in

You Only Need to Ask | Rohit Vashisht from WhizAI

Whether it’s baseball or bioengineering, data is everything now. We’re awash in dashboards showing us so much information from so many perspectives, it can sometimes make us wonder what we’re supposed to be looking at and why. But the goal has always been to get the information we need—and quickly.

Rohit Vashisht wondered if there might be an easier way. His solution became Whiz.ai, a system designed for life sciences that uses AI and natural language processing to deliver nearly any information requested  in milliseconds.

In this edition of UpTech Report, Rohit tells us where this idea originated, how it all works, why he chose to focus on life sciences, and the broader impact it could have on the industry.

More information: https://whiz.ai/


Rohit Vashisht is the co-founder and CEO of WhizAI, the first and only purpose-built cognitive insights platform for life sciences, empowering users to get answers to their business questions by simply asking via voice, text on web and mobile.

WhizAI is trained on life sciences data [867, IQVIA, CTMS, and APLD] and business terminologies [NRx, TRx, PBMs, Plans, TAs, etc.], enabling it to answer even the most complex questions from billions of records in seconds. Fast, easy, and scalable, WhizAI is the trusted partner of choice at the top global life sciences companies. Asked. Answered. Instantly.

WhizAI enjoys a 100% adoption by all business users, contributes to 60% reduction in calls to IT and 50% reduction in operational overheads within first few months of implementation for enterprise pharmaceuticals.

Rohit has over 20 years of experience in enterprise software sales, product management, development, and strategy. As a tech entrepreneur, Rohit has built successful businesses and was CEO and co-founder of Sverve – an influencer marketplace that was acquired by Bloglovin and rebranded as Activate. At Sverve, he led the revenue growth and development of a self-serve platform connecting brands and influencers for marketing campaigns. At Activate, he spearheaded product, sales, and marketing strategy to grow the influencer marketing business.

DISCLAIMER: Below is an AI generated transcript. There could be a few typos but it should be at least 90% accurate. Watch video or listen to the podcast for the full experience!

Rohit Vashisht 0:00
We are building an intelligent automation machine, which understands Life Sciences, industry, data, analytics and vernaculars. So well that it can actually do things, which would take weeks now in seconds on the fly.

Alexander Ferguson 0:20
Welcome to UpTech Report. This is our applied tech series UpTech Report is sponsored by TeraLeap. Learn how to leverage the power of video at Teraleap.io. Today, I’m very excited to be joined by my guest, Rohit Vashisht, who’s based in New Jersey. He’s the co founder and CEO at Whiz.ai, welcome, it’s good to have you on. Happy to be here. Thank you for having me. So Whiz.ai, if I understand correctly, is a pre trained conversational AI platform, that your whole focus is to connect business users with their enterprise data, insights and workflows. So in a nutshell, if someone in the life sciences, I believe you’re particularly focused on if that worker there says I want to know a specific answer to my question, and they can just type it or talk to your platform, and then it goes through all the data that’s inside the enterprise organization, and brings back the answers that I get that right? No, you’re spot on. Okay, so tell me how did how did the this whole concept begin? What was the problem that you initially saw and say, Hey, we need to solve this?

Rohit Vashisht 1:19
Sure. Let me connect the dots here. Like how we arrived at this right, me and my co founder, we’ve been in business intelligence space for almost 20 years now. So we build a product, you know, in our private life, which is used by Fortune 500, companies for the management, price management kind of analytics. So in 2017, you know, we were, we were just talking and B, we were looking at the BI market in general. And we thought, you know, nothing much has changed in last 1015 years, in 2004, or five people were building dashboards in 2017, people were building even more dashboards while the whole world became even more data driven. Every job is now data driven job, which does not mean every person has to be a data analyst. Right? So our idea was that, you know, can we bridge that gap? Can we disrupt this industry, you know, where end business used to be the salesperson, marketing person, executive, you know, r&d, or you name it? Can they simply get information that they need to do their job, without going through tons of reports and dashboards and complex software? That was a thesis, right for further study AI. And that’s what we are trying to accomplish. So you know, we have, we have found a lot of success. In last two or three years, we decided to focus on life sciences. And, and our AI is now smart enough, it’s so smart that you know, our end business users, they can actually get the information they need in seconds in milliseconds. And many of the times, you know, they don’t even have to ask for it. It just basically finds them.

Alexander Ferguson 2:52
What’s so interesting about the space and time we’re in right now is, we’re awash in data. Like, there’s so much data, and we’re tracking so much everywhere. And in many cases, we’re tracking it in many different platforms. And so we have to log in over here to find this answer. It looks through this content over here. And it can be annoying to search through to find it. But this is where aggregation or curation actually, to the point of, where’s I need this, that’s where you guys are really focused on someone. And across the industry, the company from marketing, you said HR to, to engineer or design to sciences, they can ask a question at all can find the answers.

Rohit Vashisht 3:31
Yeah, I want to clarify that. Yes, we do. And I we work with like different functions or different groups in life sciences companies. But we are very focused on structured data. It’s like insights, right? Like, for example, if you’re, if you’re if you’re a commercial team, you want to know, how are you doing in the market in different territories? Which are the doctors who are writing more or less about you? If you are market access team, you want to know, you know, how the payer universals is working? Like, are we in different plants? You know, our drug is covered in different plans, and how and all that kind of stuff, right? It’s very complicated information. I mean, so think about it. We work with top global pharma in the in the world, right? And they probably have billions and billions of transactions. And, and, you know, there is no way that they can actually before ways that they could actually put it in one place and make it available to everybody, right. So they would piecemeal, there are the small data here, the small data there, and people are going to like 500 different dashboards to do it. Now with this, we are sitting on those billions of records and anybody can actually ask question, open ended question and was smart enough that it can answer that from that data?

Alexander Ferguson 4:40
Know, walk me through the answer. So somebody types in or or speaks to, like, Hey, I don’t want to say Siri or Alexa, I have the devices. The two words together, but they could talk to the device or type in and how does the response How is it given that is it just you Is it textual? Is it just show a map like the chart and go over here? Look at the data? How does it show?

Rohit Vashisht 5:05
It’s visual? It’s purely visual. So for example, you know, if I’m, if I’m VP of sales, and I want to know, you know, hey, which are my top growing territories, let’s say last quarter for a particular truck, right? visible, I understand that question. And in this case, you know, we’ll pull that data, make that calculation. And then we have something, what we call as visualization AI, which is a pre trained machine learning algorithm, which actually looks at the question of the data that came back. And this then decides what is the best way to represent that data, whether it should be a pie chart, or map or table, and then it will create that on the fly and give it back to the user. And all of this happens in a split second, you know, so it’s the the responses are highly visual, they will have added commentary, if it needed to be, they will highlight things that you know, which needed to be highlighted that, hey, look, you know, these three regions, probably, you know, they need your attention, that kind of stuff. So, so yes, you know, the long The short answer is that, you know, visual visual responses, that’s what we generate.

Alexander Ferguson 6:08
We’re all visual. The large majority of people are live visual learners, we all appreciate visual. So being able to, I mean, having the data and looking through, it is nice, but eventually you just want to see it in a nice chart. Or traditionally, though, how would this had been done a probably a VP, who was working here, say, hey, I need this information that go to one of their managers or assistants and say, hey, go find me this answer. Is that how it would have been done?

Rohit Vashisht 6:34
Yeah, I mean, so you know, think about it, it’s a paradigm shift. Today, and we call it legacy, you know, people who are using legacy tools or solutions, the way the whole chain works is, I’m a VP of sales, I want to see this, if it is not available, I’m going to call somebody like business ID or ID and say, This is the kind of information I need. Now, they will go and start working with the IP team, data team who will put together data, then some engineers will create that analytics, and then the change management happens. And by the time it comes to me, I’m the VP of sales, it’s like a few weeks. And who knows, I mean, so by that time, I need something more now. And then then the cycle repeats, right? So that’s how it happens today, it takes weeks and weeks, a lot of manual effort. And by the way, these people, extremely smart people who work on it, and they’re extremely talented and skilled people, right? Like if somebody, these are engineers, data, data management, people, people who understand the business very well, right. So that’s the kind of caliber of people who get involved in this stuff. But with us, if you think about it, our AI is now automating the, automating that whole process.

Alexander Ferguson 7:39
So are you effectively taking away jobs? I have to be very, like, Yeah, so the question like, how is it then changing the work for all these folks?

Rohit Vashisht 7:47
Sure, yeah, I’m gonna say, you know, the, the future of work is very different. Right? So, so think about it. You know, 30 years ago, probably data entry was a great job, you know, you could go and make money and have a good career. But, you know, now fast forward 30 years, not so. But those people, they retrain the skill, and now they’re doing like better jobs. So so what we are doing here is, we are building an intelligent automation machine, which understands Life Sciences, industry, data, analytics, and vernaculars. So well, that it can actually do things, which would take weeks now in seconds on the fly, right? But this machine, you know, which we have built, this is AI powered. So people, you know, what we are telling people, instead of you just answering same question over and over again, building that pie chart again, over and over again. Instead of that, you should actually re skill or actually the, you know, these people are smart people, obviously, they’re they’re knowledge workers, right? They’re already learning AI and machine learning, they should be actually managing our models. There are tons of machine learning models, which are running in our product, natural language processing that is happening in our model visualization here that is happening in our model. You know, the idea is that you know, now how can you actually manage and evolve those models to your business needs? Because the plumbing work will be done by the machines anyways. Yeah,

Alexander Ferguson 9:16
no one wants to do. Let me push a button here, put your bed we will all want to grow and be able to have more interesting things to do. So I’m curious, like, what do you see is the future of is, is everyone just now going to work on more critical thinking mindsets of content? Or what do you see?

Rohit Vashisht 9:34
Creativity, critical thinking? You know, building intelligent machines. That’s accouter how smart

Alexander Ferguson 9:45
how smart Can I can AI truly be like that’s another question is like, okay, we’re AI is helping here it’s able to provide answers, but can it truly understand the data and where where does it balance off of the new need human insight to provide input

Rohit Vashisht 10:02
in a narrow domain, around the problem here could be very, very smart. The artificial general intelligence AGI as people say, we are very far from that I’m gonna see, you know, obviously, big guns like, you know, Elon Musk and Mark Zuckerberg, they have their own point of view. But, you know, in my in my purview, what we see is we did not, and so you know, think about this, we, we are not competing with Siri or Alexa, we are not trying to build a machine, you know, which can which a toddler can talk to, and maybe, you know, a young adult or even older people can talk to about different, different topics that’s very hard to build. But we actually pick this area, because we knew this area very well. And then the next and then even within analytics, we focused on one industry, which is life sciences. So you can see, like, how we are defining the box, I’m going to say, you know, it’s it licenses a large industry and trillions of dollars. So it’s a big problem that we are solving. But we know that, you know, in this box, our AI could be very, very smart. And it can do that job very well. I would say even then today, when we go live, we can answer like, more than 95% questions so that you know, accurately. And that is like in days, and then it kind of gets smarter and smarter as a US like, like the last mile. And that’s the kind of accuracy that we can build with our product.

Alexander Ferguson 11:26
How long does it take for this model to let’s say, a new life science company wants to be able to sign up? I mean, do they have to you have to effectively ingest all of their data? And how long does it take them for your platform people understand and start giving answers.

Rohit Vashisht 11:42
We can go live in a very complicated Life Sciences and, you know, a very complicated Life Sciences client. From couple of days to a few weeks. It’s fast, it depends. Yeah, that’s, that’s it, I haven’t seen, you know, the way our product works as that’s what we call it pre trained. It knows lot of those concepts already out of the box. So that’s a starting point. And then once we connected with their data sources, it automatically pulls metadata and non data metadata. So for example, if you have sales records, you know, a sales record might say, customer XYZ, this time, this tour, this product, this price, this volume, right. So what we’ll pull is like, these are products, these are your customers, these are your geographies, that’s it that is married, so ingest that metadata and train itself. So that process is automatically connected with data source. And now it understands your business. And that’s it after that it is all about you know, like, what do you want to do, like, small thing here a small thing here, but 80 85% of the questions sought to answer right there. You know, so

Alexander Ferguson 12:49
when there is a specific company that has more unique terms, or internal terms, whatever, how does that training model have is Is it is it user interface that then like internal folks and start pushing and say, Oh, this is what this means and clarifying.

Rohit Vashisht 13:04
Yeah, I’m gonna say, you know, we have built like, different types of training models. So the very, I wouldn’t say simple, but, uh, you know, very common, like, for example, Robert Wood Johnson is a hospital, but we call it our diplegic. Right? If you just want to do that, that’s one second change, you can go to our interface. And you can say, you know, what, Robert Wood Johnson is also a DJ, Liz loves it. That’s it. And after that, it knows that other DJs Robert Johnson, by the way, other DJ could also mean Robert Wood, rW j, Barnabas, which is another entity. And if this sees those two things, it is intelligent enough that it will ask you if you say, hey, what were my sales that are wha it’ll say, What do you mean are the Robert Wood Johnson hospital or other carnivals and you said this, and then it starts to learn your preferences and can do that. So a lot of smartness can happen, you know, is built into our product. But that’s like very simple, which people can do it in seconds, all the way where you can literally teach with your business, we call it business cognition. So, which is pretty unique to us. And we believe that you know, ai needs to be at that level to kind of answer these kind of questions. So for example, you will think about life sciences ecosystem and a lot of people probably understand this right? There is a form of company which has a drug, there is a doctor who writes a prescription that is patient who consumes that product, and then there is a payer or insurance company that pays for it. Right? So there are like different elements like so you can literally teach this and this already knows this now, but let’s say if there is something some nuanced about your company, you can literally teach this, doctors write prescriptions. pairs pay for prescriptions. And and now you know, you can you can pitch this to this, and now people can start to ask open ended questions, who sold the most, who bought the most, who wrote the most prescriptions, who are the new writers, these are very open Going into questions. And this starts to understand the domain and, and starts to answer these questions. So when you will ask the question who wrote the most this knows doctors write prescriptions. So, indirectly, this user is asking about doctors, maybe I should look at doctors and give, give that answer. That’s very, very unique and differentiated for us for our AI.

Alexander Ferguson 15:18
Yeah, maybe to create that training is it’s them simply writing it like just Yep. Just say, save it. And it understands the context of what’s being stated.

Rohit Vashisht 15:29
Absolutely. Zero coding skills required to make that training happen. Zero coding skills, you can literally a business user can go to our interface and say, doctors write prescriptions.

Alexander Ferguson 15:39
It’s almost like you bring on a new hire who’s unfamiliar with your industry and saying, hey, doctors write prescriptions. Okay, great. Yeah, no,

Rohit Vashisht 15:49
that’s exactly how we thought about it. When we started building it, we are like, it’s a toddler, and we are teaching a toddler, you know? And that’s exactly how we think about it.

Alexander Ferguson 15:59
So you saying that? I’ve heard that reference before? I mean, in many ways it is. Computers are able to learn like toddlers, but there are people that are concerned because they’re like, do I want to get my business over to a toddler? Like, like, Where? Where is the comfort level of the knowledge of is this data accurate? How do I know if I can really trust the answers that is giving back? Sure.

Rohit Vashisht 16:21
And that’s a process I’m in? So that’s a very common question in our sales cycles, right? Our customers will say, okay, you’re gonna answer, but how do I know that this data is accurate, calculated properly, assumptions are built in properly. So that’s the process that we go through, like when we set up and you know, we do like those matches, we convinced them to like, this is how it is happening. We In fact, share like, every time I answer is created or generated by this, we actually even describe how we arrived at that answer. So if somebody wants to actually look at that, so we can actually see that too. But I think, you know, what we are seeing is when we go live, users are smart, people are smart, I’m gonna say, you know, they, I would say, you know, and you know, like, for example, you look at sales people, and you might think like, oh, that person is not very technically savvy, but they know their business very well. a sales rep knows how many prescriptions are written in his or her territory every week. And if it is off, they can very quickly figured out that, you know, this is this is not right. Right. So so they get that comfort factor over a period of time. Also, very quickly,

Alexander Ferguson 17:24
actually, your target market, obviously, Life Sciences, this is where you build the product for Is it a particular size of a fabric company? or whatever? Or or what, who, what size? Are you working with?

Rohit Vashisht 17:37
Yeah, I mean, so we started at the, the top of the funnel we are working with, like, right now seven of the top and global pharma in some capacity. So that’s where we went first. Then we started going mid market. Now we have some clients who are like, for us mid market is like, anywhere between one to 5 billion in revenue. So now those are the farmers. So very recently, like two or three weeks ago, we actually have launched our self survey platform for very small pharma and biotechs. Yeah, so those companies, you know, they don’t need IP, they don’t need like services providers, they can actually come to our SAS put in their data. And that’s it. Everything works. So So yeah, so we want to work with every segment of the industry,

Alexander Ferguson 18:22
you got me interested, because I’ve spoken to a lot of folks that are applying some amazing AI tools like this conversational AI, but it’s only for enterprise because of the complexities that come with having to tune it to that company and the massive data that comes in, but how have you been able to then make it work for self service and still have that quality?

Rohit Vashisht 18:44
Yeah. So I’m gonna say, you know, like, that’s the best part about life sciences business that we figure, right. Even the industry is very unique and differentiated. They have their own terminology and a lot of nuances, you know, but once we got them into our product, there is a lot of uniformity when it comes to how a business operates. Right? So think about life sciences, maybe licenses one to one, especially on the sales side, right? Who, who basically sells a product, or who is the influencer? who sells a product that’s a doctor doctor prescribes, and patient consumes right? Now, it doesn’t matter which pharma company you go to. It’s the same number of doctors, same doctors who are prescribing. So if we understand like, 5 million doctors in the US, that’s a universal, every pharma company, you get it. And think about pairs, think about hospitals, we understand all of those and their variations today. We know all the metrics like a farmer business needs, and our access to access and Dr. X is not so there is humongous amount of uniformity from client to client. That’s why we are able to kind of offer this kind of self serve platform.

Alexander Ferguson 19:55
Every every new client you bring on you’re able to understand it further and further further. So As the final amount that do come on, it’s no big difference. Because because it’s

Rohit Vashisht 20:04
so 90 95% is already baked in and 5%, we have already given them the interfaces that they can actually tweak it. So

Alexander Ferguson 20:12
you guys started in 2017. And I have 6070 employees or team members and you’re in, you’re growing, what would you say over the last four years? Or have you been most surprised of how people are using the product?

Rohit Vashisht 20:26
It’s kind of interesting, I’m going to figure like, we just had a vision four years ago that this kind of thing can happen. And, and it’s fascinating, you know, once you go live with a product, and you start seeing people like how they use it, it’s just amazing. It’s just fascinating. So for example, Case in point, right, like our very first implementation, to a pharma company, was here in Jersey 303, and people’s sales team, hospital based sales team. And, and, you know, their sales team was not able to use dashboards, you know, which they had previously. And in fact, the company had a hotline, that the, the rep sitting in a doctor’s office or a hospital lobby will call and get those answers, right. So our first deployment was simple text messaging, SMS, we sat on their data, and we said, okay, this is a phone number, you basically just text and we were sending those visual responses on the fly as a text message, like, MMS message, right? Like if a chart was there, it’s a picture of a chart, it will come it will text the picture, you see. Yeah. So we got our first testimonial, you know, in the very first week, and the salesperson, she said, I was sitting in this lobby of this particular hospital, and I could get my information out in 20 seconds. She’s like, an anti use words, like this app is idiot proof. This is so amazing. It changed my life. It was just so amazing. And so you know, just just just that whole thing. It’s, it’s pretty cool. And then, so that was just the beginning. And obviously, like, you know, teams of 5000 sales reps, 10,000 sales reps, we are live in 27 different countries in five different languages, the nuances that are out there, it’s just amazing. And it’s it’s growing leaps and bounds.

Alexander Ferguson 22:18
Well, it’s an exciting an application of the technology that that you really honed in on. And I’m intrigued to see where you guys go next. Now I imagine that there’s a whole story to the journey that you’ve been on. For those that want to hear that journey. Stick around for part two of our discussion, where we’re going to hear Roy’s journey of starting this and before as as a tech leader, so stick around for that. But for those that want to learn more about Whiz, you can go to Whiz.ai. Right. And then what’s a good first step? Is there like a I guess they can reach out for a demo?

Rohit Vashisht 22:49
Sure, yeah. That’s the best thing. And the more when you see our demo in five minutes, you will know what we do.

Alexander Ferguson 22:55
And you’re like, yes, this this makes sense. Well, thank you so much for hit and everyone. We’ll see you on the next episode of UpTech Report. That concludes the audio version of this episode. To see the original and more visit our UpTech Report YouTube channel. If you know a tech company, we should interview you can nominate them at Uptechreport.com. Or if you just prefer to listen, make sure you’re subscribed to this series on Apple podcasts, Spotify or your favorite podcasting app.

Bots for Bookkeeping | Enrico Palmerino from Botkeeper

From Halo to CEO | Renji Bijoy from Immersed