered my question about the purpose of the feature because I think in a perfect world that’s what a lot of agents would certainly want. So my final question is this, what are you most excited about seeing in the new future as far as real estate technology?


Well there is a big push now to leverage all the data in real estate and create AI predictive models. And so we’re spending a lot of time right now and I’m pretty excited about one, being the largest brand in real estate we have access to more data and more both historical data going back 35 years and current data generated by our agents and their clients.

We have a stockpile of data, that data is the new oil for AI initiatives and we have identified many dozens of AI outcomes that we want to bring to market for our agents to better enable them to serve clients. And so these outcomes are think of us predictive models that we’re creating. I’m really excited about the ability to enhance what an agent does in their day to day operations either to reduce the amount of work that they have to do, or reduce the amount of cost in their business, or to even make them a better fiduciary for their customers.

So we’ve created things that allow an agent to start up an app on their phone which starts their video camera. They scan it across a home as they’re walking through and we can identify everything that’s been picked up in the video. So we know it can take a quick inventory of, “Hey, you have wood floors, you have granite countertops, stainless steel appliances.” This type of entry where you have a contemporary home with a big oak tree in the front yard. We can identify all that automatically categorize it and then use that information.

If this is a home seller of the agents then a walkthrough for to automatically generate a micro site for that home that the agent can then use on social media, share with the homeowner so that they can share it on their social media and create a quick marketing plan, all within 30 minutes of meeting that homeowner and doing a walkthrough of their home. That used to take and still does now could take upwards of a week to generate all those same assets.

And so using AI to again advantage the agent in their business as well as advantage the consumer by having an interaction with a smarter agent is pretty exciting for us, because we’ve got quite a few things that we’re working on that I can’t talk about right now because of IP protection. We’re patenting a lot of the things that we’re working on so we can’t talk too much about it, but just know that you’re going to be seeing a lot of announcements around what we’re doing in AI the rest of 2019.