November 18, 2025 • 25 mins
What if benefits work leaped from horse‑and‑buggy speed to highway pace? We sit down with Julian Lago, co‑founder and CEO of BenePower, to explore how AI is already compressing days of quoting, enrollment, and service into minutes—and what that means for brokers, HR leaders, and members who need clear answers right now.
We dig into the practical side of agentic AI and retrieval‑augmented generation: how specialized models fetch plan rules, accumulators, and pending claims to answer real‑world questions like “what’s left on my deductible?” with accuracy. Julian explains why virtual care is no longer just telemedicine, how ambient documentation can free clinicians to focus on patients, and why unified front‑door experiences beat a jumble of point solutions. Along the way, we talk about emotion AI that detects stress, switches languages seamlessly, and brings empathy to urgent moments—like getting an ID card to a parent driving to urgent care—without losing speed or precision.
For employers, the conversation moves from tools to outcomes: steering to high‑value care, reducing surprise bills, and designing plans people actually use. For advisors, it’s a playbook for differentiation—showing clients how to use data, automation, and hyper‑personalization to improve health and lower costs while keeping a human in the loop for sensitive decisions. We also address the hard edges: privacy, HIPAA alignment, model hallucinations, and emerging state regulations that set boundaries around clinical advice.
If you’re ready to trade legacy friction for clear, measurable gains, this episode lays out where to start, what to watch, and how to scale. Subscribe, share with a colleague, and leave a review telling us which workflow you want AI to fix first.
This episode is sponsored by Benepower, the platform of choice for a modern benefits experience. Benepower is an AI-powered benefits platform offering access to top products and services, enabling consultants and employers to create customized plans, optimize usage, and measure effectiveness. www.benepower.com
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
David Saltzman (00:00):
If the leap from horses to automobiles reshaped
an entire century, what will AIdo to our industry?
We'll find out on this episodeof Shift Shapers.
Announcer (00:13):
Change either energizes or paralyzes.
The choice is yours.
This is the Shift ShapersPodcast, bringing the employee
benefits industry interviewswith individuals and companies
who are shaping the industryshifts.
And now, here's your host,David Saltzman.
David Saltzman (00:33):
And to help us answer that question, we've
invited my old friend JulianLago, who is co-founder and CEO
at Benepower.
Hey Julian, how are you doingtoday?
Julian Lago (00:42):
Hey, David, great to be with you today.
David Saltzman (00:44):
My pleasure, my pleasure.
So you often use that very samehorse to automobile analogy.
Why is that moment in historythe best parallel for what's
happening with AI today?
Julian Lago (00:55):
You know, David, it it's sometimes it's visual.
I remember seeing uh an imageyears back of um, I think it was
Avenue of the Americas in NewYork, and there were two
vehicles and several horse andcarriages there, and then they
go less than five years forward,and it's the reverse.
There's basically one to twohorse and carriages, and they're
(01:18):
all automobiles trying tofigure out how to maneuver.
And it had always struck methat in such a short window,
technology, which was back inthe day, in fact, it wasn't the
auto industry, right?
It was the horseless carriageindustry somehow evolved and how
many people got displaced andso many things had happened.
So that that's been one ofthose things that are discussed
at a lot of business schools,but the image of that was always
(01:40):
uh interesting, that it wasjust a five-year window, and so
dramatically the changeoccurred.
David Saltzman (01:46):
You know, I when you talk about change though, I
mean, AI is not really realreplacing humans, but it's kind
of amplifying them, isn't it?
Julian Lago (01:55):
Yeah, it's an interesting dynamic.
I talked to my my engineers,and we're all using AI, right?
It's the it's the hot topic.
But machine learning, um, I'vegot an engineer on our team that
was with a pretty sizableorganization for seven, eight
years, and he laughs.
He says, Yeah, everybody'stalking AI like it's it's a new
candy, new wrapper, it's a brandnew name.
But it's been around, machinelearning has been around.
(02:16):
I think we've been exploringand seeing it in movies and how
it's going to take over, youknow, and be careful for that
robot that takes over theuniverse.
The reality is that almostwithout exception, through the
history of human mankind,technology makes um uh presence,
whether it's the wheel, whetherit's the you know, automobile
industry, whether it'selectricity turning on, in this
case, AI, the next progressionof the digital experience.
(02:40):
And human in the loop is notsomething that's gonna go away.
You need those components.
So will there be reshifting?
Will people be doing differenttypes of jobs?
Will some of the day-to-dayfunctions or responsibilities
they have go away?
But that means it frees them upto do other things.
And and that's the excitingpart.
You know, we basically uh donew things or have access to new
(03:02):
information.
David Saltzman (03:03):
So let's let's focus this in on our our our
industry and and daily,day-to-day benefits work.
What would a true human AIpartnership kind of look like?
Julian Lago (03:14):
Well, it it you know, if you think about the
process of this, you know, we'rewe're in the sales business in
some ways, where we're bringinga product to the marketplace, a
better solution.
There's shopping, there'squoting.
Um I almost a best way to lookat what the future is going to
be is to look backwards, right?
There's been uh you shared withme a great book one time.
(03:34):
We talked about the thegentleman turning on the
lanterns, right?
And um in the streets.
Think about what it was toorder a simple proposal.
In the back, I used to rememberto fill out a proposal request
form, you submit it, and itwould take a day or two for that
to just be back in your inboxto go present it to a client.
So what was the expectation?
You met with the client, youdid a fact finder, you found out
(03:56):
what their issues were, and youlet them know three to four
days, I will have some proposalsback.
So let's schedule something fornext week.
That process is happening inminutes today, right?
Because we can go in and put ina request for proposal, the
system might digest it, itidentifies it, and literally
within minutes you could bedoing that real time.
So that movement allowed us todo more.
(04:17):
We can we can do multiplequotes within a day, whereas we
were floating and we're movingphysically to go deliver that
presentation, etc.
So our efficiencies increase.
It's basically a scalingcapability.
We're able to scale ourbusiness faster and more
efficiently.
Move that into open enrollment,move that into customer
service, which AI plays asignificant role in.
(04:38):
Move that to the educationcomponent.
We're dealing with thecomplexity.
The complexity of healthcarehas not gone away because
technology is here.
So there's still a complexityof it.
It's not a subject that we dealwith every single day.
When we get ill, it's it's it'sfirst of all, it's important.
It's something we need toresolve, and many times it's
touching our family or ourlifestyle.
So it's a very escalated levelof importance.
(05:01):
And you want to bring all thesolutions to the table.
So, in that, um, that is wheresome of the technology, we're
going to want to maximize it andbe as efficient as we can.
So, efficiency and thosecomponents play a big role.
And I think every aspect ofthat, all the way through the
renewal cycle and ongoing, butmostly that customer service
(05:22):
experience and that education, Ithink AI is going to be a huge
differentiator.
David Saltzman (05:27):
Well, I mean, we're not we're not only seeing
it in those areas, we're alsoseeing it in things like
compliance monitoring, um,broadly enrollment support.
So if it's already reshapingcore processes, like which of
these areas do you think will besee the biggest transformation
with AI in the next 12 to 24months and why?
Julian Lago (05:45):
Well, I think if you think about the ability to
take information and turn itinto a digital component, um,
that bit of information um atour fingertips becomes
invaluable.
For the most part, other thanwith our pains and aches and our
ID card, we didn't walk intothe doctor's office with much
(06:05):
more information.
Today we're empowered.
We literally have all of ourdoctors, all of our networks.
We can do a lot of duediligence early on.
I hate to use the terminology,but you Google your uh your
medical procedure, at least youhave some idea what's gonna
happen at that doctor's office.
We can walk in much moreempowered.
That wasn't the case yearsback.
At best, you talk to a familyfriend or someone of trust and
(06:28):
you say, What's this experiencegonna be like?
And they gave you a vagueindication of that.
We're much more empowered.
Data is also allowing us totake our previous illnesses, our
previous comorbility issues.
If you have borderlinediabetes, that impacts what
you're being treated for, whatmedications you're on.
So information, in essence, isgonna be significantly easier to
(06:51):
come by.
Um, where to go, how to seekthe top physicians, how to get
the best level of care, mostimportantly at the lowest price
or most cost-effective price.
David Saltzman (07:00):
Well, we've talked about advisors a little
bit and employees a little bit.
What's this look like at theemployer side of the equation?
If if you're an advisor andyou're you're working with a a
C-suite person or an HR person,what does this look like?
How is AI helping them?
Julian Lago (07:16):
Uh I think it's the ultimate game changer.
The average consultant we workwith is looking to differentiate
themselves at that employerlevel.
Employers are looking at this.
Um, over the past year, we'veall heard that we're moving the
conversation from the HR to theC-suite, right?
Where it's a financialdecision, there's much more
(07:37):
information, and there's toolsto do predictive analytics, and
that's wonderful.
But what if we can actuallytake a deeper dive into how
their health plan and the waythat it's been built is actually
equaling or serving thepopulation that it's intended to
serve.
Just because you build it thisway, I can build a car with a
convertible, but you know, if ifif the person riding it wears a
(08:01):
toupee, they're not going toopen the convertible top, right?
Because the hair's gonna blowoff.
I'm using the crazy example.
But if if we're building plansin such a way that it doesn't
accompany the population that istrying to serve, it's not a
great solution.
If you're putting in solutionsthat people don't know about and
aren't aware how to use it,virtual health care, people
(08:22):
still perceive that astelemedicine.
And it's so evolved over justthe last several years.
Think about what COVID did withallowing physicians to practice
medicine digitally.
That's a transformation bothfrom their billing to their
methodology and the way they cancommunicate and all the HIPAA
compliance that goes along withit.
We have in some ways thoughthat's an outcome of what we've
(08:46):
experienced through COVID.
So today it's no longertelemedicine, it's virtual care.
That can extend to primarycare, even specialists,
certainly counseling andtreatments, is you know, medical
devices, home care, all of thatcan be touched now with the
digital component.
David Saltzman (09:03):
Do you think that all that stuff that kind of
got quote unquote forced on usduring COVID has opened the door
for a more ready acceptance ofsome of the things that are
coming down the pike?
Julian Lago (09:15):
Yeah, I think in some ways, unfortunately, um
we're creatures of habit.
If it's not broken, we're notgonna go out of our way to fix
it.
If you talk to the averageperson pre-COVID, if they had a
chance to speak a doctor face toface as it get to on the phone
and a Zoom, they'd probablyprefer it.
Um, think of the days where weall had expense accounts, right?
And everybody's, if they'regonna have a meeting, let's go
(09:37):
make it a meeting, I'll fly intotown, take you to dinner, and
so on and so forth.
Today, that becomes a 30-minuteZoom call.
We're here doing a podcast, andI'm not sitting in front of
you, we're using technology todo this.
Um, much more efficient.
Um, in fact, I can do threeother things throughout the day.
Whereas if I was traveling forthat one-on-one meeting,
physicians are in that medicalpractice is in the same way.
(09:59):
Think about what we complainabout having very limited time
with our physicians.
Physicians have to see morepatients to receive similar
reimbursements, and their timeis limited.
In this case, here we can bemuch more efficient because a
lot of the data is there.
Doctors are no longer worriedabout making sure they jot down
all the notes.
They sit there, turn on theirphone, and the AI agent hears
(10:22):
the entire conversation and isready to create and fill in the
ERM and also plan for your thatx-ray, read it, reduce the
medication, you know, lay outyour surgery, and if it's done
properly, all that comes in andbecomes part of your day-to-day
calendar and scheduling withouta lot of thought on either part.
That's where we're headed withsome of this great technology.
David Saltzman (10:41):
So, you know, a lot of the physician
relationships, vastly themajority of physician
relationships now are corporateowned.
They're owned by hospitalsystems and such.
Do you see an accelerator beingthe hospital systems themselves
bringing more AI to bear sothat their doctors can see more
patients in a more in a moretime-effective way?
Julian Lago (11:00):
That's a really interesting question, David.
It actually has two sides tothat coin.
I think we why were doctorsbeing purchased by practices?
Because they themselves, as asmall regional practice, was not
able to keep up with all thetechnology and all the changes
that were happening.
Um, if they got you know placedon an HMO, they got the volume
(11:21):
of patients, but that also meansthey have to take care of
billing.
And think about about a thirdof their entire staff is is
there just to collect thereceivables.
When they join an ASO model,they join a large practice, some
of those efficiencies come intoplay.
Today, physicians are migratingand reverting back to direct
primary care, which allows themto move in a cash-paying model.
(11:42):
Um, they're actually topsurgeons like the Oklahoma
Surgical Center.
They're doing bundled surgicalprocedures on a direct cash,
very transparent.
So I'm seeing the ability forthat doctor or that provider
that still wants to have thatdirect patient interaction and
practice and be in control ofhis practice, where technology
is freeing them up.
(12:03):
It's just the same idea thatyou know you can now run your
own practice and do somethingfor any type of business with
technology as opposed to havingyou leverage people and power
and position and power.
So nowadays I think it's goingto have a positive impact.
We may see more and moreindividual physicians practicing
(12:24):
because they don't have to moveinto those large practices to
leverage technology or data orinformation.
David Saltzman (12:32):
And now a word from our sponsor.
If you're like most advisors,you've used platforms that
promised to help, but only addedwork.
Outdated legacy systems,weeks-long setups, endless data
formatting headaches.
BenePower changes all that.
Built by Benefits Pros orBenefits Pros, it was designed
(12:54):
for AI from day one, soeverything works faster and
smarter.
They integrate vendors in days,not weeks, and take any data
format, Excel, PDF, whatever,and make it usable instantly.
VennaPower's Truvo app givesemployees real-time AI-powered
help by chat, email, or voice,and pre-authorization alerts
(13:16):
prevent costly surprises.
You get performancemeasurements and one-click
reporting, so you spend lesstime on admin and more time
growing your book.
Move from frustration tofreedom, from vendor to trusted
advisor with Benepower.
For more information or toschedule a demo of Venepower
Advantage, go to BenePower.com.
(13:37):
And now, back to ourconversation.
It's really more of a networkof specialized models.
(13:58):
Can you explain what that meansto the average employer and
benefits advisor?
Julian Lago (14:02):
Yeah, I think fundamentally, if we understand
a basic computer has a harddrive, Dave, and has storage of
information, um AI is theability for an agentic process
for agents to go and do what'scalled RAG, they retrieve,
augment, and bring thatinformation out of the system.
If we're thinking about aninsurance, the the who, what,
(14:26):
where is covered is written intoan insurance policy.
That's our governing document.
And if you think about asking aquestion to just Google, which
is like going to you know thelibrary and opening up
encyclopedia, it's gonna it'sjust gonna search out
information.
What AI is able to do is go inand retrieve specifically
(14:46):
information as it pertains to asubject.
And that agentic process allowsmultiple agents to work
simultaneously.
Let me let me go into a littledeeper.
Um, if I ask a simple question,um, what's my deductible or
what's remaining of mydeductible?
There's several things.
Number one, I've got to go tothe plan document, find out what
the original deductible was.
You have a $2,500 deductible,$5,000 per family.
(15:09):
It's covered at 80-20 fromthat.
We all kind of follow that.
It has a maximum lifetime X.
But we would also need to sendan agent in there to go read how
much of my deductible have Iaccumulated, right?
And we would need to go findout open EOBs that are pending
against that.
Um, so the question, the simplefundamental question, is
perfect for the way AI is built,which is agentic.
(15:32):
Multiple sources of informationget brought back, and then
through a rag system, itretrieves it, augments that, and
responds back to the questionat hand.
And that process is a perfectprocess for the way AI
functions.
So that's why I believe we'reseeing such a conversation and
such um adoption, early adoptionto AI, unlike other
(15:55):
technologies, David, that um mayhave circled around the
insurance space, but weren'tadopted as as briefly.
I think it's very logical forour process to go.
I could talk about a claimjudication process in the
similar way.
It has multiple moving parts.
You have to pre-authorize, youhave to identify and see, and
then you pay the claim andadjudicate the claim and then
communicate it back to themember.
(16:16):
Um, even on pre-cert, themember has to ask a question and
a go cert.
So there's multiple ways for AIto really streamline that
process.
And at the enterprise level, Ithink that's the most exciting.
You know, TPAs, insurancecompanies, and certainly brokers
that are managing larger booksof business.
Um, this is not somethingthey're doing tomorrow.
This is something they didyesterday already and are
(16:37):
actively putting in place today.
So if you haven't done it,you're already almost behind the
eight-ball.
David Saltzman (16:43):
Well, speaking of behind the eight-ball, I
mean, there are there are modelsof AI that folks aren't even
aware of.
There's one that's calledemotion AI.
And it's it's really, to me,it's kind of one of the most
fascinating parts.
Is it true that it can evendetect the tone or mood that
you're in and then ask if you'reokay?
Julian Lago (16:59):
Yeah, I mean, that it's amazing the amount.
Again, remember all thoselittle agents that I kind of
talked about that are out there,they're listening and they're
hearing, and um they're beingthey're programmed to understand
human emotions.
Um, they're programmed tounderstand the severity of your
voice, the tone of your voice.
Um, they're there to understandif you're struggling with a
(17:20):
native language, if you haveSpanish as your more preferred
language, it will knowintuitively to ask if you prefer
to speak in another language.
Um, so those components are allthere.
Um, and a lot happens behindthe scene, David.
This is where all theprogramming is occurring for
these types of agents and andbringing them together.
One that it listens properly towhat is the true question,
(17:42):
where does it have to go seekthe information, and and and
really in an augmentation, howdo we address the communication
at hand with this person?
Um, are we speaking to thatperson, you know, in the right
tone, in the right methodology?
So all that is is part of whatlarge language capability does.
Unlike a chat, right?
A chat, you just plug it in,you type it in, and it's going
(18:03):
to respond.
It's not listening to the toneof your voice.
David Saltzman (18:07):
Is now you you talk about something called
hyper-personalization.
Is that kind of where that'sgoing?
You you mentioned when whenI've heard you talk about it,
you talk about it kind of as theholy grail.
What does that look like whenyou combine natural language
modeling and emotionalintelligence?
Julian Lago (18:21):
Well, you get very, very close to a human
interaction, right?
Um, think about when we wrote aletter to someone and you put a
big exclamation point, oryou're on the phone and you
pause and you let silence bepart of that conversation.
You can interpret two differentmethodology, two different
messages there.
(18:42):
And um, you know, think abouthow powerful dad was when he you
asked him a question, he turnedto you, he looked at you, but
didn't say anything.
Right?
That that echoes.
And AI has that capability tobring that type of
personalization intoconversations, empathy.
You know, um sometimes it'sit's bringing a sense of urgency
(19:02):
to a situation.
A mom who's on the phonedriving home and and and can't
find the ID card and she's onthe way to a doctor's office,
that's an emergency for her.
She's gonna arrive at thatemergency room, she wants her
child seen, and she knowsthey're gonna ask her for the ID
card.
Doesn't seem like a 911 fireemergency, but in that person's
case, so AI should respondadequately.
(19:22):
Ma'am, focus on your driving.
Let me go ahead and look insideyour text.
We'll be emailing you the IDcard.
It'll be there when you arriveat the doctor's office.
Empathy, efficiency,capabilities that are there.
David Saltzman (19:34):
Now, every technology, it doesn't matter
what it is, every and everytechnological leap comes with
caveats.
What do you see as the biggestrisks or roadblocks, whether
it's privacy, integration, etcetera, that organizations
should be preparing for now?
Julian Lago (19:48):
Well, the interesting part is that
legislatively we catch up totechnology.
We don't predict technology.
HIPAA was written, you know,with a certain amount of you
know, ideas of what it wouldplay out.
We're we're crossing into someof these spaces, David, that we
just don't have.
Um, you know, it's almost likea need for self-governing to
(20:11):
some extent.
And thank goodness that um, youknow, we run through processes
and procedures that allow us tomake sure that we have the
integrity of the individualmedical records, all of their
privacy.
Um, so that there's acautionary tale that's there
because it's it's happened inthe past.
Um, we really don't know towhat extent the legislative
(20:35):
component has to adapt or thetechnology has to adapt.
But I know that there's stateslike California.
Um, I was just on a conferenceand they're already looking at
legislating some of the thingsthat AI could basically um
hallucinate or impersonate aphysician and give information
that is required by law to be alicensed practitioner.
(20:57):
Can it be done?
Yes.
Should it be done?
No, it's illegal, right?
You don't want to impersonate adoctor.
So California's already startedto address, so we're gonna
start seeing some of thesefunctions happen throughout AI
and throughout our industry.
Um and we just have to be awareof what those what those
components are.
David Saltzman (21:16):
I know one of the analogies that you talk
about a lot when you talk togroups is about the buggy whip
makers makers who adapted versusthose who disappeared.
What does that adapting looklike for benefits brokers, for
carriers, and for HR teams inthis new AI era?
Julian Lago (21:31):
Dave, I I I love that analogy for so many reasons
because there's, you know,let's play it out very, very
quickly.
We have limited time, but it'simportant that we kind of play
it out.
You know, all of a sudden Iwake up one morning and all my
orders for buggy whips are gone,right?
And my little shop in town,there's another guy across town.
He got the same request thathey, buggy whips are slowing
(21:51):
down, our current inventory isprobably keep up, so sorry to
have you.
One person takes it and hesays, Oh my God, I'm being
displaced, I'm out of business,you know, time to move out of
state, whatever, whatever thelaws are, or whatever the
situation.
The other guy looks at it, hmm,you know what?
Um, what have I been doing forseveral years?
I've been manipulating leatherand building buggy whips.
And I was visiting the shop umand they've got leather steering
(22:12):
wheels that are there, and Isee the seats are going to be
made out of cloth, but theywould look better in leather.
And uh, let me put in a bid andsee if I can become the new
leather, you know, the upgraderfor these Model T Fords.
Somewhere along the line thathappened, and someone adopted,
and guess what?
That buggy whip manufacturerhad multiple locations, is now
(22:34):
doing for all the vehicles thatare coming off the assembly
line.
Their job quadrupled uh 10x andthey became extremely
successful in that environment.
Did they change fundamentallywhat they were doing?
To some extent, but they werestill in the box of they were
manipulating leather.
Um the other guy said, Oh,completely out of business.
So a lot has to do with how youperceive this.
And if we're confident in whatwe deliver and we're confident
(22:57):
in our capabilities, one of thethings we should never have is
the ability to stop learning andadapting.
And if that's missing in anentrepreneur, you're missing one
of the key components of beingan entrepreneur.
So in my mind, that's anexciting dynamic.
You look at it with how do weadapt?
How do we bring?
And you're not going to haveall the answers.
David Saltzman (23:15):
Well, you can't have all the answers because all
the technology isn't all rolledout yet.
And even the new technology isbegatting, to use a biblical
term, even newer technology.
And so that will continue toevolve.
But, you know, I think a lot ofbrokers are afraid of these
changes.
But boy, there you'd behard-pressed to find an uh an
occupation that has done moreadapting in the last 20 years
(23:38):
than brokers.
I mean, you and I have been atthis a long time.
I remember when we hadindemnity policies where there
was a schedule of what would bepaid for each procedure, period,
end of story, that was it.
And we've gone through changesafter changes after changes.
So I think most brokers will bein a good place.
They just need somebody to kindof help explain it to them and
a partner to work with them.
That's kind of what you guysare doing, isn't it?
Julian Lago (24:00):
Yeah, I mean, that's exactly it, David.
We haven't lost our corecompetency.
We understand how important therole of a consultant and broker
is.
Um, and they're there as aconsultant for these employers.
Interpreting how technology isgoing to make that, you know,
providing healthcare for theiremployees more efficiently and
giving better results is is justpart of this.
(24:20):
There's there's a tremendousnumber of solutions that are
coming into our marketplace.
We could break it down, the olddoctor-hospital medicine, but
in pharmaceutical, there's atremendous amount of new tools
that can function whether you'rea diabetic or not.
Introducing technology so wecan communicate that to Joe the
forklift operator in thewarehouse and making sure that
he can interpret it and use it,that's a key component of the
(24:42):
broker's role.
And they do that really, reallywell.
So we're providing all thetools and all the technology so
they don't have to be the onebuilding it, but they will be
empowered to say, we've got itin our, we've got it, you know,
in our in our uh, you know, ourlist of products and services,
and we can introduce technologythere.
David Saltzman (24:58):
So that's a great place to end our
conversation for today.
Julian Lago, co-founder and CEOat Benepower.
Julian, thanks for afascinating conversation.
Julian Lago (25:07):
Always, always great to catch up, David.
Thank you.
Announcer (25:11):
The Shift Shapers Podcast is a production of Shift
Shaper strategies and may notbe reproduced or quoted in whole
or in part without our expresswritten permission.
Copyright 2020, all rightsreserved.
If the leap from horses to automobiles reshaped
an entire century, what will AIdo to our industry?
We'll find out on this episodeof Shift Shapers.
Announcer (00:13):
Change either energizes or paralyzes.
The choice is yours.
This is the Shift ShapersPodcast, bringing the employee
benefits industry interviewswith individuals and companies
who are shaping the industryshifts.
And now, here's your host,David Saltzman.
David Saltzman (00:33):
And to help us answer that question, we've
invited my old friend JulianLago, who is co-founder and CEO
at Benepower.
Hey Julian, how are you doingtoday?
Julian Lago (00:42):
Hey, David, great to be with you today.
David Saltzman (00:44):
My pleasure, my pleasure.
So you often use that very samehorse to automobile analogy.
Why is that moment in historythe best parallel for what's
happening with AI today?
Julian Lago (00:55):
You know, David, it it's sometimes it's visual.
I remember seeing uh an imageyears back of um, I think it was
Avenue of the Americas in NewYork, and there were two
vehicles and several horse andcarriages there, and then they
go less than five years forward,and it's the reverse.
There's basically one to twohorse and carriages, and they're
(01:18):
all automobiles trying tofigure out how to maneuver.
And it had always struck methat in such a short window,
technology, which was back inthe day, in fact, it wasn't the
auto industry, right?
It was the horseless carriageindustry somehow evolved and how
many people got displaced andso many things had happened.
So that that's been one ofthose things that are discussed
at a lot of business schools,but the image of that was always
(01:40):
uh interesting, that it wasjust a five-year window, and so
dramatically the changeoccurred.
David Saltzman (01:46):
You know, I when you talk about change though, I
mean, AI is not really realreplacing humans, but it's kind
of amplifying them, isn't it?
Julian Lago (01:55):
Yeah, it's an interesting dynamic.
I talked to my my engineers,and we're all using AI, right?
It's the it's the hot topic.
But machine learning, um, I'vegot an engineer on our team that
was with a pretty sizableorganization for seven, eight
years, and he laughs.
He says, Yeah, everybody'stalking AI like it's it's a new
candy, new wrapper, it's a brandnew name.
But it's been around, machinelearning has been around.
(02:16):
I think we've been exploringand seeing it in movies and how
it's going to take over, youknow, and be careful for that
robot that takes over theuniverse.
The reality is that almostwithout exception, through the
history of human mankind,technology makes um uh presence,
whether it's the wheel, whetherit's the you know, automobile
industry, whether it'selectricity turning on, in this
case, AI, the next progressionof the digital experience.
(02:40):
And human in the loop is notsomething that's gonna go away.
You need those components.
So will there be reshifting?
Will people be doing differenttypes of jobs?
Will some of the day-to-dayfunctions or responsibilities
they have go away?
But that means it frees them upto do other things.
And and that's the excitingpart.
You know, we basically uh donew things or have access to new
(03:02):
information.
David Saltzman (03:03):
So let's let's focus this in on our our our
industry and and daily,day-to-day benefits work.
What would a true human AIpartnership kind of look like?
Julian Lago (03:14):
Well, it it you know, if you think about the
process of this, you know, we'rewe're in the sales business in
some ways, where we're bringinga product to the marketplace, a
better solution.
There's shopping, there'squoting.
Um I almost a best way to lookat what the future is going to
be is to look backwards, right?
There's been uh you shared withme a great book one time.
(03:34):
We talked about the thegentleman turning on the
lanterns, right?
And um in the streets.
Think about what it was toorder a simple proposal.
In the back, I used to rememberto fill out a proposal request
form, you submit it, and itwould take a day or two for that
to just be back in your inboxto go present it to a client.
So what was the expectation?
You met with the client, youdid a fact finder, you found out
(03:56):
what their issues were, and youlet them know three to four
days, I will have some proposalsback.
So let's schedule something fornext week.
That process is happening inminutes today, right?
Because we can go in and put ina request for proposal, the
system might digest it, itidentifies it, and literally
within minutes you could bedoing that real time.
So that movement allowed us todo more.
(04:17):
We can we can do multiplequotes within a day, whereas we
were floating and we're movingphysically to go deliver that
presentation, etc.
So our efficiencies increase.
It's basically a scalingcapability.
We're able to scale ourbusiness faster and more
efficiently.
Move that into open enrollment,move that into customer
service, which AI plays asignificant role in.
(04:38):
Move that to the educationcomponent.
We're dealing with thecomplexity.
The complexity of healthcarehas not gone away because
technology is here.
So there's still a complexityof it.
It's not a subject that we dealwith every single day.
When we get ill, it's it's it'sfirst of all, it's important.
It's something we need toresolve, and many times it's
touching our family or ourlifestyle.
So it's a very escalated levelof importance.
(05:01):
And you want to bring all thesolutions to the table.
So, in that, um, that is wheresome of the technology, we're
going to want to maximize it andbe as efficient as we can.
So, efficiency and thosecomponents play a big role.
And I think every aspect ofthat, all the way through the
renewal cycle and ongoing, butmostly that customer service
(05:22):
experience and that education, Ithink AI is going to be a huge
differentiator.
David Saltzman (05:27):
Well, I mean, we're not we're not only seeing
it in those areas, we're alsoseeing it in things like
compliance monitoring, um,broadly enrollment support.
So if it's already reshapingcore processes, like which of
these areas do you think will besee the biggest transformation
with AI in the next 12 to 24months and why?
Julian Lago (05:45):
Well, I think if you think about the ability to
take information and turn itinto a digital component, um,
that bit of information um atour fingertips becomes
invaluable.
For the most part, other thanwith our pains and aches and our
ID card, we didn't walk intothe doctor's office with much
(06:05):
more information.
Today we're empowered.
We literally have all of ourdoctors, all of our networks.
We can do a lot of duediligence early on.
I hate to use the terminology,but you Google your uh your
medical procedure, at least youhave some idea what's gonna
happen at that doctor's office.
We can walk in much moreempowered.
That wasn't the case yearsback.
At best, you talk to a familyfriend or someone of trust and
(06:28):
you say, What's this experiencegonna be like?
And they gave you a vagueindication of that.
We're much more empowered.
Data is also allowing us totake our previous illnesses, our
previous comorbility issues.
If you have borderlinediabetes, that impacts what
you're being treated for, whatmedications you're on.
So information, in essence, isgonna be significantly easier to
(06:51):
come by.
Um, where to go, how to seekthe top physicians, how to get
the best level of care, mostimportantly at the lowest price
or most cost-effective price.
David Saltzman (07:00):
Well, we've talked about advisors a little
bit and employees a little bit.
What's this look like at theemployer side of the equation?
If if you're an advisor andyou're you're working with a a
C-suite person or an HR person,what does this look like?
How is AI helping them?
Julian Lago (07:16):
Uh I think it's the ultimate game changer.
The average consultant we workwith is looking to differentiate
themselves at that employerlevel.
Employers are looking at this.
Um, over the past year, we'veall heard that we're moving the
conversation from the HR to theC-suite, right?
Where it's a financialdecision, there's much more
(07:37):
information, and there's toolsto do predictive analytics, and
that's wonderful.
But what if we can actuallytake a deeper dive into how
their health plan and the waythat it's been built is actually
equaling or serving thepopulation that it's intended to
serve.
Just because you build it thisway, I can build a car with a
convertible, but you know, if ifif the person riding it wears a
(08:01):
toupee, they're not going toopen the convertible top, right?
Because the hair's gonna blowoff.
I'm using the crazy example.
But if if we're building plansin such a way that it doesn't
accompany the population that istrying to serve, it's not a
great solution.
If you're putting in solutionsthat people don't know about and
aren't aware how to use it,virtual health care, people
(08:22):
still perceive that astelemedicine.
And it's so evolved over justthe last several years.
Think about what COVID did withallowing physicians to practice
medicine digitally.
That's a transformation bothfrom their billing to their
methodology and the way they cancommunicate and all the HIPAA
compliance that goes along withit.
We have in some ways thoughthat's an outcome of what we've
(08:46):
experienced through COVID.
So today it's no longertelemedicine, it's virtual care.
That can extend to primarycare, even specialists,
certainly counseling andtreatments, is you know, medical
devices, home care, all of thatcan be touched now with the
digital component.
David Saltzman (09:03):
Do you think that all that stuff that kind of
got quote unquote forced on usduring COVID has opened the door
for a more ready acceptance ofsome of the things that are
coming down the pike?
Julian Lago (09:15):
Yeah, I think in some ways, unfortunately, um
we're creatures of habit.
If it's not broken, we're notgonna go out of our way to fix
it.
If you talk to the averageperson pre-COVID, if they had a
chance to speak a doctor face toface as it get to on the phone
and a Zoom, they'd probablyprefer it.
Um, think of the days where weall had expense accounts, right?
And everybody's, if they'regonna have a meeting, let's go
(09:37):
make it a meeting, I'll fly intotown, take you to dinner, and
so on and so forth.
Today, that becomes a 30-minuteZoom call.
We're here doing a podcast, andI'm not sitting in front of
you, we're using technology todo this.
Um, much more efficient.
Um, in fact, I can do threeother things throughout the day.
Whereas if I was traveling forthat one-on-one meeting,
physicians are in that medicalpractice is in the same way.
(09:59):
Think about what we complainabout having very limited time
with our physicians.
Physicians have to see morepatients to receive similar
reimbursements, and their timeis limited.
In this case, here we can bemuch more efficient because a
lot of the data is there.
Doctors are no longer worriedabout making sure they jot down
all the notes.
They sit there, turn on theirphone, and the AI agent hears
(10:22):
the entire conversation and isready to create and fill in the
ERM and also plan for your thatx-ray, read it, reduce the
medication, you know, lay outyour surgery, and if it's done
properly, all that comes in andbecomes part of your day-to-day
calendar and scheduling withouta lot of thought on either part.
That's where we're headed withsome of this great technology.
David Saltzman (10:41):
So, you know, a lot of the physician
relationships, vastly themajority of physician
relationships now are corporateowned.
They're owned by hospitalsystems and such.
Do you see an accelerator beingthe hospital systems themselves
bringing more AI to bear sothat their doctors can see more
patients in a more in a moretime-effective way?
Julian Lago (11:00):
That's a really interesting question, David.
It actually has two sides tothat coin.
I think we why were doctorsbeing purchased by practices?
Because they themselves, as asmall regional practice, was not
able to keep up with all thetechnology and all the changes
that were happening.
Um, if they got you know placedon an HMO, they got the volume
(11:21):
of patients, but that also meansthey have to take care of
billing.
And think about about a thirdof their entire staff is is
there just to collect thereceivables.
When they join an ASO model,they join a large practice, some
of those efficiencies come intoplay.
Today, physicians are migratingand reverting back to direct
primary care, which allows themto move in a cash-paying model.
(11:42):
Um, they're actually topsurgeons like the Oklahoma
Surgical Center.
They're doing bundled surgicalprocedures on a direct cash,
very transparent.
So I'm seeing the ability forthat doctor or that provider
that still wants to have thatdirect patient interaction and
practice and be in control ofhis practice, where technology
is freeing them up.
(12:03):
It's just the same idea thatyou know you can now run your
own practice and do somethingfor any type of business with
technology as opposed to havingyou leverage people and power
and position and power.
So nowadays I think it's goingto have a positive impact.
We may see more and moreindividual physicians practicing
(12:24):
because they don't have to moveinto those large practices to
leverage technology or data orinformation.
David Saltzman (12:32):
And now a word from our sponsor.
If you're like most advisors,you've used platforms that
promised to help, but only addedwork.
Outdated legacy systems,weeks-long setups, endless data
formatting headaches.
BenePower changes all that.
Built by Benefits Pros orBenefits Pros, it was designed
(12:54):
for AI from day one, soeverything works faster and
smarter.
They integrate vendors in days,not weeks, and take any data
format, Excel, PDF, whatever,and make it usable instantly.
VennaPower's Truvo app givesemployees real-time AI-powered
help by chat, email, or voice,and pre-authorization alerts
(13:16):
prevent costly surprises.
You get performancemeasurements and one-click
reporting, so you spend lesstime on admin and more time
growing your book.
Move from frustration tofreedom, from vendor to trusted
advisor with Benepower.
For more information or toschedule a demo of Venepower
Advantage, go to BenePower.com.
(13:37):
And now, back to ourconversation.
It's really more of a networkof specialized models.
(13:58):
Can you explain what that meansto the average employer and
benefits advisor?
Julian Lago (14:02):
Yeah, I think fundamentally, if we understand
a basic computer has a harddrive, Dave, and has storage of
information, um AI is theability for an agentic process
for agents to go and do what'scalled RAG, they retrieve,
augment, and bring thatinformation out of the system.
If we're thinking about aninsurance, the the who, what,
(14:26):
where is covered is written intoan insurance policy.
That's our governing document.
And if you think about asking aquestion to just Google, which
is like going to you know thelibrary and opening up
encyclopedia, it's gonna it'sjust gonna search out
information.
What AI is able to do is go inand retrieve specifically
(14:46):
information as it pertains to asubject.
And that agentic process allowsmultiple agents to work
simultaneously.
Let me let me go into a littledeeper.
Um, if I ask a simple question,um, what's my deductible or
what's remaining of mydeductible?
There's several things.
Number one, I've got to go tothe plan document, find out what
the original deductible was.
You have a $2,500 deductible,$5,000 per family.
(15:09):
It's covered at 80-20 fromthat.
We all kind of follow that.
It has a maximum lifetime X.
But we would also need to sendan agent in there to go read how
much of my deductible have Iaccumulated, right?
And we would need to go findout open EOBs that are pending
against that.
Um, so the question, the simplefundamental question, is
perfect for the way AI is built,which is agentic.
(15:32):
Multiple sources of informationget brought back, and then
through a rag system, itretrieves it, augments that, and
responds back to the questionat hand.
And that process is a perfectprocess for the way AI
functions.
So that's why I believe we'reseeing such a conversation and
such um adoption, early adoptionto AI, unlike other
(15:55):
technologies, David, that um mayhave circled around the
insurance space, but weren'tadopted as as briefly.
I think it's very logical forour process to go.
I could talk about a claimjudication process in the
similar way.
It has multiple moving parts.
You have to pre-authorize, youhave to identify and see, and
then you pay the claim andadjudicate the claim and then
communicate it back to themember.
(16:16):
Um, even on pre-cert, themember has to ask a question and
a go cert.
So there's multiple ways for AIto really streamline that
process.
And at the enterprise level, Ithink that's the most exciting.
You know, TPAs, insurancecompanies, and certainly brokers
that are managing larger booksof business.
Um, this is not somethingthey're doing tomorrow.
This is something they didyesterday already and are
(16:37):
actively putting in place today.
So if you haven't done it,you're already almost behind the
eight-ball.
David Saltzman (16:43):
Well, speaking of behind the eight-ball, I
mean, there are there are modelsof AI that folks aren't even
aware of.
There's one that's calledemotion AI.
And it's it's really, to me,it's kind of one of the most
fascinating parts.
Is it true that it can evendetect the tone or mood that
you're in and then ask if you'reokay?
Julian Lago (16:59):
Yeah, I mean, that it's amazing the amount.
Again, remember all thoselittle agents that I kind of
talked about that are out there,they're listening and they're
hearing, and um they're beingthey're programmed to understand
human emotions.
Um, they're programmed tounderstand the severity of your
voice, the tone of your voice.
Um, they're there to understandif you're struggling with a
(17:20):
native language, if you haveSpanish as your more preferred
language, it will knowintuitively to ask if you prefer
to speak in another language.
Um, so those components are allthere.
Um, and a lot happens behindthe scene, David.
This is where all theprogramming is occurring for
these types of agents and andbringing them together.
One that it listens properly towhat is the true question,
(17:42):
where does it have to go seekthe information, and and and
really in an augmentation, howdo we address the communication
at hand with this person?
Um, are we speaking to thatperson, you know, in the right
tone, in the right methodology?
So all that is is part of whatlarge language capability does.
Unlike a chat, right?
A chat, you just plug it in,you type it in, and it's going
(18:03):
to respond.
It's not listening to the toneof your voice.
David Saltzman (18:07):
Is now you you talk about something called
hyper-personalization.
Is that kind of where that'sgoing?
You you mentioned when whenI've heard you talk about it,
you talk about it kind of as theholy grail.
What does that look like whenyou combine natural language
modeling and emotionalintelligence?
Julian Lago (18:21):
Well, you get very, very close to a human
interaction, right?
Um, think about when we wrote aletter to someone and you put a
big exclamation point, oryou're on the phone and you
pause and you let silence bepart of that conversation.
You can interpret two differentmethodology, two different
messages there.
(18:42):
And um, you know, think abouthow powerful dad was when he you
asked him a question, he turnedto you, he looked at you, but
didn't say anything.
Right?
That that echoes.
And AI has that capability tobring that type of
personalization intoconversations, empathy.
You know, um sometimes it'sit's bringing a sense of urgency
(19:02):
to a situation.
A mom who's on the phonedriving home and and and can't
find the ID card and she's onthe way to a doctor's office,
that's an emergency for her.
She's gonna arrive at thatemergency room, she wants her
child seen, and she knowsthey're gonna ask her for the ID
card.
Doesn't seem like a 911 fireemergency, but in that person's
case, so AI should respondadequately.
(19:22):
Ma'am, focus on your driving.
Let me go ahead and look insideyour text.
We'll be emailing you the IDcard.
It'll be there when you arriveat the doctor's office.
Empathy, efficiency,capabilities that are there.
David Saltzman (19:34):
Now, every technology, it doesn't matter
what it is, every and everytechnological leap comes with
caveats.
What do you see as the biggestrisks or roadblocks, whether
it's privacy, integration, etcetera, that organizations
should be preparing for now?
Julian Lago (19:48):
Well, the interesting part is that
legislatively we catch up totechnology.
We don't predict technology.
HIPAA was written, you know,with a certain amount of you
know, ideas of what it wouldplay out.
We're we're crossing into someof these spaces, David, that we
just don't have.
Um, you know, it's almost likea need for self-governing to
(20:11):
some extent.
And thank goodness that um, youknow, we run through processes
and procedures that allow us tomake sure that we have the
integrity of the individualmedical records, all of their
privacy.
Um, so that there's acautionary tale that's there
because it's it's happened inthe past.
Um, we really don't know towhat extent the legislative
(20:35):
component has to adapt or thetechnology has to adapt.
But I know that there's stateslike California.
Um, I was just on a conferenceand they're already looking at
legislating some of the thingsthat AI could basically um
hallucinate or impersonate aphysician and give information
that is required by law to be alicensed practitioner.
(20:57):
Can it be done?
Yes.
Should it be done?
No, it's illegal, right?
You don't want to impersonate adoctor.
So California's already startedto address, so we're gonna
start seeing some of thesefunctions happen throughout AI
and throughout our industry.
Um and we just have to be awareof what those what those
components are.
David Saltzman (21:16):
I know one of the analogies that you talk
about a lot when you talk togroups is about the buggy whip
makers makers who adapted versusthose who disappeared.
What does that adapting looklike for benefits brokers, for
carriers, and for HR teams inthis new AI era?
Julian Lago (21:31):
Dave, I I I love that analogy for so many reasons
because there's, you know,let's play it out very, very
quickly.
We have limited time, but it'simportant that we kind of play
it out.
You know, all of a sudden Iwake up one morning and all my
orders for buggy whips are gone,right?
And my little shop in town,there's another guy across town.
He got the same request thathey, buggy whips are slowing
(21:51):
down, our current inventory isprobably keep up, so sorry to
have you.
One person takes it and hesays, Oh my God, I'm being
displaced, I'm out of business,you know, time to move out of
state, whatever, whatever thelaws are, or whatever the
situation.
The other guy looks at it, hmm,you know what?
Um, what have I been doing forseveral years?
I've been manipulating leatherand building buggy whips.
And I was visiting the shop umand they've got leather steering
(22:12):
wheels that are there, and Isee the seats are going to be
made out of cloth, but theywould look better in leather.
And uh, let me put in a bid andsee if I can become the new
leather, you know, the upgraderfor these Model T Fords.
Somewhere along the line thathappened, and someone adopted,
and guess what?
That buggy whip manufacturerhad multiple locations, is now
(22:34):
doing for all the vehicles thatare coming off the assembly
line.
Their job quadrupled uh 10x andthey became extremely
successful in that environment.
Did they change fundamentallywhat they were doing?
To some extent, but they werestill in the box of they were
manipulating leather.
Um the other guy said, Oh,completely out of business.
So a lot has to do with how youperceive this.
And if we're confident in whatwe deliver and we're confident
(22:57):
in our capabilities, one of thethings we should never have is
the ability to stop learning andadapting.
And if that's missing in anentrepreneur, you're missing one
of the key components of beingan entrepreneur.
So in my mind, that's anexciting dynamic.
You look at it with how do weadapt?
How do we bring?
And you're not going to haveall the answers.
David Saltzman (23:15):
Well, you can't have all the answers because all
the technology isn't all rolledout yet.
And even the new technology isbegatting, to use a biblical
term, even newer technology.
And so that will continue toevolve.
But, you know, I think a lot ofbrokers are afraid of these
changes.
But boy, there you'd behard-pressed to find an uh an
occupation that has done moreadapting in the last 20 years
(23:38):
than brokers.
I mean, you and I have been atthis a long time.
I remember when we hadindemnity policies where there
was a schedule of what would bepaid for each procedure, period,
end of story, that was it.
And we've gone through changesafter changes after changes.
So I think most brokers will bein a good place.
They just need somebody to kindof help explain it to them and
a partner to work with them.
That's kind of what you guysare doing, isn't it?
Julian Lago (24:00):
Yeah, I mean, that's exactly it, David.
We haven't lost our corecompetency.
We understand how important therole of a consultant and broker
is.
Um, and they're there as aconsultant for these employers.
Interpreting how technology isgoing to make that, you know,
providing healthcare for theiremployees more efficiently and
giving better results is is justpart of this.
(24:20):
There's there's a tremendousnumber of solutions that are
coming into our marketplace.
We could break it down, the olddoctor-hospital medicine, but
in pharmaceutical, there's atremendous amount of new tools
that can function whether you'rea diabetic or not.
Introducing technology so wecan communicate that to Joe the
forklift operator in thewarehouse and making sure that
he can interpret it and use it,that's a key component of the
(24:42):
broker's role.
And they do that really, reallywell.
So we're providing all thetools and all the technology so
they don't have to be the onebuilding it, but they will be
empowered to say, we've got itin our, we've got it, you know,
in our in our uh, you know, ourlist of products and services,
and we can introduce technologythere.
David Saltzman (24:58):
So that's a great place to end our
conversation for today.
Julian Lago, co-founder and CEOat Benepower.
Julian, thanks for afascinating conversation.
Julian Lago (25:07):
Always, always great to catch up, David.
Thank you.
Announcer (25:11):
The Shift Shapers Podcast is a production of Shift
Shaper strategies and may notbe reproduced or quoted in whole
or in part without our expresswritten permission.
Copyright 2020, all rightsreserved.
Popular Podcasts
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
Are You A Charlotte?
In 1997, actress Kristin Davis’ life was forever changed when she took on the role of Charlotte York in Sex and the City. As we watched Carrie, Samantha, Miranda and Charlotte navigate relationships in NYC, the show helped push once unacceptable conversation topics out of the shadows and altered the narrative around women and sex. We all saw ourselves in them as they searched for fulfillment in life, sex and friendships. Now, Kristin Davis wants to connect with you, the fans, and share untold stories and all the behind the scenes. Together, with Kristin and special guests, what will begin with Sex and the City will evolve into talks about themes that are still so relevant today. "Are you a Charlotte?" is much more than just rewatching this beloved show, it brings the past and the present together as we talk with heart, humor and of course some optimism.
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.