Geoff Moore is the Chief Information Officer at Valmark Financial Group, a leader in the financial services industry since 1963, specializing in life insurance, equity, investment advisory, risk management, and retirement plan products for top financial advisors. With over 20 years of experience in financial services and technology, Geoff works to enhance operational efficiency and streamline processes at Valmark. He leads initiatives across cybersecurity, expansion through recruitment, and operational ease, utilizing a team of skilled developers, testers, project managers, and business analysts. Passionate about staying abreast of the latest technological tools and methodologies, Geoff drives Valmark’s continuous improvement and innovation.
In this episode…
Artificial intelligence (AI) is reshaping industries across the globe, but what about its impact on the wealth management sector? Amidst the buzz and growing applications, can AI truly revolutionize how wealth firms operate, or is it just another fleeting trend?
According to Geoff Moore, a seasoned technologist with over two decades in the field, AI holds significant promise despite its surrounding hype. He compares the current excitement about AI to the early days of the World Wide Web, noting both periods were marked by groundbreaking potential and inevitable challenges. Geoff highlights AI's practical benefits in enhancing operations and compliance within wealth management, yet he remains cautiously optimistic. He emphasizes the realistic improvements AI can provide, like modest boosts in developer productivity and strategic enhancements in regulatory compliance, suggesting that while AI may not revolutionize the industry overnight, it is paving the way for substantial progress.
In this episode of The Beacon Flash Podcast, host Chip Krispert and Geoff Moore, Chief Information Officer at Valmark Financial Group, explore the intriguing world of artificial intelligence in wealth management. They discuss AI’s hype, its realistic benefits and potential pitfalls, how firms are strategically deploying AI, and the critical role of regulatory and security frameworks in its adoption.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:02):
Welcome to the podcast that is dedicated to
providing insight, intelligenceand thought-provoking
conversations with members ofthe enterprise wealth management
community.
Speaker 2 (00:17):
Chip Kispert.
Here I'm the host of the BeaconFlash podcast, where I discuss
the current state as well as thefuture of enterprise wealth
management with its leaders.
Past guests have includedSalesforce's Michelle Feinstein,
sid Yenimendra, the CEO ofSurge Ventures, and Frank
McAleer from Raymond James.
(00:37):
This podcast is brought to youby Beacon Strategies LLC, a
go-to resource for enterprisewealth management firms and
providers.
We are a consulting firm thatis dedicated to the wealth
management industry.
We help our customersunderstand, adapt and be more
relevant in today's dynamicenterprise wealth business.
(00:58):
We bring three services to bearwe host focus roundtables,
deliver consulting services andprovide gap products.
In short, we help wealth andprovider firms foster business
growth and improve profitability.
Recently, we helped one of ourcustomers choose a new clearing
firm and investment platform.
(01:19):
Add to those bodies of work.
We aided them as they resettheir cash management economics.
Our work helped them addmillions to their margin.
If you need some industryperspective or help, visit
BeaconStrategiesLLCcom and don'tmiss out on our exclusive
podcast, the Beacon Flash forindustry news and trends.
(01:41):
Today I'm excited to have JeffMoore on with me.
He is the Chief InformationOfficer of Valmark Financial
Group.
He brings an unbelievableamount of experience to the
table as well.
He's a friend.
I've known Jeff for over 15years.
Notable recent highlights inJeff's career include holding
positions on the FSI Operationsand Technology Council.
(02:05):
Additionally, in 2020, theValmark MyCase team was a
runner-up for Tech Team of theYear from the Greater Cleveland
Partnership and finally, in 2022, he was selected to the Surge
Ventures Industry Council forhis expertise in cybersecurity
(02:25):
and financial service.
Oh, jeff, I am so excited thatyou're here today.
We get to talk about AI.
Speaker 1 (02:33):
Indeed, I'm excited too, chip.
I know everyone's been kind oftalking about this.
I'm glad we finally get tocatch up and discuss this.
Speaker 2 (02:40):
So, Jeff, we've had a bunch of our roundtables
recently.
We've had a lot of discussionon AI.
We've had a ton of discussionjust on a daily basis.
I'm getting that discussion, soyou know, right off the bat,
tell me a little bit about yourthoughts on AI.
Speaker 1 (03:00):
Well, I mean overall, I'm pretty excited.
As a technology, I think thisis pretty exciting.
It almost reminds me back alittle bit when the World Wide
Web started, you know, or atleast people were starting to
use it, maybe like 20 years ago,kind of more commercially.
There was a lot of hype, butthere was also a lot of stuff
out there that was just wasn'tvery good too.
So I feel like we're I don'tknow if we're quite at the top
(03:22):
of the hype cycle just yet, butI think it is exciting.
You know some of theannouncements, like Microsoft's
co-pilot features, but I alsothink we're still in the process
of sorting out what's going toprovide real value still and
what is just something thatmight be overhyped.
Speaker 2 (03:37):
So I agree with you 100%.
You know, it's fascinating whenwe sit down with wealth firms,
right, they're all excited,they're super excited, but then
you kind of start doing a littlepick and shovel work and saying
, all right, where, where areyou going to use it?
Right, and how are you going touse it?
Um, and and that that draws apause.
(04:00):
There's a pause there.
So I think we are you know,we're not even we are really
young into this AI process.
Do you see a lot of firms?
Because you know this is one ofthe things that I noticed about
you and maybe this is thebetter question I know you
created an AI committee.
Can you tell me?
Speaker 1 (04:21):
about that a little bit.
Yeah, so, interestingly enough,I mean I wish I could take
credit for the concept.
It was actually a youngerperson in the organization that
was in charge of a business lineand he he was.
He was reading all kinds ofinformation about AI and he said
, jeff, I, I really I want tostart a committee, I want to get
some other people involved.
And I said, great, as long asyou're the one that's willing to
(04:42):
chair it.
And he said, for sure thing.
So it's actually worked outreally well, because I think one
of the other things with someof this innovation and AI, I
don't think it's going tonecessarily come from, you know,
somebody at the top of thehouse.
I think it really is going totake some innovation from all
different areas of theorganization, right, somebody
that you know knows their jobreally well and can then maybe
(05:02):
apply some of this technology orthink about it and how they can
apply it to their job and thenalso teach their peers how to
use this.
I mean, one example is just evenwith what they call prompt
engineering, or you can think ofthis as just you know how to
put what type of terms do youput in the box and what kind of
results are you're getting.
That's a skill Like people arestill learning how to do that,
and some people are better atthat than others, so I think
(05:25):
it's good to have committeesthat can have just a broad sense
of what their jobs are like.
And then also as part of thiscommittee, I think it was really
important that we includedlegal and compliance.
I'm sure we'll talk a littlebit more about that later, but
there's all kinds of questionsthat are starting to come up in
terms of sort of the legal andcompliance components of AI as
well.
I guess, yeah, we'll just leaveit there for now.
Speaker 2 (05:53):
You know it's fascinating because, you know,
having had, you know, 20enterprise wealth firms and
seven providers last week inAustin, it provided me an
opportunity to really thinkabout, for about a day, what we
see out there.
As I said when we said a littlebit earlier, people are super
excited, but we also we look atit almost in three buckets.
We look at bucket one being usecases, bucket two being
(06:17):
regulatory and bucket threebeing security, and I think a
lot of us see, oh, chat, gpt,and we get excited.
It can write letters for us, itcan do this.
There are also some guardrailsthat we need to be thinking
about.
Speaker 1 (06:32):
Yeah, yeah, yeah, for sure I'm thinking right.
As you mentioned, there's liketwo topic areas kind of popped
off in my head right away, andthe first is around, I'll call
it.
We'll start with the kind ofthe regulatory and then we'll
move into security.
And so, on the regulatory front, there was a great case study
that just came out from AirCanada, and if you haven't heard
(06:52):
about this story, what it was,there was a woman.
Unfortunately she needed toattend a funeral and she needed
to book her ticket very quicklyand, for whatever reason, she
didn't have time to get thebereavement situation sorted out
.
So she asked the chatbot at AirCanada can I just book my
ticket now and apply for thebereavement rate?
Later Chatbot said sure, noproblem, you can do that.
(07:15):
She booked her ticket and thenlater Air Canada came back and
said no, actually that's not ourpolicy, you needed to have that
sorted out ahead of time.
Well, she went ahead and suedAir Canada and won, and the
finding was basically that AirCanada was responsible for the
results of the chatbot.
(07:37):
Now, why that's so interestingnow is we have a legal precedent
saying that these chatbots areresponsible for their answers,
even if that differs from theofficial policy of the
organization.
So that's, I think, could be anexample of where people are
starting to, you know, get somepause.
The other part on the securityfront I think is interesting is
it kind of ties in withregulatory is if people are
wanting to use some of theselarger systems like open AI and
(07:59):
they're going to use their APIsand that sort of thing, there's
a question of like okay, well,does my data go back into the
model?
And they're putting somelanguage around saying like, oh,
it's not coming back in themodel, but do we know and do we
have evidence that that'shappening?
Is there a good audit functionthat we can trust?
That I'm not sure is 100% inplace.
So if somebody is thinkingabout really sensitive uh that
(08:22):
they have that they want to putthrough some sort of system, um,
how do you know for sure, otherthan you know?
They say they're doing it, butare they really?
Speaker 2 (08:29):
we just don't know yet yeah, you know you're spot
on because you know I look atthe security aspect of it scares
a heck of um, honestly, andthat's you know.
I, I believe and I and I thinkyou know, like you guys have
already done, you created yourAI committee right?
I think that's super necessaryfor firms, I think that helps
(08:50):
with strategy, I think it helpswith governance, all those
pieces of the puzzle, and I hopeto see more firms putting this
type of the type of committeethat you have in place.
I think it's great Thanks.
So question there as part ofthat committee, obviously you do
talk about governance, talkabout security.
How are you, you know, how areuse cases coming?
Speaker 1 (09:13):
up.
It's interesting.
So a lot of it's just you know,we'll have like an informal
communication group where, aspeople are reading the news and
coming up with things, they'llsay, hey, did you see this or
did you see that, or I have thisuh thing I want to try or I
think could be useful.
Can we experiment with this andtest it out?
Um, I will say I feel like it'smixed.
(09:35):
I feel like sometimes we we hitsomething and we feel like, oh,
this is really really good.
And then other times we'll trysomething out and it's like, ah,
it didn't quite live up to thehype.
So I just feel I feel verymixed.
Speaker 2 (09:47):
Sometimes it's like man spot on amazing, and other
times it's like oh, this is notwhat I thought it was.
That's funny, because, butwhat's different than anything
else?
We have that in our daily lives, right?
Oh, hey, great idea.
Well, all right that one didn'treally resonate all that well
yeah, yeah, absolutely.
Speaker 1 (10:08):
It's like I'm so excited about microsoft copilot
in so many ways, and then someof the things I've done with it
blow my mind and others.
I'm just like, oh man, I, Ishould have just done this the
old-fashioned way.
Speaker 2 (10:19):
It would have been faster well, you know it's funny
because just little thingsright.
So, looking at, you know, whenwe obviously we do a lot of
video right, we look at teamsand we look at zoom right, and
the crazy thing, you gotMicrosoft co-pilot right, but
yet to be able to get an AIsummary of a meeting doesn't
seem to be in there yet.
Speaker 1 (10:40):
We, we don't record our meetings, so we haven't, and
that's more for regulatorylegal compliance reasons.
So, uh, I wouldn't know, but II do.
I do know those tools are outthere for, like zoom and other
other.
Speaker 2 (10:53):
Yeah, it's just the summary aspect we really like
and everybody I talk to hey, youget a summary of a meeting.
It's great, um so anotherquestion for you, and this one
is more Out of curiosity, and itcomes from this takeaway we had
from last week, which was thatwealth firms are really looking
to their partners, theirprovider partners, to help lead
(11:16):
them in the conversation of AI.
Speaker 1 (11:18):
Do you agree with that?
Yeah, I think so.
I think a lot of these toolsare going to have to be built by
vendors for the majority offirms.
I will tell you, I've spentsome.
I am a bit of a closet nerd andI've been trying to build.
Speaker 2 (11:32):
I never knew that.
Speaker 1 (11:33):
I know, I know I literally I'll check it.
If I move my camera I'll showyou.
I got it.
One of those Nvidia 4090 cards,put it in my uh, my, my desktop
at work to try and build out myown GPT system based on
documents.
We have to kind of see what itcould answer.
And I'll say, even though thereare tons of open source items
(11:54):
available, apis, all that stuffit is still not easy to then
also do it in a secure, scalableway for enterprises.
So I can definitely see, andthat's just one use case, right,
that's just like what they calla RAG system or a retrieval
augmented generation type system.
That's just one type case,right, that's just like what
they call a RAG system or aretrieval augmented generation
type system.
That's just one type of AI andthere's so many others and I
think it's hard, you know.
(12:15):
I'll give you another example.
Like there's a vendor we'vebeen looking at that's
specifically in the lifeinsurance space.
It's been doing some kind ofinteresting things and part of
it it was they had to buildthese models, right, that takes
a long time to just build themodel.
So I can see why vendors, a lotof it's just going to take time
to like train these models towork the right way.
I know another one, rich Walker, that you're friends with.
(12:39):
You know he's got a new productthat they've come out with that
they've trained their model onforms and helping to
automatically pull data valueout of forms and stuff.
So I think that's a goodexample of where you need
partners Like that would bereally hard for a single firm to
do on their own.
Something like that wherethere's a lot of mapping
involved, or something like thatwhere you need a lot of model
training.
Speaker 2 (12:55):
Yeah, no, it's fascinating to me.
I think, you know, one of thethings we also kind of have come
away with over the last fewmonths is there almost needs to
be a primer on all this, right,an overview, you know, an
(13:15):
overview piece where peoplecould really learn about the
different aspects of AI and andyou know the different types of
models.
You know I always love to askthose questions, so you know I,
I obviously I talked to a lot ofpeople in the industry, right,
we're, you know, we're lookingfor insight as well.
I've had the opportunity totalk about, you know, what
Salesforce's strategy is withMichelle Feinstein Really,
really interesting where they'regoing with it.
Being able to spend some timewith our mutual friend, sid
(13:37):
Yedimendra, who with SurgeVentures and they're doing some
neat things with compliance,artificial intelligence who are
some folks out there that you'reseeing, both as a provider and
maybe, additionally, as you know, people that are are solving,
utilizing ai and solvingproblems uh.
Speaker 1 (13:56):
So just a couple people maybe to think about.
Uh, mark butler with his wealthmanagement gpt and I know they
recently they had that in pilotfor quite some time uh, a chat
gpt like tool that has somespecific prompt for advisors to
use.
So I know he just recently wasadded to the Kitsies FinTech map
.
Jonathan Michael he justrenamed his company AdvisorX AI.
(14:19):
He's been putting out a lot ofmaterial on LinkedIn and
specifically around AI.
Two other names I'd mentionedI'm local to Cleveland, so
KeyBank is in our backyard.
I'm not going to say his name,right.
Azeem Shaikh.
He puts out a to Cleveland, soKeyBank is in our backyard.
I'm not going to say his name,right.
Azeem Shaikh he puts out a lotof material from KeyBank, just
kind of what they're doing.
He's kind of on their strategyteam, so he puts out a lot of
(14:40):
really interesting thoughtpieces around AI.
And then, lastly, I just gotintroduced to the Responsible AI
Institute, which I think isalso really interesting.
So, less on the tech front butmore on the sort of regulatory
compliance front.
I think they're starting to putout some really good
information.
And I know Jackson NationalLife is in our space, the
(15:03):
finance space, and they're amember of the Responsible AI
Institute, as a member of FOOT,so they've got some really good
I think just thought piecesabout some of the Responsible AI
Institute as a member, sothey've got some really good.
I think just thought piecesabout some of the more
regulatory things we should bethinking about as we're
evaluating AI systems.
Speaker 2 (15:16):
Got it, got it.
So you know, just big rocksright when you're looking at
what AI could do for your firm.
What are a couple items thatyou're taking away as hey, this
is something we're superinterested in.
And then, on the flip side,what are some things that maybe
you've thought about but haveseen data proving otherwise?
Speaker 1 (15:41):
So I will start with expectations around software
developers, right, so I'll putthis in the overhyped category.
I have a friend who they have apretty large software
development organization, sodeveloper productivity is really
, really important for them.
And they were, you know theirexecutive team was initially
seeing you know marketing slidesthat were come out and saying
(16:02):
things like you know, you'll get40 to 50% developer
productivity.
I mean that's huge becausetypically these are, you know,
high dollar resources.
So what they did is theyactually went out and had
studies done by McKinsey,microsoft and others to see, ok,
if we have, you do, anengagement, what can we actually
expect as a productivity liftfor our developers?
(16:23):
And when they had to put theirreputation on the line, those
numbers came back at like 2% to%.
So I mean not zero.
So there's definitely a rolefor some of these tools, but it
might not be as big as people orsome of these.
You know people are touting butdefinitely, you know, I think
even a two to 4% lift insomebody's you know work output
(16:44):
could be really good and I thinkwhat we've seen is especially
with, like, younger developers,it's very helpful, right, they
can quickly, you know, don'thave to sort through a lot of
different websites or YouTubevideos, they can get the answer
pretty quickly for kind of morebasic things.
As it comes to more likearchitecture, engineering, like
that's a complete differentballgame.
So I think where I'm probablymost excited still a little bit
(17:07):
cautious, but starting to putsome more effort into is, again,
I mentioned a little bitearlier this RAG or retrieval,
augmented generation.
So this is like you can putyour own organization's
documents into a chat GPT likeresponse.
So, for example, let's say youhave, you know, in our world
broker dealer world, we've gotwritten supervisory procedures,
right.
Some of these can be very, verylong.
(17:28):
What if you could put adocument like this inside a
system and people could justsimply ask the question?
You know, and the good thing iswith these systems too, they
often reference the document andeven sometimes the page number
from which it came.
So there's a little bit moretracking involved.
(17:48):
With somebody's asking thesystem it can be verified like
okay, that's, the document camefrom, this was the page number
to go and check it.
So I still think it's kind ofearly to see if, if that would
actually prove out, and some ofit's going to depend on the
quality of the organization's.
You know inputs right Ifthey've got good inputs or if
they've got outdateddocumentation floating around.
Speaker 2 (18:13):
Fascinating, fascinating Closing question.
You know we've we have a greatdeal of rhetoric going on with
AI right now.
As I said earlier, I don't eventhink we've touched the surface
.
We're hearing a lot of wealthfirms are you know?
They want to be the theproviders to be their AI
catalyst?
Completely get it.
You know.
They want to be the providersto be their AI catalyst?
Completely get it.
You know, is AI going to gettraction or is it going to be
(18:35):
relegated to another one ofthose buzzwords?
What do you think?
Speaker 1 (18:38):
It's going to get traction, I'm going to go.
I'm going to go on record assaying it's going to get
traction.
I, I feel so.
I'm just giving you in myperspective, like when we heard
a lot about RPA robotic processautomation I always kind of
thought like, ah, this is aBand-Aid.
Really somebody wants to doAPIs.
I get, maybe you've got somelegacy systems.
You know, that was an example Ithought was kind of overhyped.
(18:59):
I still think there's a ton ofpotential out there with AI to
help really accelerate both ourlearning and just how we do our
day-to-day jobs.
Speaker 2 (19:10):
Fabulous, Jeff.
Thank you for being a guest onthe Beacon Flash podcast.
Always love to have you here.
Speaker 1 (19:16):
All right, thank you, chip.
Speaker 2 (19:18):
We've been talking to Jeff Moore, who is the CIO of
Valmark Financial Group, and ithas been a real pleasure.
If you need some industryperspective or help, visit
BeaconStrategiesLLCcom and don'tmiss out on our exclusive
podcast, the Beacon Flash forindustry news and trends.
Have a great day.
Speaker 1 (19:40):
Thank you for listening to the Beacon Flash
podcast.
We'll see you next time and besure to click subscribe to get
future episodes.
Welcome to the podcast that is dedicated to
providing insight, intelligenceand thought-provoking
conversations with members ofthe enterprise wealth management
community.
Speaker 2 (00:17):
Chip Kispert.
Here I'm the host of the BeaconFlash podcast, where I discuss
the current state as well as thefuture of enterprise wealth
management with its leaders.
Past guests have includedSalesforce's Michelle Feinstein,
sid Yenimendra, the CEO ofSurge Ventures, and Frank
McAleer from Raymond James.
(00:37):
This podcast is brought to youby Beacon Strategies LLC, a
go-to resource for enterprisewealth management firms and
providers.
We are a consulting firm thatis dedicated to the wealth
management industry.
We help our customersunderstand, adapt and be more
relevant in today's dynamicenterprise wealth business.
(00:58):
We bring three services to bearwe host focus roundtables,
deliver consulting services andprovide gap products.
In short, we help wealth andprovider firms foster business
growth and improve profitability.
Recently, we helped one of ourcustomers choose a new clearing
firm and investment platform.
(01:19):
Add to those bodies of work.
We aided them as they resettheir cash management economics.
Our work helped them addmillions to their margin.
If you need some industryperspective or help, visit
BeaconStrategiesLLCcom and don'tmiss out on our exclusive
podcast, the Beacon Flash forindustry news and trends.
(01:41):
Today I'm excited to have JeffMoore on with me.
He is the Chief InformationOfficer of Valmark Financial
Group.
He brings an unbelievableamount of experience to the
table as well.
He's a friend.
I've known Jeff for over 15years.
Notable recent highlights inJeff's career include holding
positions on the FSI Operationsand Technology Council.
(02:05):
Additionally, in 2020, theValmark MyCase team was a
runner-up for Tech Team of theYear from the Greater Cleveland
Partnership and finally, in 2022, he was selected to the Surge
Ventures Industry Council forhis expertise in cybersecurity
(02:25):
and financial service.
Oh, jeff, I am so excited thatyou're here today.
We get to talk about AI.
Speaker 1 (02:33):
Indeed, I'm excited too, chip.
I know everyone's been kind oftalking about this.
I'm glad we finally get tocatch up and discuss this.
Speaker 2 (02:40):
So, Jeff, we've had a bunch of our roundtables
recently.
We've had a lot of discussionon AI.
We've had a ton of discussionjust on a daily basis.
I'm getting that discussion, soyou know, right off the bat,
tell me a little bit about yourthoughts on AI.
Speaker 1 (03:00):
Well, I mean overall, I'm pretty excited.
As a technology, I think thisis pretty exciting.
It almost reminds me back alittle bit when the World Wide
Web started, you know, or atleast people were starting to
use it, maybe like 20 years ago,kind of more commercially.
There was a lot of hype, butthere was also a lot of stuff
out there that was just wasn'tvery good too.
So I feel like we're I don'tknow if we're quite at the top
(03:22):
of the hype cycle just yet, butI think it is exciting.
You know some of theannouncements, like Microsoft's
co-pilot features, but I alsothink we're still in the process
of sorting out what's going toprovide real value still and
what is just something thatmight be overhyped.
Speaker 2 (03:37):
So I agree with you 100%.
You know, it's fascinating whenwe sit down with wealth firms,
right, they're all excited,they're super excited, but then
you kind of start doing a littlepick and shovel work and saying
, all right, where, where areyou going to use it?
Right, and how are you going touse it?
Um, and and that that draws apause.
(04:00):
There's a pause there.
So I think we are you know,we're not even we are really
young into this AI process.
Do you see a lot of firms?
Because you know this is one ofthe things that I noticed about
you and maybe this is thebetter question I know you
created an AI committee.
Can you tell me?
Speaker 1 (04:21):
about that a little bit.
Yeah, so, interestingly enough,I mean I wish I could take
credit for the concept.
It was actually a youngerperson in the organization that
was in charge of a business lineand he he was.
He was reading all kinds ofinformation about AI and he said
, jeff, I, I really I want tostart a committee, I want to get
some other people involved.
And I said, great, as long asyou're the one that's willing to
(04:42):
chair it.
And he said, for sure thing.
So it's actually worked outreally well, because I think one
of the other things with someof this innovation and AI, I
don't think it's going tonecessarily come from, you know,
somebody at the top of thehouse.
I think it really is going totake some innovation from all
different areas of theorganization, right, somebody
that you know knows their jobreally well and can then maybe
(05:02):
apply some of this technology orthink about it and how they can
apply it to their job and thenalso teach their peers how to
use this.
I mean, one example is just evenwith what they call prompt
engineering, or you can think ofthis as just you know how to
put what type of terms do youput in the box and what kind of
results are you're getting.
That's a skill Like people arestill learning how to do that,
and some people are better atthat than others, so I think
(05:25):
it's good to have committeesthat can have just a broad sense
of what their jobs are like.
And then also as part of thiscommittee, I think it was really
important that we includedlegal and compliance.
I'm sure we'll talk a littlebit more about that later, but
there's all kinds of questionsthat are starting to come up in
terms of sort of the legal andcompliance components of AI as
well.
I guess, yeah, we'll just leaveit there for now.
Speaker 2 (05:53):
You know it's fascinating because, you know,
having had, you know, 20enterprise wealth firms and
seven providers last week inAustin, it provided me an
opportunity to really thinkabout, for about a day, what we
see out there.
As I said when we said a littlebit earlier, people are super
excited, but we also we look atit almost in three buckets.
We look at bucket one being usecases, bucket two being
(06:17):
regulatory and bucket threebeing security, and I think a
lot of us see, oh, chat, gpt,and we get excited.
It can write letters for us, itcan do this.
There are also some guardrailsthat we need to be thinking
about.
Speaker 1 (06:32):
Yeah, yeah, yeah, for sure I'm thinking right.
As you mentioned, there's liketwo topic areas kind of popped
off in my head right away, andthe first is around, I'll call
it.
We'll start with the kind ofthe regulatory and then we'll
move into security.
And so, on the regulatory front, there was a great case study
that just came out from AirCanada, and if you haven't heard
(06:52):
about this story, what it was,there was a woman.
Unfortunately she needed toattend a funeral and she needed
to book her ticket very quicklyand, for whatever reason, she
didn't have time to get thebereavement situation sorted out
.
So she asked the chatbot at AirCanada can I just book my
ticket now and apply for thebereavement rate?
Later Chatbot said sure, noproblem, you can do that.
(07:15):
She booked her ticket and thenlater Air Canada came back and
said no, actually that's not ourpolicy, you needed to have that
sorted out ahead of time.
Well, she went ahead and suedAir Canada and won, and the
finding was basically that AirCanada was responsible for the
results of the chatbot.
(07:37):
Now, why that's so interestingnow is we have a legal precedent
saying that these chatbots areresponsible for their answers,
even if that differs from theofficial policy of the
organization.
So that's, I think, could be anexample of where people are
starting to, you know, get somepause.
The other part on the securityfront I think is interesting is
it kind of ties in withregulatory is if people are
wanting to use some of theselarger systems like open AI and
(07:59):
they're going to use their APIsand that sort of thing, there's
a question of like okay, well,does my data go back into the
model?
And they're putting somelanguage around saying like, oh,
it's not coming back in themodel, but do we know and do we
have evidence that that'shappening?
Is there a good audit functionthat we can trust?
That I'm not sure is 100% inplace.
So if somebody is thinkingabout really sensitive uh that
(08:22):
they have that they want to putthrough some sort of system, um,
how do you know for sure, otherthan you know?
They say they're doing it, butare they really?
Speaker 2 (08:29):
we just don't know yet yeah, you know you're spot
on because you know I look atthe security aspect of it scares
a heck of um, honestly, andthat's you know.
I, I believe and I and I thinkyou know, like you guys have
already done, you created yourAI committee right?
I think that's super necessaryfor firms, I think that helps
(08:50):
with strategy, I think it helpswith governance, all those
pieces of the puzzle, and I hopeto see more firms putting this
type of the type of committeethat you have in place.
I think it's great Thanks.
So question there as part ofthat committee, obviously you do
talk about governance, talkabout security.
How are you, you know, how areuse cases coming?
Speaker 1 (09:13):
up.
It's interesting.
So a lot of it's just you know,we'll have like an informal
communication group where, aspeople are reading the news and
coming up with things, they'llsay, hey, did you see this or
did you see that, or I have thisuh thing I want to try or I
think could be useful.
Can we experiment with this andtest it out?
Um, I will say I feel like it'smixed.
(09:35):
I feel like sometimes we we hitsomething and we feel like, oh,
this is really really good.
And then other times we'll trysomething out and it's like, ah,
it didn't quite live up to thehype.
So I just feel I feel verymixed.
Speaker 2 (09:47):
Sometimes it's like man spot on amazing, and other
times it's like oh, this is notwhat I thought it was.
That's funny, because, butwhat's different than anything
else?
We have that in our daily lives, right?
Oh, hey, great idea.
Well, all right that one didn'treally resonate all that well
yeah, yeah, absolutely.
Speaker 1 (10:08):
It's like I'm so excited about microsoft copilot
in so many ways, and then someof the things I've done with it
blow my mind and others.
I'm just like, oh man, I, Ishould have just done this the
old-fashioned way.
Speaker 2 (10:19):
It would have been faster well, you know it's funny
because just little thingsright.
So, looking at, you know, whenwe obviously we do a lot of
video right, we look at teamsand we look at zoom right, and
the crazy thing, you gotMicrosoft co-pilot right, but
yet to be able to get an AIsummary of a meeting doesn't
seem to be in there yet.
Speaker 1 (10:40):
We, we don't record our meetings, so we haven't, and
that's more for regulatorylegal compliance reasons.
So, uh, I wouldn't know, but II do.
I do know those tools are outthere for, like zoom and other
other.
Speaker 2 (10:53):
Yeah, it's just the summary aspect we really like
and everybody I talk to hey, youget a summary of a meeting.
It's great, um so anotherquestion for you, and this one
is more Out of curiosity, and itcomes from this takeaway we had
from last week, which was thatwealth firms are really looking
to their partners, theirprovider partners, to help lead
(11:16):
them in the conversation of AI.
Speaker 1 (11:18):
Do you agree with that?
Yeah, I think so.
I think a lot of these toolsare going to have to be built by
vendors for the majority offirms.
I will tell you, I've spentsome.
I am a bit of a closet nerd andI've been trying to build.
Speaker 2 (11:32):
I never knew that.
Speaker 1 (11:33):
I know, I know I literally I'll check it.
If I move my camera I'll showyou.
I got it.
One of those Nvidia 4090 cards,put it in my uh, my, my desktop
at work to try and build out myown GPT system based on
documents.
We have to kind of see what itcould answer.
And I'll say, even though thereare tons of open source items
(11:54):
available, apis, all that stuffit is still not easy to then
also do it in a secure, scalableway for enterprises.
So I can definitely see, andthat's just one use case, right,
that's just like what they calla RAG system or a retrieval
augmented generation type system.
That's just one type case,right, that's just like what
they call a RAG system or aretrieval augmented generation
type system.
That's just one type of AI andthere's so many others and I
think it's hard, you know.
(12:15):
I'll give you another example.
Like there's a vendor we'vebeen looking at that's
specifically in the lifeinsurance space.
It's been doing some kind ofinteresting things and part of
it it was they had to buildthese models, right, that takes
a long time to just build themodel.
So I can see why vendors, a lotof it's just going to take time
to like train these models towork the right way.
I know another one, rich Walker, that you're friends with.
(12:39):
You know he's got a new productthat they've come out with that
they've trained their model onforms and helping to
automatically pull data valueout of forms and stuff.
So I think that's a goodexample of where you need
partners Like that would bereally hard for a single firm to
do on their own.
Something like that wherethere's a lot of mapping
involved, or something like thatwhere you need a lot of model
training.
Speaker 2 (12:55):
Yeah, no, it's fascinating to me.
I think, you know, one of thethings we also kind of have come
away with over the last fewmonths is there almost needs to
be a primer on all this, right,an overview, you know, an
(13:15):
overview piece where peoplecould really learn about the
different aspects of AI and andyou know the different types of
models.
You know I always love to askthose questions, so you know I,
I obviously I talked to a lot ofpeople in the industry, right,
we're, you know, we're lookingfor insight as well.
I've had the opportunity totalk about, you know, what
Salesforce's strategy is withMichelle Feinstein Really,
really interesting where they'regoing with it.
Being able to spend some timewith our mutual friend, sid
(13:37):
Yedimendra, who with SurgeVentures and they're doing some
neat things with compliance,artificial intelligence who are
some folks out there that you'reseeing, both as a provider and
maybe, additionally, as you know, people that are are solving,
utilizing ai and solvingproblems uh.
Speaker 1 (13:56):
So just a couple people maybe to think about.
Uh, mark butler with his wealthmanagement gpt and I know they
recently they had that in pilotfor quite some time uh, a chat
gpt like tool that has somespecific prompt for advisors to
use.
So I know he just recently wasadded to the Kitsies FinTech map
.
Jonathan Michael he justrenamed his company AdvisorX AI.
(14:19):
He's been putting out a lot ofmaterial on LinkedIn and
specifically around AI.
Two other names I'd mentionedI'm local to Cleveland, so
KeyBank is in our backyard.
I'm not going to say his name,right.
Azeem Shaikh.
He puts out a to Cleveland, soKeyBank is in our backyard.
I'm not going to say his name,right.
Azeem Shaikh he puts out a lotof material from KeyBank, just
kind of what they're doing.
He's kind of on their strategyteam, so he puts out a lot of
(14:40):
really interesting thoughtpieces around AI.
And then, lastly, I just gotintroduced to the Responsible AI
Institute, which I think isalso really interesting.
So, less on the tech front butmore on the sort of regulatory
compliance front.
I think they're starting to putout some really good
information.
And I know Jackson NationalLife is in our space, the
(15:03):
finance space, and they're amember of the Responsible AI
Institute, as a member of FOOT,so they've got some really good
I think just thought piecesabout some of the Responsible AI
Institute as a member, sothey've got some really good.
I think just thought piecesabout some of the more
regulatory things we should bethinking about as we're
evaluating AI systems.
Speaker 2 (15:16):
Got it, got it.
So you know, just big rocksright when you're looking at
what AI could do for your firm.
What are a couple items thatyou're taking away as hey, this
is something we're superinterested in.
And then, on the flip side,what are some things that maybe
you've thought about but haveseen data proving otherwise?
Speaker 1 (15:41):
So I will start with expectations around software
developers, right, so I'll putthis in the overhyped category.
I have a friend who they have apretty large software
development organization, sodeveloper productivity is really
, really important for them.
And they were, you know theirexecutive team was initially
seeing you know marketing slidesthat were come out and saying
(16:02):
things like you know, you'll get40 to 50% developer
productivity.
I mean that's huge becausetypically these are, you know,
high dollar resources.
So what they did is theyactually went out and had
studies done by McKinsey,microsoft and others to see, ok,
if we have, you do, anengagement, what can we actually
expect as a productivity liftfor our developers?
(16:23):
And when they had to put theirreputation on the line, those
numbers came back at like 2% to%.
So I mean not zero.
So there's definitely a rolefor some of these tools, but it
might not be as big as people orsome of these.
You know people are touting butdefinitely, you know, I think
even a two to 4% lift insomebody's you know work output
(16:44):
could be really good and I thinkwhat we've seen is especially
with, like, younger developers,it's very helpful, right, they
can quickly, you know, don'thave to sort through a lot of
different websites or YouTubevideos, they can get the answer
pretty quickly for kind of morebasic things.
As it comes to more likearchitecture, engineering, like
that's a complete differentballgame.
So I think where I'm probablymost excited still a little bit
(17:07):
cautious, but starting to putsome more effort into is, again,
I mentioned a little bitearlier this RAG or retrieval,
augmented generation.
So this is like you can putyour own organization's
documents into a chat GPT likeresponse.
So, for example, let's say youhave, you know, in our world
broker dealer world, we've gotwritten supervisory procedures,
right.
Some of these can be very, verylong.
(17:28):
What if you could put adocument like this inside a
system and people could justsimply ask the question?
You know, and the good thing iswith these systems too, they
often reference the document andeven sometimes the page number
from which it came.
So there's a little bit moretracking involved.
(17:48):
With somebody's asking thesystem it can be verified like
okay, that's, the document camefrom, this was the page number
to go and check it.
So I still think it's kind ofearly to see if, if that would
actually prove out, and some ofit's going to depend on the
quality of the organization's.
You know inputs right Ifthey've got good inputs or if
they've got outdateddocumentation floating around.
Speaker 2 (18:13):
Fascinating, fascinating Closing question.
You know we've we have a greatdeal of rhetoric going on with
AI right now.
As I said earlier, I don't eventhink we've touched the surface
.
We're hearing a lot of wealthfirms are you know?
They want to be the theproviders to be their AI
catalyst?
Completely get it.
You know.
They want to be the providersto be their AI catalyst?
Completely get it.
You know, is AI going to gettraction or is it going to be
(18:35):
relegated to another one ofthose buzzwords?
What do you think?
Speaker 1 (18:38):
It's going to get traction, I'm going to go.
I'm going to go on record assaying it's going to get
traction.
I, I feel so.
I'm just giving you in myperspective, like when we heard
a lot about RPA robotic processautomation I always kind of
thought like, ah, this is aBand-Aid.
Really somebody wants to doAPIs.
I get, maybe you've got somelegacy systems.
You know, that was an example Ithought was kind of overhyped.
(18:59):
I still think there's a ton ofpotential out there with AI to
help really accelerate both ourlearning and just how we do our
day-to-day jobs.
Speaker 2 (19:10):
Fabulous, Jeff.
Thank you for being a guest onthe Beacon Flash podcast.
Always love to have you here.
Speaker 1 (19:16):
All right, thank you, chip.
Speaker 2 (19:18):
We've been talking to Jeff Moore, who is the CIO of
Valmark Financial Group, and ithas been a real pleasure.
If you need some industryperspective or help, visit
BeaconStrategiesLLCcom and don'tmiss out on our exclusive
podcast, the Beacon Flash forindustry news and trends.
Have a great day.
Speaker 1 (19:40):
Thank you for listening to the Beacon Flash
podcast.
We'll see you next time and besure to click subscribe to get
future episodes.
Popular Podcasts
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.
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
The Breakfast Club
The World's Most Dangerous Morning Show, The Breakfast Club, With DJ Envy, Jess Hilarious, And Charlamagne Tha God!