Putting the "AI" in Analytics with Vishal Bhalla & Noel Zamot

In this episode, Vishal Bhalla, Founder & CEO of AnalytAIX, and Noel Zamot, COO of AnalytAIX, share insights into how their company is bridging the AI adoption gap in healthcare by making data insights more accessible, improving safety, and supporting human-centered care. Learn more at https://analytaix.com/

Episode Transcript

(00:00):
This is Gracelyn Keller with the Becker's Healthcare
podcast, and I'm excited today to be joined
by two leaders. First, we have
Vishal Bala, who is the founder and CEO
at analytics,
and we also have Noel Zamat, who is
the chief operating officer at Analytics. So thank
you both for joining me today. You also
let me know that you were classmates at

(00:21):
MIT, so I love that piece of background
history for you too. And I would love
to start our conversation by having you both
introduce yourself a little bit further for the
audience.
Great. Thank you, Grace.
It's a pleasure to be here.
We founded analytics
to get people to be able to frictionlessly

(00:42):
access their data. My background is primarily hospitality
and then health care.
Over to you, Noel.
Yeah. Grace, thanks for, having us here. Vishal
and I met at, at business school at
Sloan. He has a background in health care.
I do not. My background is, I was
an engineer, did MIT beforehand as a as
an engineer, and then flew jets for the

(01:03):
Air Force for a long time where I
did a bunch of work in what we
used to call autonomy in those days, and
now we more broadly call it artificial intelligence
and machine learning in aviation.
So we connected and it just made sense
to expand on some of the classes, professional
development courses that we found ourselves
taking again in AI.

(01:24):
And what we're trying to do, like Vish
said, is
make it easier for organizations to use a
lot of the hidden insights that are locked
away in their data. Most organizations just collect
and collect data, and they never get an
opportunity to actually do something about it. We're
trying to fix that. And based on the
experience of some of our teammates, we thought
that engaging in health care would be a

(01:45):
great place to start.
Absolutely. Well, thank you for taking the time
to join us today on the podcast.
And you said something before we started the
recording that I loved. You said you guys
put the AI in analytics.
So to start the conversation, I would love
to kind of hear your guys' thoughts about
all this buzz that's going on around AI

(02:05):
and where you think organizations should start when
they want to apply AI in a practical
sense.
Yeah. So I'll start off on this, Grace.
It's very important to
understand, you know, what is the problem
that they're trying to solve and look at
high friction points,
like employee engagement

(02:26):
or
understanding
reporting from various
like Noel said, you know, our data is
locked up in various platforms, and we need
to make decisions. We need all that information
instantly
in a frictionless manner
and in human language. And so that's key.
And in order to do that, just to
have the
identification
of the key friction points

(02:48):
and understand where AI agents can plug in
easily.
Yeah. One of the things, you know, I
I look at this from, from the point
of view of other technologies that have come
in and and changed industries, but we still
have fundamentally a lot of the same challenges.
So one example that we give to our
clients and our partners often is that the

(03:08):
fact that we created
our tools
did not make carpenters
obsolete. It just made the better carpenters better.
And more than any other technology before this,
that's what AI does. It gives us the
ability to solve problems, I think, at a
deeper level. So when we look at people
saying, oh, this is gonna be, you know,
the panacea, the solution to all problems,

(03:30):
That's probably not what it's gonna be like.
What it is going to be like is
that
you're going to be able to use these
tools in a way that's gonna be able
to solve your problems perhaps in a better
way than you've had in the past or
solve some challenges
that you haven't been able to solve in
the past. And that's how we're looking at
this. We're not looking at this as some

(03:51):
miraculous
cure for everything. We're looking at it as
a tool, as an assistant that can solve
some pretty significant challenges that we have in
hospital administration and health care today.
Absolutely. That makes perfect sense.
And going off of that, can you give
a a real life example of AI in
action inside of a health system?

(04:13):
I'll give you, one example of a client
that we have right now that we're working
with,
and that's workplace violence prevention in hospitals. And
we all know that that is not only
a pervasive problem, but also a horrific problem.
It has human implications. It has personal implications.
It has fiscal financial performance,
customer experience implications.

(04:34):
It touches everything.
And it turns out that, in many cases,
the way that hospitals have been responding to
this crisis
has been the way they've been responding to
it for years before. So we were very
fortunate to connect with a partner who said,
can you help us on this? And we're
taking measured steps to get to where they
want,

(04:54):
because of delays and, you know, getting stuff
behind the wall in IT is always a
challenge, but they were so desperate for the
solution that they said, can you just you
know, we'll give you data, provide us a
solution predictions,
and we've done just that. And the response
has been really,
unexpected.
The the first part is just being able
to identify and quantify the problem

(05:17):
and working with them to see how we
adapt these AI and machine learning tools to
their challenge was very instructive on both sides
of the fence. Then when we actually provided
the solutions,
two things came up. One was their most
experienced people said, yeah. That makes a ton
of sense. By the way, you went further
than we thought we needed to go,

(05:40):
but and then this was the second part.
Now you have data and science to back
it up. So we, you know, we were
leveraging,
you know, all told a century of experience
that they had in in hospital administration.
They thought that the problem was gonna look
this way. We were able to literally in
days
tell them, yes, and not only do you

(06:01):
need to go, you know, to ten, you
need to go to eleven,
but we have the data to back it
up. And it's their data. It's not our
data. It's collecting a bunch of information that
otherwise would have sat there and not been
used. And now they're able to actually
help their people, Not just their
patients, but the people who work at these

(06:22):
ERs who are gonna feel much better taken
care of because now the organization has actually
made an investment in their safety.
Absolutely. That makes so much sense.
And on the flip
side of that, I would love to hear
maybe some of the mistakes or issues that
arise
for companies when they are trying to roll
out new AI integrations.

(06:45):
Yeah. So there is a lot of
siloing, I guess, of the data,
and the data resides in different platforms. And
the definitions
may not be consistent across the board or
even the terminology.
So the foundation
of any technology I'm sure you've heard garbage
in, garbage out, but the foundation of any

(07:05):
technology is making sure we understand
what we call the metadata.
What does that mean? So if you're looking
to connect employee engagement
with patient experience, both qualitative and quantitative,
and then connect it to safety outcomes and
value based purchasing
or ACO outcomes and say, okay. Which of

(07:26):
my
ZIP codes or which of my DRGs
is you know, track it all the way
through? While technology can do that, AI can
do that very easily today, the problem is
if we don't understand the metadata.
And so metadata,
imagine columns and
rows. So the heading of each column just
understanding the definition of that and how it

(07:48):
relates to the headings in other columns. So
understanding the relationship
between the headings of each of the columns,
I'm oversimplifying,
but I think that is a key
problem to be solved.
Yeah. And I'll I'll add to that. You
know, Vishal and I are coming to this
from,
perhaps a slightly different point of view, and
that's because
we have actually

(08:09):
been educated, had professional education in AI,
both in the strategic
business side
and on the technical side. So we can
speak to both, but we can actually speak
based on experience because we know how these
tools work. To answer your question, you know,
what's one of the biggest mistakes
is that people
you know, what we have seen in some
cases is people applying these tools like if

(08:32):
they were a panacea. Right? This I'm gonna
sprinkle AI on my omelette, and all of
a sudden, I'm gonna turn it into, I
don't know, a turkey pot pie.
That doesn't work. There's you know, and and
Sloan has a a course on AI strategy,
and one of the things that they talk
about is using a continuum of capability for
artificial intelligence to these machine learning tools. And

(08:55):
one of the ones where where you can
create the most benefit with the least disruption
is using them as tools or assistance
and assistance.
So helping people do better work, but it
doesn't replace people. It doesn't necessarily,
in many cases, replace processes.
It helps you understand

(09:16):
those processes at a deeper point. It helps
you be more productive so that then you
can have better information
to change the processes
and, you know, maybe even the data that
you gather. So it's not necessarily
disruptive.
Done best, we firmly believe that it is
a growth where
we help the client

(09:37):
learn more by doing and therefore, they have
this, you know, the the adoption curve is
not a stepwise function with a lot of
disruption, but rather an accelerated
beneficial, you
know, approach where they get better by being
better at what they do because they're leveraging
these tools appropriately.

(09:58):
I love how that's phrased. Grace, if I
may just add one example of doing that.
So, you know, people don't necessarily understand their
benefits a % of the time,
nor do they read the emails that come
from benefits or a % of the time.
Right?
So what if
you could talk to a tech human

(10:18):
who understands you and your benefits, and you
can talk to them in your language
247
and say, hey. What about this? Or
and so on and so forth,
that perspective. So currently, you have call centers
answering phone calls for employees
that are not twenty four seven, obviously, those
call centers.

(10:38):
You can take all that load away so
that
if there is a high emotion situation,
then the human tech person can be involved.
And so that's a simple example of how
we have tools that you can leverage today
to make that happen.
And and I'll I'll pull on that story
because
we had a we're working with a with

(11:00):
a partner right now.
And in the first engagement, when we rolled
out this capability that we have, which is
essentially a tailored LLM
speaking through an AI agent, a visual agent.
We call her Pai. And Pai is answering
questions having to deal with all sorts of
benefits questions. We essentially programmed a benefits package,

(11:20):
and she can answer any and all questions
and find the information
throughout this wide body of information.
When we first presented it to this client,
they said, oh, that's
this is taking away my job. We said,
no. Tell me tell me what your pain
points are right now. And they came back
and said, well, you know, I have to
manage all these benefits
ambassador

(11:40):
whose
job as benefits ambassadors is a unpaid additional
duty that they have in the company. And
many times, they don't have the information and
it's challenged and it takes too much of
their time.
So we said, hey. Maybe this agent could
actually be the first stop. Somebody has a
question. Hey. You know, what about a smoking
cessation program? Is that covered in my in

(12:01):
my package? And if so, how much do
I save? And, you know, they have all
these questions. They can start asking those. And
then at some point, that agent will say,
you know, are you ready? If you have
additional questions, here's a person you can talk
to. So what we did was not replace
somebody. What we did was have an assistant
that makes the job easier. So

(12:22):
the nuts and bolts, the grunt work of
answering some very basic questions
was taken by an agent so that then
the person can solve problems at a higher
level. And and at the end, that's really
the hallmark of this technology. It's not really
replacing people. It's allowing done well, and this
is what we wanna do at analytics,

(12:43):
done well, it allows us to focus people
on problems what that they are much better
at solving.
I love the the way that you just
described that. I think that's so helpful for
listeners and it really anybody who's interested in
integrating AI, and wanting it to go well
in their system. So thank you for that.
And then how does analytics

(13:05):
turn agent interactions
into insights
that leadership can act on? Like, how are
we taking this data and really looking at
it in a way and utilizing AI to
do so to make informed decisions?
I'll start off with that. We have a
slide on our client presentation deck that basically
shows the overall process.
And it starts off with getting the data

(13:25):
as close as we can to the source.
It shows, you know, a little bit of
our AI layer, that we call and then
our our very
bespoke integration
that we use for clients. And what's interesting
is that, what we show is there's three
main areas or three main methods that we
communicate with clients. One is through forms.
Another one is especially, you know, an

(13:48):
an optimized,
large language model,
and then our our live AI agents.
And a lot of people go, man, we're
we're in the twenty first century. What are
you doing creating forms? Well, it turns out
that that's how our clients,
that's how their organizations
run.
So instead of us saying, let's reinvent the
wheel here. I have this tool, but I'm

(14:09):
gonna force all of you guys and gals
to actually change the way you work so
you can now start working with my tool.
We said, no. That's that's gonna make it
harder on the client. What if we use
our capability
to provide much better information
in the way that they already know how
to digest it. So now that agent interaction

(14:33):
is in a manner that they expect,
but the information that it provides
is much more meaningful
because it's tailored for them. It's tailored for
the way they think and the way that
they operate, and it's based on their data.
It's not somebody else telling them, you know,
I read up case studies about somewhere else

(14:55):
that happened. No. We are actually using the
data that's in your system to solve your
problems. You have spent, in many cases, millions
of dollars trying to collect this data. Let's
put it to use.
Let's put it to use in a way
that is easy for you to implement and
digest and implement and and, and execute,

(15:16):
and then we do it all over again.
That's our goal.
Well said. No. If I may add one
perspective, Grace. You know, I'll take a 30,000
foot view.
What we are basically doing is we bolt
on to any existing systems.
In one of my previous roles,
I had to pull together 26

(15:37):
disparate teams, all
had good intentions at heart,
to get to a single platform so that
we can look at employee engagement, patient experience,
safety, and financial outcomes, and say, okay. What
are the three things we need to focus
on
to improve our value based purchasing outcomes? Right?
Today, we don't need to go on a

(15:58):
single survey platform, so there's no disruption needed.
What we have done is we have built
a AI
tool layer.
So you can think of it as a
magic carpet that you put over your various
silos of excellence,
and then you can ask it questions. And
it'll give you answers

(16:18):
instantaneously.
And you can ask it in simple language,
and it'll give you responses
across qualitative and quantitative data in simple language.
And you can export reports.
So that's the intention
so that we can turn
interactions into insights.

(16:38):
If we were to look at the benefits
agent, so Pai is answering questions for, I
don't know, six, seven, 10 thousand employees,
then through the organization,
Pai can give reports that say, hey. Employees
are asking about option a or option b,
which we're not offering. Right? Maybe they're looking
for
gym memberships, but we are offering something else.

(17:01):
Or employees of this ZIP code or this
demographic profile are looking for x or y
and really provide those insights. Secondly,
is to be able to demonstrate.
Remember, we are a AI analytics firm, so
we help you also do predictions.
So we can help predict the consumption

(17:23):
of the benefits for the following year, thereby
enabling the organization
to drive
strategic
decision making
and to have stronger
footing when negotiating those
deals for those benefits.
Wonderful. Well, thank you for sharing that. And
my final question today as we wrap up
our conversation

(17:44):
is what are you both looking forward to
most in the next coming
months and years as this continues to to
become even further implemented and AI becomes an
even larger part of our workforce and lives.
I'm gonna give you a
somewhat
sobering but ultimately hopeful
view.
If you look at the state of AI

(18:05):
adoption today,
there's a small mark segment
of our
market, of our industries
that have already integrated AI and ML into
their operations, and they're already reaping the benefits
out of those.
The better capitalized you are, the more tolerant
perhaps you are. There's some friction

(18:26):
and, you know, and and some barriers to
internal adoption, but in general, if you have
the money, you
invest in these technologies
and off you go. Meanwhile, there's a significant
chunk of the economy that is frankly being
left behind because they don't understand how to
integrate these technologies
into their operations.

(18:47):
We have very
mindfully
focused ourselves
to fill in that gap.
We are trying to make it easy
for
stakeholders.
Right now we're focused in the hospital administration
and health care segment to integrate AI and
machine learning into their operations
in a cost effective,

(19:08):
disciplined,
non disruptive manner so that it doesn't become
an obstacle to their progress, but rather allows
them to keep up with their peer competition
and serve their client base better. That's my
strategic view for the future. It's it's using
this technology
for the greater good and ensuring

(19:28):
that a lot of people don't get behind
because they see this as that's either too
hard, too complex, too costly,
too difficult.
We are at the point in the market
where we can provide these solutions,
make them seamless, make them tailored, make them
cost effective,
and make them very effective. And that's what

(19:49):
I see into the future.
I'll just add two perspectives.
One is
we insist on building an ROI dashboard for
every single project. Because if we're not delivering
an ROI, then what's the point?
Secondly, in terms of the future, we have
to remember
in whose service. Right?
In health care, I say there are those

(20:09):
we serve and there are those who serve.
Right?
AI should be serving both in service of
the human. And so keeping that in mind,
keeping
our ethical
framework
in mind,
I think we, together, can ensure that we
reduce the friction
and get people to operate, as we say,

(20:30):
at the top of their license, but really
take away all the mundane stuff. And as
time goes on, give people more and more
opportunity
to connect with their purpose.
That's what health care is built on. Well,
that's what I love about serving in health
care is every individual is so connected to
the patient they serve, and we need to
remove all the red tape that prevents that

(20:52):
beautiful bird who wants that cake, the cake
being the patient. We put glass around the
cake. We we need to break that glass,
and AI is an awesome tool to help
make that happen.
Well, wonderful. Thank you, Noel and Vishal, for
joining me today on the Becker's Healthcare podcast
and sharing more about what analytics is and
how AI in general is transforming

(21:12):
the way that we work, the way that
we live, and the way that data is
computed and analyzed. So thank you for joining
me today.
Appreciate the time. Yeah. Thank you for your
hospitality,
and,
look forward to continuing connection and making a
difference in the lives of those who serve.
Wonderful. Well, thank you so much.