Skip to playerSkip to main content
HousingWire
  • 1 hour ago
Servicing is where technology meets homeowners on their worst days, and the systems behind it have spent decades falling behind. Omer Farooque, CTO of Sagent, joins Power House to explain how the team built Dara, a cloud-native servicing platform meant to replace legacy constraints with real-time processing and connected data. The first years went to fundamentals, not AI: mapping roughly 300 servicing processes and 3,000 attributes before layering anything on top. That groundwork is what makes agentic tools like Ask Dara work, handling plain-language questions and walking a homeowner through an escrow analysis (hypothetical in the example he gives). They also get into where AI could help borrowers in hardship and loss mitigation, why empathy and misuse safeguards need constant attention, and how Sagent keeps product features aligned with changing regulatory requirements. Farooque’s read on what’s next: more tailored AI, voice and visual interfaces, and live transcription, so homeowners and servicing teams spend less time fighting software.

Related to the episode:

⁠Zeb Lowe’s LinkedIn⁠
https://www.linkedin.com/in/zebulon-lowe-a02353a4/

Omer Farooque's LinkedIn
https://www.linkedin.com/in/omerfr/

Sagent
https://sagent.com/

Dara
https://sagent.com/dara/

Want more from Zeb? Don’t forget to subscribe to LendingLife.

The Power House podcast brings the biggest names in housing to answer hard-hitting questions about industry trends, operational and growth strategy, and leadership. Join HousingWire’s Zeb Lowe every Thursday morning for candid conversations with industry leaders to learn how they’re differentiating themselves from the competition. Hosted and produced by the HousingWire Content Studio.

Category

🗞
News
Transcript
00:00Welcome back to Powerhouse.
00:01Servicing is where the industry meets homeowners on their worst day.
00:05And it's running on technology that feels like it's been lagging behind for decades.
00:09On this sponsored episode, my guest is Omar Farouk, Chief Technology Officer at Sageint.
00:15He has spent the past four years building the platform meant to change all of that.
00:20Today, we get into what AI actually does for a borrow in hardship, as well as what it shouldn't touch.
00:36Hi, Omar. Thank you for joining me.
00:39Good afternoon, Zeb. How are you doing?
00:41Doing well. You ready to talk about any and all things servicing technology today?
00:46Hit me.
00:47Okay, will do.
00:48So you've been leading the development of Dara for over four years now, I believe.
00:54Can you walk me through, one, walk me through what that process has been like for you and your team?
01:00And can you share what the most, I guess, what the most rewarding outcome has been for you so far?
01:05Yeah, so this journey has actually started about seven years ago.
01:12I was cutting my teeth on getting a company and a data management product out to the market.
01:20And we just had finished getting our Series A.
01:25And Sridhar, who was a colleague of mine in years past, had reached out on what he was doing at
01:33Cooper, at Mr. Cooper.
01:34And it sounded interesting, and I'm like, I've covered, and I like to look at career as slices of a
01:43pie.
01:43So the first slice of my pie was an architect and a developer tinkering with stuff at a consulting company.
01:51And then from there, moving into product, from product, wanted to see what being a member of the firm would
02:00be.
02:00So that was my next logical step, and joining Mr. Cooper and trying to do something that was fairly complicated,
02:12fairly long-ranging,
02:15where, and it hasn't been done for about three or four decades, was appealing.
02:22And I chase those kind of challenges, or they attract me, whichever way you want to look at it.
02:27So when this came about, we started with trying to put together the pieces to modernize the mortgage servicing core.
02:36And at that point, I didn't even realize what I was getting into.
02:39Because back then, it was like I was coming from retail, aerospace, logistics, defense.
02:46And I'm like, there are so many transactions online.
02:49I mean, mortgage just has one.
02:51How complicated can this be?
02:53So I was kind of cocky when I started.
02:57And the first three months really humbled me and quick in terms of the complexities involved here and what it
03:05takes.
03:06So that was kind of my both hazing as well as onboarding into the mortgage space.
03:13People warned me that once you get into the mortgage space, it's like Hotel California.
03:17You can't really get out.
03:19And lo and behold, seven and a half, almost eight years.
03:24And I've learned a lot as both a member of the firm and then back into the product and technology
03:31seats of building and shipping product,
03:35which is probably my most comfortable space, most familiar space, and the one I release the most.
03:41So from Cooper into Sageant, we acquired the Cooper technology stack.
03:50Keep in mind, at that time, Sageant already had the LoneServe stack.
03:55So in addition to the LoneServe stack, the Cooper stack, and with the expectation that we're going to be starting
04:03a brand new modern cloud platform to handle all of this.
04:08So pretty audacious goals on day one.
04:12And we literally doubled the organization the day I joined.
04:16And then from there, over the next year, again, you know, added 100% capacity to support what we now
04:25call Dara.
04:26Right. So I had three literal tribes that we've kind of brought together and seen what kind of synergies we
04:35can leverage in terms of their learned experiences.
04:38Because if you look at this group of people, collectively, they have the largest and longest tenure of mortgage experience,
04:47right?
04:47Mortgage servicing experience.
04:49So figuring out how to use that, how to bring that to bear, and also applying it into how to
04:57do mortgage servicing in the 21st century was pretty interesting.
05:03In terms of learnings and what we've been able to do diligently, we started out with a simple concept.
05:11If you're familiar with the iceberg analogy, and everyone in a technology ecosystem talks about the iceberg analogy where they
05:19want to do the next best thing,
05:21but are hobbled by everything below the line where the legacy constraints are the ones that are impeding them.
05:28Right. So when you think about all of that, what is that?
05:31That's batch interfaces, file-based systems, fixed position capabilities, and monolithic systems.
05:41What if you don't have any of those constraints?
05:43What if you built from the ground up an API-centric event-first platform that is real-time in nature
05:52and is able to handle payments the moment they were posted directly onto the ledger?
05:57What if you did that, and the rest of your ecosystem is integrated end-to-end?
06:02It's a tall order.
06:04That's why it took the amount of time it took.
06:06But we have a couple of sayings or phrases that I use internally that have now become moniker.
06:15I'm looking at my laptop, and there are quite a few of these stickers on my laptop, which are our
06:20phrases or catchphrases that we use internally.
06:22And one of them is, and you do this why, don't panic, darification, and you got babacked.
06:29I won't go there right now, but my point is, the phrase that comes to mind here is, we eat
06:37our own dog food.
06:39So what I mean by that is, everything we've built is built on top of this data, events, and API
06:45layer.
06:46And we started from there making sure everything has a structure, everything has a place, and everything is connected the
06:52way it's meant to be.
06:53So if you think about mortgage servicing, it's about 3,000 attributes, and about, give or take, 300 processes.
07:00So if you have a place for each and every one of those 300 processes, and the attribution that describes
07:06them, and you built them ground up with that mindset, the rest, as they say, will follow.
07:12So that's what we did.
07:15And in these last three years, I've also had to be fairly oxymoronic.
07:22And by that, what I mean is, the first two years I said, we're not chasing the new shiny.
07:27We're not doing AI, we're not doing any of this.
07:30We're focused on brass tacks.
07:33I want to clarify, you said you're not chasing AI?
07:37You're not pursuing AI?
07:38In the first two years of this journey, if you just divide this up from the last four years, right?
07:44The first two years, I was fairly dictatorial that we're not going to chase anything new and shiny.
07:52By that, I mean, we're going to focus on making sure the fundamentals are done right.
07:57Right?
07:58And what I mean by that is, those 300 processes, the 3,000 attributions that describe mortgage servicing, needs to
08:05be done the way a servicer needs them to be done.
08:08Right?
08:09So we focused on those.
08:11We made sure they were a thing.
08:13We made sure they were connected in the way.
08:16And I'll give you examples, right?
08:18How you do recasts.
08:19How you do PMI removals.
08:20These are very well-known things that need to work a certain way.
08:26And we just, we need to make sure they are implemented that way.
08:31And that's what we did.
08:32So now these are all well-known things all the way through our screens, APIs, events, and data structures.
08:42Doing that, doing that enabled, back to my iceberg analogy, what it enabled is fast forward to last year.
08:50I said I was oxymoronic, right?
08:53So for the first two years, I said no new shiny.
08:56These next two years, I said absolutely everything new shiny, we're going to apply it on top of this.
09:02Because now that we built that foundation, the agentic capabilities, the MCP, the tools and resources that you would normally
09:09want to apply or layer onto all this, just logically fits.
09:15And lo and behold, once we started doing that, our agentic framework was able to infer everything that our developers,
09:24our product folks were doing naturally.
09:26So it was a very easy transition for our whole team.
09:30And it was very, very refreshing to see.
09:33That was actually going to be my next question.
09:35How are you guys tapping into technologies like AI to deliver value for both servicers and the homeowners that they
09:43serve?
09:43Yeah, so since late 24 but early 25, what we've also done is most of our development folks were the
09:57first ones to start with agentic capabilities to assist and help them accelerate their outputs.
10:03But at the same time, we also had our product folks focus on making sure they're able to leverage their
10:10agentic capabilities as needed.
10:13So think Copilot, think Cloud, think a host of other utilities in the middle that would assist them in doing
10:21what they need to so that what we call the PDLC lifecycle, the product development lifecycle, is a well-oiled
10:31machine.
10:32And can be, right, and can be as continuously optimized as possible.
10:36So when we started this, it was still SDLC, software development lifecycle.
10:41And we were more focused on making sure the stages within that pipeline are as optimal as possible so we
10:49can ship product really fast.
10:51In 2026, we've also switched to a bunch of our capabilities being completely agentic PDLC.
11:00So that's one end of the how we ship product paradigm.
11:04The other dimension here is about what we're offering to the market.
11:09So when you think about what we're offering to the market, every capability that our product comes with out of
11:15the box is now available through AskDara.
11:19So AskDara is a conversational experience.
11:22It's a voice experience that can answer pretty much any question that could be product user guide oriented, could be
11:31functionality oriented, could be transactional in nature.
11:35And the examples I'll give you and what kind of really surprised me was, I actually prefer talking to our
11:41AskDara voice agent than talking to a human.
11:45Because the question I asked AskDara was, hey, if my insurance changed, I was playing a customer and asking it,
11:55if my insurance changed, if my taxes for the year changed by $10,000, what would my payment look like?
12:04Now keep in mind, that question I just framed has layers to it.
12:10What it's really doing is it's a mock escrow analysis.
12:13You're asking us to perform a mock escrow analysis and respond with the what ifs.
12:18But I didn't use any of those words in that question.
12:22And what we were able to do was everything internally that we are already doing to perform all of those
12:28functionalities back to my 300 processes and 3000 attribution that's available through the APIs and the events.
12:35The agentic layer was able to learn and able to respond in kind very, very quickly and naturally.
12:41So it was able to tell me and it was able to do all the math, right?
12:45So even though all of these things are available on the screen, as well as humans can do it, they'd
12:50have to do math.
12:52And I have a thing about humans.
12:54They're not good at math.
12:56So the agent was able to rattle off these numbers fairly quickly, fairly naturally.
13:02And I was able to have a cogent conversation with him about what my changes would be and why.
13:09And then it would round up that whole conversation by asking me,
13:13So do you want me to actually do an escrow analysis if that is the case, an order evaluation?
13:20So it led me to a decision which normally would have taken multiple calls, letters, correspondences, and a pretty crappy
13:31customer experience, right?
13:33And all of this in less than three minutes.
13:35So that's the kind of advantage that a modern servicing platform with the data structures, the processes, and the attribution
13:46that allows for all of this to be communicated, integrated in the current age and be nimble and pliable for
13:57an agentic framework.
13:57When you layer all of these things up together, it is really refreshing to see what you can do.
14:03Let me ask you this.
14:05Since I've been in the industry, I got into the business as a loan originator over 15 years ago.
14:12And ever since I'm from day one, I have been hearing about how technology is going to upend the industry,
14:21change the industry.
14:21And on the origination side, on the lending side, it's do more loans faster, right?
14:27And reduce the cost of origination.
14:30And thus far, in reality, that just has not been the case.
14:35I think now with where we're at with AI specifically or most specifically, we're starting to see some lenders really
14:45kind of figure that out now.
14:47The technology has gotten to that point.
14:50But I'm curious just to know from your perspective on the – because I think especially LOs and lenders, when
14:58the loan is closed and it goes over to servicing, it's one of those out of sight, out of mind.
15:03But because of technology transforming everything just within the last two or three years so rapidly, so much to a
15:11deeper level than it has in the past decade, how do you see technology or modern technology transforming the servicing
15:23industry specifically?
15:24Yeah, so again, both the originations as well as the servicing industries are ripe for modernization.
15:33I mean, the only difference is how much older the technologies in each of these segments are.
15:42One's 30, 40, 50 years old.
15:44The other is 10, 20 years old.
15:46But it's still – I mean, the last time I checked, the origination systems are still thick client-based, right?
15:52And what I mean by that is you need a desktop application to work it.
15:55I mean, go to a millennial, go to a – go to somebody who's 20-something and ask them what
16:02a thick client looks like.
16:03The only thing they're going to tell you is the applications on their computer that come out of the box.
16:09They do not expect to install anything else.
16:11So my point about that is there are existential constraints that you're hobbled with.
16:18That was my iceberg analogy where I was talking about how if you have those constraints trying to e-code
16:24efficiencies, you're doing it at the peripherals.
16:26You're not able to address it at the core.
16:28And you make choices.
16:29You make trade-offs that are only going to get you that far, right?
16:34RPA was that.
16:35So a lot of RPA and bots was all of that where you were just bandaging on the peripherals versus
16:42in this environment, the one that I'm talking about for servicing specifically, you don't have to have any of those
16:48impedances, any of those constraints.
16:50You're not shackled.
16:52You can go from the start.
16:55And a very simple thing, payments.
16:59Payments are not real-time in mortgage servicing platforms.
17:02The reason is, on the mainframe, you process a transaction.
17:08It gets ledgered by the end-of-day processing.
17:11It has to wait for the end-of-day processing to finish in order for all of those transactions to
17:16tarry and balance.
17:17That's just the way these technologies were built.
17:21In our ecosystem, you don't have any of that because we ledger in real-time.
17:28We post in real-time.
17:29So there is no black box.
17:31There is no black hole of, I posted a payment.
17:34Did it go through?
17:35Do I need to check back tomorrow?
17:37Day after?
17:38Oh, heaven forbid if it's a weekend.
17:40That means I'm going to get processed on a month.
17:42None of that, right?
17:43So the compounding effects of all of those go out the window because a lot of that scaffolding that you
17:51built around that ecosystem no longer applies.
17:54You mentioned this a bit earlier.
17:56Why do you think that the servicing technology has lagged behind historically to the extent that it has?
18:03Well, that's a very simple answer.
18:05Most of the servicing entities are market or public-facing, have quarterly and annual statements, which means they have to
18:12make justifications of said investments to the market.
18:17And when you're operating on a six and seven basis point business, it is very hard to justify a multi
18:23-million dollar, multi-year effort to modernize something where not a lot of people, not many have succeeded, right?
18:31So making that kind of bet and expecting the market to say, yeah, we're going to cheer you on and
18:37champion you.
18:39The market is not forgiving that way.
18:41So for servicing entities to be able to, market servicing entities to be able to take that up, no bueno.
18:47So that's where you need the perfect unicorn to be able to take this up.
18:53Something like SageGent, something like Dara that was almost engineered to be able to manifest this role we're in and
19:02this position we're in.
19:04Right.
19:05You had mentioned earlier that you preferred, you were testing and you were kind of role-playing as a customer
19:10and you preferred speaking to AI as opposed to a person in that scenario.
19:15Can you expand on that?
19:17What are the AI agent capabilities and how can they be or are they able to be tailor-made to
19:24meet the needs of every servicer?
19:28Yeah, so it's an amazing question because we're at the cusp of natural language conversations and natural language conversations to
19:38the point of it does not feel like you're talking to something automated or an agent.
19:42It does feel like you're talking to some human on the other end, but you also know, and this is
19:48almost like, I compare this to the line that you have for resolutions on movies, especially animation movies.
19:57You can make animation movies really real, but if you do so, the viewers are going to make, are going
20:06to have a very visceral reaction of being scared.
20:10I don't know if you know this analogy.
20:13So therefore, animators, make sure they don't go past that boundary so that people can know that this is an
20:20animation and not real.
20:22It's the uncanny value, right?
20:25Yes, yes.
20:26Same way, I was struggling with the phrase, but you got it right there.
20:31Same way with the agentic frameworks and the agentic conversations.
20:35You want to be borderline effective, but not cross that to say, I don't even need to tell you I'm
20:43a human because that will scare people away.
20:45We're there.
20:46We're there right now.
20:47And I believe we should stay there because I'll speak for myself.
20:52As far as I'm concerned, once I know that the agent has the capabilities to be able to engage with
20:58me, I'm going to use that agent effectively to get to the outcome in the shortest path possible.
21:04So that allows me to ask my questions in whatever manner I need to.
21:09It will do the deciphering that it needs to and get to the outcomes we need to and will be
21:14done with the call in the shortest amount of time possible.
21:17Whereas with a human, you have a welcome script, you have a feedback script, and you have everything in the
21:25middle.
21:26I'm sorry, I'm not a social person by design.
21:29So that's why I prefer agents than humans.
21:34It's actually why I prefer machines to humans, even in my professional life.
21:40But don't tell my team that.
21:42No, that's all right.
21:42I'd rather be sitting out alone in the woods myself, but they gave me a podcast where I talk to
21:46people anyway for some reason.
21:50I was thinking through that and I'm thinking through your answer and curious to know how you guys are planning
22:00or working to improve the experience for, I don't know,
22:06really difficult moments for customers like delinquency or any other sort of a hardship.
22:14How are you designing around that?
22:17Yeah.
22:18So two aspects or two facets to this.
22:20One is the empathy garnering and empathy management as well as abuse management.
22:29So there are both sides to the coin where we need to figure out how to manage this.
22:34And it is a continuous effort.
22:35It is not a one and done.
22:36What I mean by that is the tools and resources that we have at our disposal allow us to continuously
22:42refine this and keep gaining and keep keeping up with both the regulatory environment and the changes there,
22:49as well as the overall human, you know, or I'll call it worldwide, you know, news environment that we're operating
23:01in.
23:02So both of those, what used to be called like a sentiment score, et cetera, now is a lot more
23:07to do with the empathy score, right?
23:10So you can actually manage those things.
23:12You can toggle those things.
23:13That's one.
23:14And then the abuse aspect of like intrusion for and something as benign as tell me the recipe for a
23:25chocolate cake.
23:26You could insert this into a conversation to determine what the response from the agent could be and use that
23:34to intrude into the overall, you know, experience.
23:38So protecting from that front, as well as being empathetic on the other are constants.
23:46We have to do it continuously.
23:47And this era that we're in is ripe for both protecting as well as abusing.
23:55And you just got to find the right balance.
23:56I wanted to kind of also take the time for another key theme we're seeing in the industry, right?
24:03While we've talked about the agentic and the AI capabilities, interop and optionality are also key fundamental value props that
24:13we feel as a platform and a SaaS fintech solutions provider we need to be at.
24:20And let me explain what we've been finding in the market.
24:23If you're a bank versus an IMB, your appetite for risk and how much you want to be able to
24:31adopt early or late will differ.
24:37As a platform provider, for us, it is crucial to be able to balance both and say, you know what,
24:44we can handle by giving you our capabilities out of the box as an agentic framework on top of everything
24:51else we provide for those 300 processes
24:53and the 3000 attributes that describe mortgage servicing end-to-end, we can do that.
24:59You can bring your own agents or you can use none at all.
25:04Just use our application online and be able to do your job completely manually.
25:09That whole spectrum is something we believe very, very strongly needs to be a value prop for any modern platform,
25:16especially given the spectrum of our target market customers.
25:22So, let's tie this into, well, I guess, Dara, this is kind of a toss-up between the two, but
25:32can you tie that into something more specific like how you're using AI to help homeowners navigate something like a
25:45loss mitigation waterfall?
25:48So, we already offer our consumer experience, which is completely online.
25:53You can use the out-of-the-box self-service capabilities right there by pressing a button, or you can
26:02have a conversational experience that can allow you to get there.
26:05It'll do the QRPC and all of the other, you know, gating points for you, send you text messages for
26:12uploading docs, all of that good stuff where you don't have to even engage with a human.
26:17Part of the empathy angle that we just talked about a little while earlier is in those situations of hardship,
26:22you probably don't want to talk to a human as much as possible.
26:25So, you just want to get done with the application and move on with your life, right?
26:31And the capabilities we're providing allow our customers, our clients, to be able to manage that as much as possible
26:40without a human involved.
26:42Obviously, underwriting and the other decision-making steps are crucial for a human to be involved, but they're back-end
26:49facing.
26:49They're not really client-facing anyways.
26:51So, I think we're at the cusp of addressing and doing the right balance for all of these with the
26:57agentic capabilities that I just talked about,
26:59layered on top of everything else we've been building for the last three, four years.
27:03So, it's just, it's the perfect, I don't want to call it storm.
27:07It's the perfect environment to be in where we're able to support anyone where they need us to meet them.
27:15And what about, you know, compliance changes?
27:18Yeah, yeah.
27:18So, again, you've got to think about, when we built something like this, right,
27:26we had to think about all of those aspects from the ground up to make sure they're cross-referenceable.
27:31So, I live in a world where matrixes, and again, the analogy with the Matrix movie is not too far
27:40off
27:40because all of these things have to be matrixed and connected and be a, you know, multidimensional mapping that we
27:49have to manage.
27:50So, Reg IQ, as an example in Dara, is exactly that.
27:54As we are shipping features of our product, every one of those features is cross-referenced through our own agent.
28:03Remember, I said we eat our own dog food.
28:04So, we use the exact same capabilities that I just talked about and are able to map those features to
28:13two things.
28:16One, what we do in our platform, and two, what the regulatory bodies like the GSEs, Fannie, Freddie, etc., are
28:24saying or changing or evolving on those on their web fronts.
28:29So, we scrape all of that information, create a knowledge graph, map all of that into our matrix,
28:36and provide updates to our compliance teams so they can evaluate with every release that's going out how compliant we
28:45are with features.
28:46And the answer there has to always be 100%.
28:49So, if the reg says something, our feature says something, it has to match.
28:56So, without that gate, we don't release.
28:59So, this whole mechanism of agentic capabilities allowing us to also manage the compliance angle of what happens when regulations
29:09change on an ongoing basis to our feature development velocity is crucial for us to keep up.
29:15And that's also why we now release in weeks as opposed to six-month yearly cycles with Dara.
29:23So, that's another key enabler that allows all of these things to happen.
29:27It just so happens that RegIQ also becomes the basis for us to go out in the market and say
29:33to our client base,
29:35you no longer have to keep up with all of this anywhere else.
29:39Back to my end-to-end platform analogy, you can do it all within the same platform.
29:44I think one of the themes and one of the most used words in this conversation so far has been
29:51evolution or evolve.
29:52And I'm curious, my parting question for you is, how do you expect servicing technology and AI to evolve over
30:01the next five years?
30:02And along with that, what is shaping your priorities as a CTO?
30:10So, I'm excited that what we're doing and what we're putting into the hands of our client base,
30:19they're going to be able to take it to places that we didn't even imagine.
30:24So, that's the first thing that I want to make sure your listeners, your viewers understand that
30:31this is probably the most exciting times for mortgage servicing.
30:36And again, interest rates and the market, the way it is, being what it is,
30:43this space couldn't have asked for a better moment to shine and were there.
30:50Second, the agentic capabilities.
30:57I don't think these are an afterthought.
31:01I don't think these are optional.
31:03In terms of adoption itself, you're going to have your earlies, your late and your laggards.
31:09It doesn't really matter, but it's here.
31:11It's here and it is a first grade capability that everybody should take note
31:16and start figuring out how to wield or weave into their ecosystem.
31:21When you look at, at least, I can't see five years out.
31:25I can only see, there's another phrase that we use in the office, right?
31:31One step, one day at a time.
31:33So, I'll take that and tell you at least what I see for the next year or two.
31:37What I see for the next year or two is everything that we've built,
31:41how can agentic capabilities that we have currently in a large language model
31:48be encapsulated in a small language model and be fairly contained
31:53so that your expectations from it are exactly tailor-built, tailor-made
31:59for what our clients need.
32:01And our clients are every medium to large to very, very large mortgage services
32:08and they're going to have their own unique needs.
32:10So, how do we make sure that we can tailor these things
32:15to give them the exact capabilities that they need and nothing more?
32:20That's one.
32:21But the second is also, I don't think we need to have a user interface reliance
32:30as much as we have.
32:31I think voice, I think visual aids are probably where the experiences
32:37are going to go next, not just for the customer.
32:40But even internally for your team members and how they are going to start using stuff.
32:45voice and visual enablers are probably where we're going next.
32:52So, when I look at our client base, I'm talking about the team members
32:59or back office, front office users of our platform.
33:04The amount of typing they do with the level of sophistication that transcription has just taken off.
33:11It just negates the point.
33:12I mean, I see my mom using this and saying what she wants to me in a message
33:18and it transcribes it and sends it.
33:20So, now I finally understand how she can type so fast.
33:25So, imagine the exact same thing when you're having a conversation with an agent
33:29and the whole conversation is transcribed, captured as a correspondence or a communication
33:36in your platform live.
33:38You don't need to do a thing.
33:40And your SLAs, everything else is measured for you.
33:46You can just focus on what you need to, which is engage with the customer.
33:50So, that's the world we're going to.
33:53So, the user interface may not even be a thing.
33:57And when I say user, I'm talking about the web app itself.
34:00Yes, it's going to be there.
34:01But the reliance on it, I am looking for the age where that becomes a non-factor.
34:09Does that happen in two years?
34:12All appendages crossed.
34:15Omar, thank you so much for coming on the show.
34:17This was an excellent conversation.
34:19Pleasure speaking with you.
34:20And I look forward to our next one.
34:23Absolutely, Zab.
34:24Same here.

Recommended