April 29, 2025 • 21 mins
Welcome to another edition of The Disruptor Podcast!
In this episode, your host, John Kundtz, interviews Caitlyn Truong, CEO and Co-founder of Zengines.
This is part one of our two-part series.
Today's show explores how Zengines disrupts organizations' efforts to automate the end-to-end data conversion process.
Caitlyn shares information from her background in electrical and computer engineering and consulting. She often saw the problems with data conversion and migration in large organizations, especially in financial services.
This led her to co-found Zengines to help ensure data stays useful during modernization.
Points from the Discussion:
Understanding Data Migration: It includes all steps to get a new system running, from getting data to testing.
Common Pitfalls: Not accounting for all steps, lack of a plan, not using tools, and not underestimating the work effort.
Advice for Smoother Migration: Understand all steps, plan clearly, use tools (especially AI), empower business analysts, and work in steps.
Zengines' Disruptive Way:
- Zengines gives business analysts AI tools to automate mapping and changing data.
- This "shift left" approach reduces the need for large teams, making data conversion faster and more efficient by letting the business directly influence data changes.
- This approach moves the work earlier in the process, making data conversion less expensive, more productive, and quicker by letting the business directly affect how data is changed.
To learn more about Caitlyn and Zengines, visit their website (Zengines.AI) and connect with Caitlyn on LinkedIn.
Stay tuned for Part 2 of this conversation, where Caitlyn and I shift gears and explore the challenges many enterprises face with Mainframe Modernization.
You won't want to miss it.
Comments or Questions? Send us a text
***
Engage, Share, and Connect!
Spread the Word: Valuable insights are best when shared. Share this episode with peers who may benefit from it if you find it insightful.
Your Feedback Matters: How did this episode resonate with you? Share your thoughts, insights, or questions. Your engagement enriches our community.
Stay Updated: Don’t miss out on further insights.
Subscribe: You can listen to our podcast, read our blog posts on Medium, Substack, and LinkedIn, and watch our YouTube channel.
Collaborate with The Disruptor and connect with John Kundtz.
Got a disruptive story to share? We’re scouting for remarkable podcast guests. Nominate a Disruptor.
Thank you for being an integral part of our journey.
Together, let’s redefine the status quo!
Tips are welcomed and appreciated, too!
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:01):
Disruption and Data Transforming, migration and
Modernizing Mainframes.
Hi everyone, I'm your host,john Kuntz, and welcome to
another edition of the DisruptorPodcast.
For those that are new to ourshow, we started this series
back in December of 2022 as aperiodic segment of the Apex
Podcast.
Our vision was to go beyond theconventional wisdom by
(00:23):
confronting the status quo andexposing the raw power of
disruptive thinking.
Today, we talk with CaitlinTruong, ceo of Zen Engines, as
we explore how her companydisrupts organizations,
automates the end-to-end dataconversions.
We will discuss valuable adviceon the pitfalls, the mistakes
(00:45):
that many executives make whenattempting digital
transformation.
Welcome to the show, caitlin.
Speaker 2 (00:50):
Thank you so much, John.
Thank you for having me Twotopics I love to talk about.
Speaker 1 (00:55):
Great.
In prep for this show, we had alot of interesting crosses of
our professional experiences.
Speaking of that, why don't youtell us a little bit about your
background, your education,your experiences, how you came
to start your company?
And you can start anywhere youwant.
Speaker 2 (01:11):
Thanks so much, John.
So my background is inengineering.
I am an electrical and computerengineer and spent a couple
stints building circuits formobile devices way back when
then went to the dark side ofconsulting, where I spent a fair
amount most of my career in theconsulting industry helping
(01:32):
large organizations withtechnology transformation type
initiatives.
Most of that was in financialservices, John, and that was
where I crossed paths with oneof the other Zengens co-founders
.
We talked about data conversion, data migration.
We ran into it all of the timeand mostly it was because we saw
that organizations are alwayschanging, they're always
(01:52):
modernizing, they're alwayslooking to stay relevant and
keeping their technology current.
Your systems hold data.
You want to preserve and makethat data continuously usable,
and so there's always thishurdle of going through a
conversion or a migration tomake sure you can execute on
your business strategy.
So that was the pain point thatwe saw.
(02:14):
It was a challenge we saworganizations face all
organizations, all sizes, allindustry, and we said there's
got to be a better way.
So that was how this all cameto be.
I'm excited because it's asolvable problem.
Speaker 1 (02:29):
I agree.
I had similar experiences.
A good bit of my career I wasrunning part of IBM's business
where we did data centermigrations and consolidations.
The Achilles heel of thoseprojects was our ability to
migrate the data andapplications.
One little glitch would takethe whole project down.
Let's dive into data migration.
(02:49):
What are some of the mostcommon mistakes or pitfalls
organizations tend to make whentrying to take a more
traditional approach to theirdigital transformation and or
their data migration projects?
Speaker 2 (03:01):
Well, first, I always like to make sure that I
clarify.
I should not redefine datamigration, john, because I find
that, based on one's experience,you might believe or understand
data migration to only be apart of it.
When I think of data migration,I think it is all of the things
associated with helping get anew system or a new data store
(03:23):
to be live.
In other words, it is thatupfront ingestion of what might
be the sources of data thatyou're going to move.
It is analyzing it so that youunderstand what you will have to
deal with when you're goingthrough a migration.
It is mapping the data so thatyou know this is what it looks
like, this is where it's goingto.
(03:43):
It's changing or transformingthat data because you probably
will need it to fit a differentway or look a different way.
Then it's the physical portionof getting the data out,
applying changes, putting itinto a new data store and
testing it.
I say all of that's part ofdata migration because
ultimately, we're talking aboutmaking that data still be a
(04:03):
valuable asset to you, and it'simportant to make sure that's
clear.
My experience has been that,depending on where someone might
have participated in that fulljourney, you might only think of
data migration as just themapping portion or just the ETL
portion.
It's important to thinkholistically because you need to
solve for all of those things,all of those complexities, all
(04:26):
of those potential pitfalls thatcome in across all steps.
So first it was important tolevel set that, from my view,
all of those things areimportant.
Pitfall number one is notaccounting for all of it.
Sometimes, when folks aren'texperienced, they might only
think that data migration isjust that portion of oh right,
(04:47):
before go live, we need to movethat data.
But really you needed to haveprepared all of those previous
steps.
So pitfall number one is nottruly understanding that all of
that is in the picture.
And then pitfall number two isnot having a really good plan as
to how will you tackle this.
Because in having all of thosesteps as part of the effort, who
(05:07):
will do what?
Who knows what?
Is it your responsibility?
Is it the software vendor orthe servicer that's accountable
for making sure the data showsup in the new target locations?
Having a good plan, making sureyou can execute well and bring
in the right experts, is pitfallnumber two.
And pitfall number three is nottruly leveraging and
(05:29):
capitalizing on all of the toolsthat are available.
I think a lot of folks probablybecause they didn't think of
all of the steps that areincluded and then they didn't
have a good plan because, again,not everyone lives data
conversions every day.
But without having that thefirst two steps in place, then
you don't know what tools can beapplied and there are just
(05:50):
tools that can really help, asopposed to thinking you have to
go through it the old school wayof humans and spreadsheet.
So I would just start withthose three to start, and then
there's things that I thinklater we could talk about, john,
which is patterns that we knoware true around data migrations
and how you can help people getthrough it successfully.
Speaker 1 (06:11):
I wholeheartedly agree with those pitfalls.
Based on my experience, I'veseen a number of projects
underestimated the planning andthe big picture.
They just figured they justtake a bunch of servers and the
data on there and just sort oflift them, shift them into a new
data center.
Those were career-breakingprojects for some of the
(06:31):
executives I worked with, theones that avoided the pitfalls
you just explained, and theprojects were super successful.
Most of those guys and gals gotpromoted.
Many of the CIOs that took itfor granted and sort of try to
do it on their own.
As you mentioned, most peopledon't do this every day.
They've never done it before.
They underestimate the level ofeffort.
It could be a career-endingmove.
(06:53):
I've seen it on both counts.
So, based on that, whatguidance would you give our
listeners to avoid thesepitfalls, to ensure that there
is a smooth data migrationprocess?
Speaker 2 (07:03):
First, make sure there is a clear understanding
of all steps we promote.
What is the data migration 101.
I'm very clear in making surethat I think there is a picture
that summarizes this well and isvery digestible.
Really understand allactivities that go into a data
conversion Call it many thingsdata conversion.
Sometimes it's a data migration.
(07:24):
Is it part of the system'simplementation?
But make sure that there is anunderstanding of all of those
steps and then having that planso that you say who is going to
be accountable for making thecompleting the mapping steps,
who will be responsible forwriting transformation rules,
etc.
And once you can have that inplace, then you can say let's
(07:49):
talk about accelerating.
If this is how we're going todole out responsibilities, let's
talk about now not doing itmanually and how do we take
advantage of tools that are outthere.
And, more importantly, I thinkthat one of the best things to
consider here is that this isall about pattern recognition
and it's really key that teamsstart leveraging AI.
(08:11):
I know that everyone has maybediffering opinions and it's
becoming more mainstream, john.
Everyone's looking to try tothrow AI at this.
I think it's the right thing.
This is about patternrecognition and the great part
when I started sharing with youthat I believe data migrations
can be successful is becauseit's data and it's static and
it's about pattern recognition.
(08:31):
So, instead of leveraginghumans who have experience and
there's a risk to thatexperience being in a human who
might walk away from thatconversion project, who might
not be interested in staying thefull duration of that
conversion project we use toolsand software.
So in that way, I think it'sreally important that folks
(08:51):
start understanding that thereare really good tools out in the
marketplace that are powered byAI that can help you with that.
Obviously, zengens is one ofthem.
Speaker 1 (09:01):
Excellent.
Appreciate that.
I remember back when we weredoing data center consolidation
work.
There were tools, but theyweren't very good, so much of
this was manual.
We had spreadsheets and theseprojects could take five or six
months to physically consolidatea data center.
But understanding the data,understanding how the
(09:21):
applications use the data, whichapplications needed to get
which data where, when, was amajor part, and it was all done
manually.
If somebody left the project ormissed one thing, we would have
to redo the whole project planagain.
So how does your approach withyour company benefit your
(09:41):
clients in terms of core datamigration?
Speaker 2 (09:44):
One thing that we really looked at and said we
have to do differently is whywere there so many people?
Why was it always an army ofpeople to do the work?
It's all those things we saidbefore.
Right, it was a bit ofsometimes you didn't allocate it
correctly and then you startedjust adding more and more people
because you were findingexperts.
But it was that allocation.
(10:05):
The second pain point that Italked about, john, around the
not understanding who will dowhat.
I feel strongly.
This belongs to the business.
Data belongs to the business.
The business knows the dataright and what I mean by the
data belongs to the businesses.
They are the ones to ultimatelyfinalize that this is
acceptable.
This is what I want the data tolook like by the time it shows
(10:27):
up in the cloud, in my newsystem or in Snowflake and
Databricks as part of the dataproduct.
They have enough familiaritywith the data to decide what
looks right or wrong or to setthe requirements.
One of the best things to do isto empower the analyst, empower
the business team to do the dataconversion, because today we
(10:49):
found is that it was a bit ofbusiness comes in here to do, to
share.
Oh, this is where you'll findthe data.
They go away because then thetechnical team comes in to do
the okay, we're going to set upand connect to the database, and
then business comes back whensomeone starts to profile the
data and says, well, does this?
Is this right or wrong?
Should this have null values?
(11:09):
I'm looking for a patternbecause this is a fixed list of
items, but I'm seeing theserandom things.
Is that right or wrong?
Again, business has to be theone to define if that's right or
wrong.
Business is the one that startsto do some of that mapping
right, because they say well, Iknow that this is here, but I
don't know what that means overthere.
All I'm looking to emphasizehere is one of the things that
(11:30):
Zenges focuses on is to create aproduct that's for the business
the BA, the business analystand empower that person to
finish the data conversion foracross all of those steps that I
talked about, becauseultimately, they're the final
decision makers.
So take away the need to bringin a lot of people, because you
(11:50):
brought in people because theyknew bits and pieces of the
process.
So, in other words, make thetool or the product intelligent
enough to be able to supportthose other areas, accelerate
the work for the BA, because youdon't want the BA to be doing
mundane, rote, manual,rules-based types of activities.
Ie, this is where AI we putpattern recognition into
(12:13):
something like mapping and wesay you don't have to do
guesswork, let AI do theguesswork and we give you some
sense of level of confidenceassociated with that prediction.
You, as the business decisionmaker, can validate whether or
not you like that mapping.
And then we know that businessdrives transformation changes.
They say I want you to jointhese two fields, I want you to
concatenate this or split this.
(12:34):
So let's give business a toolto allow them to get that data
change applied but not have todo the technical steps of
writing the syntax.
So again, this is where Zengensoffers an LLM.
We allow the user to writeplain English on what that data
transformation rule should beand they see what syntax is
(12:54):
auto-generated and then theyapprove or disapprove that.
One of the biggest things wasalways thinking that you had to
bring the representative of thedifferent areas to the table,
but instead it really should beto cater to the decision maker,
which is the business, and givethem good tools to make sure
that they can get through thatprocess really fast, right.
Take away the technicalchallenges.
(13:15):
Take away the fact that youdon't always know what's on the
other side, so let AI predictwhat's best on the other side.
Speaker 1 (13:24):
This is great.
This is the classic what wewould call a shift left approach
.
At the end of the day, as yousaid, the business units
anything that has to do withtechnology.
The only reason it exists is sothe business can do something,
make money, get more customers,sell more stuff.
The more you can move theactual doing towards where the
responsibility lies, ie thebusinesses a shift left approach
(13:48):
.
It's less expensive, moreproductive and faster, so all
three benefits.
A classic shift left approachwhere the businesses who own the
data, the business response,the P&L, so to speak, let them
be more productive.
In the old days, we couldn't dothat because they didn't have
(14:08):
the technical capability.
That's why we had to bring inarmies of people and it's why it
took so long.
So what I'm hearing and what issuper exciting about what your
company is doing is now you'reenabling, using English language
, large language models, aiapproaches to help these
businesses do what they need todo without engaging a ton of
(14:30):
other people.
Speaker 2 (14:31):
Yes, I think that was part of our secret sauce really
understanding the problem andmaking sure that we solve the
problem for the right person.
John, and one example that cameto mind this was one of our
customers was that the businessanalysts.
As I said, they ultimately madethe final decision as to this
(14:51):
is what I want it to look likeby the time it shows up in the
production system.
And that teammate was workingwith the DBA, who would write
the SQL query, to extract it andshow it to the business before
it got approved.
Well, the business user wouldsend an email or have a
conversation and say okay, theseare the fields I want and this
is what I want you to do tothose fields.
And that was written in plainEnglish over an email sent off
(15:12):
to the DBA.
The DBA picked it up, sawsomething, interpreted it and,
in this case, as probably halfthe time, interpreted it
somewhat incorrectly, did thequery put in the transformation
rule, sent it back to thebusiness the business doesn't
see the query, just the outputand said that's not what I want.
Can you do this now thatiteration goes back and forth
(15:33):
and back and forth because, as Isaid, data is ultimately the
truth teller.
Business is the decision maker,but then you just always had
that friction because theycouldn't get to the data
directly or they didn't know howto write the right query in
order to get the data to lookthe way they wanted to.
But this second piece that Iwould say is that Data
(15:54):
conversions are alwaysunpredictable.
With AI that's patternrecognition Unfortunately, we
can't go back and make it bepredictable.
I can't go back in time 20, 30years ago and change what was
inserted into the database wayback when, but what we can do is
at least have some tool thatanalyzes it really fast and
allows you to look at it andgives you some sense of this is
(16:16):
what I think it looks like, andthe business user can iterate,
and iteration is key.
So that was the other thingthat I think is really important
is that you just take a shot atit, you run it through, you
take a look at what you'vemapped, you take a look at what
you've converted and then yousay, does that look right?
And start with a small datasample set, because then you can
build rules on top of rules, asopposed to believing that
(16:38):
you've got everything you needand now you can accommodate for
all of the variations oftransformation rules that you
might need to put in there.
So just start with a first dataset, get it through very fast
and, like I said, we're talkingabout generating a load file in
minutes.
Look at that load file and thensay do you like that?
If it looks good, all right,let's continue to add more
(16:59):
upgrades in and continue to seeif your rule applies and if you
need to change it, then put morerules on top of that.
But I'm just saying it's thefact that you just don't know
what is in the data.
So give you a tool that allowsyou to get through iterations
really fast and allows you toplay around with getting it
perfect.
Speaker 1 (17:17):
That's tremendous advice.
We've covered really essentialground on modernizing and data
migration.
I heard 60% of all digitaltransformation initiatives fail.
They don't understand whatthey're getting into or the
level of effort.
What you've discussed is a hugebenefit for digital
(17:38):
transformation.
Wow, Thanks for sharing allthat.
Why don't we wrap this up?
I appreciate your insights andexperiences.
My last question is is thereanything I haven't asked you
that you'd like to share beforewe wrap the show up?
Speaker 2 (17:51):
When we think of data migration, john, a major aha
moment for me was we need toempower the business analysts
because, again, I believe thatit's the data.
And we need to empower thebusiness analysts because, again
, I believe that it's the dataand the decision belongs to the
business.
And I think all the yearsbefore one, it was that there
was always a need to havemultiple translation points.
(18:11):
Right, oh, business only knowsone side of the transformation
the conversion.
They knew their system but theydidn't know the target state
system, and then also, businesswasn't as technically adept as
some of the SMEs.
So let's unpack that and solvefor the business.
So I just think that that wasone aspect that I think is
really important because,compared to a lot of the other
(18:34):
products in the marketplacetoday, data mapping, data
integration that's a need.
In some cases that's an internaltalk, the data pipes and that's
a technology problem.
I'm trying to solve theonboarding, the new
implementation, the post-mergerintegration, where you want that
data to truly be right.
So that's one thing that Iwould say on the data conversion
, data migration side.
(18:54):
So this is a little bit of theproduct mindset I think
organizations and teams canimagine and design and design
and talk about it for a reallylong time and I say you'll learn
and you will get so much morevalue just by doing.
You might make mistakes becauseyou didn't plan for all of it,
but you're at least alreadystarted and not waiting to
(19:16):
believe that you'll discover allof it through planning.
Speaker 1 (19:20):
Great advice, I agree .
How do you eat the elephant?
One?
Speaker 2 (19:22):
bite at a time.
Speaker 1 (19:24):
Yes.
How can people learn more aboutyou and your services?
What are your socials?
What's the best way?
We'll include all of these inthe show notes.
Speaker 2 (19:31):
Thank you, john.
So Zengens' website is wwwaiand that's Zengens with a z or a
z in front of the word engines,plural and, by the way, our
domain.
We had AI from the verybeginning, so way before a lot
of these other companies startedto change to a ai on the domain
name and I'm on LinkedIn, sofolks can always connect with me
(19:56):
or send me a direct message andwe have information as to how
to reach out to our team, to geta demo and work with us, to go
through a POC or a trial.
In this world, everyone likesto take a look and see how it
works and we would be delightedto have people give it a shot.
I really believe, when it comesto data conversions, it doesn't
(20:16):
take an army.
I really believe everyone cando it themselves.
As I said, it's to the businessuser.
So if we start to change themindset and say, hey, this can
be self-service, that's what I'mhoping to move to, so that
everyone can change faster.
Speaker 1 (20:30):
Excellent.
This has been a great interview.
You're so energetic, soknowledgeable.
It's a pleasure to have you asa guest on the Disruptor
Everybody listening.
Please don't forget to connectwith Caitlin on LinkedIn.
Check out their website.
And, caitlin, before I wrap up,I always give the guest the
last word.
Then we'll say goodbye.
Speaker 2 (20:51):
Thanks, john.
Well, first I wanted to saythank you.
We had so much fun and I thinkI capitalized on the talk time
here, but your listeners shouldknow just what an interesting
and knowledgeable person you areand I had so much fun learning
from you in our conversations.
And then also, just to wrap, Isay that it's all about
gratitude.
I appreciate so much theopportunity to spend time with
(21:14):
you, that you've made time andspace for me, and that this
experience of building a companyto deliver value to people.
I believe data is solvable.
Let's help people focus onother things and let's make the
data be not something that theyhave to worry about.
Speaker 1 (21:30):
We're going to come back with part two of our
podcast and when we return we'regoing to shift gears and we're
going to explore the challengesthat many enterprises face on
mainframe modernization.
Stick with us and we'll be backAll righty everybody.
So thanks, caitlin, for beingon our show I'm John Kuntz and
(21:50):
thanks for joining this editionof the Disruptor Podcast.
Have a great day, take care.
Speaker 2 (21:56):
Thanks, John.
Disruption and Data Transforming, migration and
Modernizing Mainframes.
Hi everyone, I'm your host,john Kuntz, and welcome to
another edition of the DisruptorPodcast.
For those that are new to ourshow, we started this series
back in December of 2022 as aperiodic segment of the Apex
Podcast.
Our vision was to go beyond theconventional wisdom by
(00:23):
confronting the status quo andexposing the raw power of
disruptive thinking.
Today, we talk with CaitlinTruong, ceo of Zen Engines, as
we explore how her companydisrupts organizations,
automates the end-to-end dataconversions.
We will discuss valuable adviceon the pitfalls, the mistakes
(00:45):
that many executives make whenattempting digital
transformation.
Welcome to the show, caitlin.
Speaker 2 (00:50):
Thank you so much, John.
Thank you for having me Twotopics I love to talk about.
Speaker 1 (00:55):
Great.
In prep for this show, we had alot of interesting crosses of
our professional experiences.
Speaking of that, why don't youtell us a little bit about your
background, your education,your experiences, how you came
to start your company?
And you can start anywhere youwant.
Speaker 2 (01:11):
Thanks so much, John.
So my background is inengineering.
I am an electrical and computerengineer and spent a couple
stints building circuits formobile devices way back when
then went to the dark side ofconsulting, where I spent a fair
amount most of my career in theconsulting industry helping
(01:32):
large organizations withtechnology transformation type
initiatives.
Most of that was in financialservices, John, and that was
where I crossed paths with oneof the other Zengens co-founders
.
We talked about data conversion, data migration.
We ran into it all of the timeand mostly it was because we saw
that organizations are alwayschanging, they're always
(01:52):
modernizing, they're alwayslooking to stay relevant and
keeping their technology current.
Your systems hold data.
You want to preserve and makethat data continuously usable,
and so there's always thishurdle of going through a
conversion or a migration tomake sure you can execute on
your business strategy.
So that was the pain point thatwe saw.
(02:14):
It was a challenge we saworganizations face all
organizations, all sizes, allindustry, and we said there's
got to be a better way.
So that was how this all cameto be.
I'm excited because it's asolvable problem.
Speaker 1 (02:29):
I agree.
I had similar experiences.
A good bit of my career I wasrunning part of IBM's business
where we did data centermigrations and consolidations.
The Achilles heel of thoseprojects was our ability to
migrate the data andapplications.
One little glitch would takethe whole project down.
Let's dive into data migration.
(02:49):
What are some of the mostcommon mistakes or pitfalls
organizations tend to make whentrying to take a more
traditional approach to theirdigital transformation and or
their data migration projects?
Speaker 2 (03:01):
Well, first, I always like to make sure that I
clarify.
I should not redefine datamigration, john, because I find
that, based on one's experience,you might believe or understand
data migration to only be apart of it.
When I think of data migration,I think it is all of the things
associated with helping get anew system or a new data store
(03:23):
to be live.
In other words, it is thatupfront ingestion of what might
be the sources of data thatyou're going to move.
It is analyzing it so that youunderstand what you will have to
deal with when you're goingthrough a migration.
It is mapping the data so thatyou know this is what it looks
like, this is where it's goingto.
(03:43):
It's changing or transformingthat data because you probably
will need it to fit a differentway or look a different way.
Then it's the physical portionof getting the data out,
applying changes, putting itinto a new data store and
testing it.
I say all of that's part ofdata migration because
ultimately, we're talking aboutmaking that data still be a
(04:03):
valuable asset to you, and it'simportant to make sure that's
clear.
My experience has been that,depending on where someone might
have participated in that fulljourney, you might only think of
data migration as just themapping portion or just the ETL
portion.
It's important to thinkholistically because you need to
solve for all of those things,all of those complexities, all
(04:26):
of those potential pitfalls thatcome in across all steps.
So first it was important tolevel set that, from my view,
all of those things areimportant.
Pitfall number one is notaccounting for all of it.
Sometimes, when folks aren'texperienced, they might only
think that data migration isjust that portion of oh right,
(04:47):
before go live, we need to movethat data.
But really you needed to haveprepared all of those previous
steps.
So pitfall number one is nottruly understanding that all of
that is in the picture.
And then pitfall number two isnot having a really good plan as
to how will you tackle this.
Because in having all of thosesteps as part of the effort, who
(05:07):
will do what?
Who knows what?
Is it your responsibility?
Is it the software vendor orthe servicer that's accountable
for making sure the data showsup in the new target locations?
Having a good plan, making sureyou can execute well and bring
in the right experts, is pitfallnumber two.
And pitfall number three is nottruly leveraging and
(05:29):
capitalizing on all of the toolsthat are available.
I think a lot of folks probablybecause they didn't think of
all of the steps that areincluded and then they didn't
have a good plan because, again,not everyone lives data
conversions every day.
But without having that thefirst two steps in place, then
you don't know what tools can beapplied and there are just
(05:50):
tools that can really help, asopposed to thinking you have to
go through it the old school wayof humans and spreadsheet.
So I would just start withthose three to start, and then
there's things that I thinklater we could talk about, john,
which is patterns that we knoware true around data migrations
and how you can help people getthrough it successfully.
Speaker 1 (06:11):
I wholeheartedly agree with those pitfalls.
Based on my experience, I'veseen a number of projects
underestimated the planning andthe big picture.
They just figured they justtake a bunch of servers and the
data on there and just sort oflift them, shift them into a new
data center.
Those were career-breakingprojects for some of the
(06:31):
executives I worked with, theones that avoided the pitfalls
you just explained, and theprojects were super successful.
Most of those guys and gals gotpromoted.
Many of the CIOs that took itfor granted and sort of try to
do it on their own.
As you mentioned, most peopledon't do this every day.
They've never done it before.
They underestimate the level ofeffort.
It could be a career-endingmove.
(06:53):
I've seen it on both counts.
So, based on that, whatguidance would you give our
listeners to avoid thesepitfalls, to ensure that there
is a smooth data migrationprocess?
Speaker 2 (07:03):
First, make sure there is a clear understanding
of all steps we promote.
What is the data migration 101.
I'm very clear in making surethat I think there is a picture
that summarizes this well and isvery digestible.
Really understand allactivities that go into a data
conversion Call it many thingsdata conversion.
Sometimes it's a data migration.
(07:24):
Is it part of the system'simplementation?
But make sure that there is anunderstanding of all of those
steps and then having that planso that you say who is going to
be accountable for making thecompleting the mapping steps,
who will be responsible forwriting transformation rules,
etc.
And once you can have that inplace, then you can say let's
(07:49):
talk about accelerating.
If this is how we're going todole out responsibilities, let's
talk about now not doing itmanually and how do we take
advantage of tools that are outthere.
And, more importantly, I thinkthat one of the best things to
consider here is that this isall about pattern recognition
and it's really key that teamsstart leveraging AI.
(08:11):
I know that everyone has maybediffering opinions and it's
becoming more mainstream, john.
Everyone's looking to try tothrow AI at this.
I think it's the right thing.
This is about patternrecognition and the great part
when I started sharing with youthat I believe data migrations
can be successful is becauseit's data and it's static and
it's about pattern recognition.
(08:31):
So, instead of leveraginghumans who have experience and
there's a risk to thatexperience being in a human who
might walk away from thatconversion project, who might
not be interested in staying thefull duration of that
conversion project we use toolsand software.
So in that way, I think it'sreally important that folks
(08:51):
start understanding that thereare really good tools out in the
marketplace that are powered byAI that can help you with that.
Obviously, zengens is one ofthem.
Speaker 1 (09:01):
Excellent.
Appreciate that.
I remember back when we weredoing data center consolidation
work.
There were tools, but theyweren't very good, so much of
this was manual.
We had spreadsheets and theseprojects could take five or six
months to physically consolidatea data center.
But understanding the data,understanding how the
(09:21):
applications use the data, whichapplications needed to get
which data where, when, was amajor part, and it was all done
manually.
If somebody left the project ormissed one thing, we would have
to redo the whole project planagain.
So how does your approach withyour company benefit your
(09:41):
clients in terms of core datamigration?
Speaker 2 (09:44):
One thing that we really looked at and said we
have to do differently is whywere there so many people?
Why was it always an army ofpeople to do the work?
It's all those things we saidbefore.
Right, it was a bit ofsometimes you didn't allocate it
correctly and then you startedjust adding more and more people
because you were findingexperts.
But it was that allocation.
(10:05):
The second pain point that Italked about, john, around the
not understanding who will dowhat.
I feel strongly.
This belongs to the business.
Data belongs to the business.
The business knows the dataright and what I mean by the
data belongs to the businesses.
They are the ones to ultimatelyfinalize that this is
acceptable.
This is what I want the data tolook like by the time it shows
(10:27):
up in the cloud, in my newsystem or in Snowflake and
Databricks as part of the dataproduct.
They have enough familiaritywith the data to decide what
looks right or wrong or to setthe requirements.
One of the best things to do isto empower the analyst, empower
the business team to do the dataconversion, because today we
(10:49):
found is that it was a bit ofbusiness comes in here to do, to
share.
Oh, this is where you'll findthe data.
They go away because then thetechnical team comes in to do
the okay, we're going to set upand connect to the database, and
then business comes back whensomeone starts to profile the
data and says, well, does this?
Is this right or wrong?
Should this have null values?
(11:09):
I'm looking for a patternbecause this is a fixed list of
items, but I'm seeing theserandom things.
Is that right or wrong?
Again, business has to be theone to define if that's right or
wrong.
Business is the one that startsto do some of that mapping
right, because they say well, Iknow that this is here, but I
don't know what that means overthere.
All I'm looking to emphasizehere is one of the things that
(11:30):
Zenges focuses on is to create aproduct that's for the business
the BA, the business analystand empower that person to
finish the data conversion foracross all of those steps that I
talked about, becauseultimately, they're the final
decision makers.
So take away the need to bringin a lot of people, because you
(11:50):
brought in people because theyknew bits and pieces of the
process.
So, in other words, make thetool or the product intelligent
enough to be able to supportthose other areas, accelerate
the work for the BA, because youdon't want the BA to be doing
mundane, rote, manual,rules-based types of activities.
Ie, this is where AI we putpattern recognition into
(12:13):
something like mapping and wesay you don't have to do
guesswork, let AI do theguesswork and we give you some
sense of level of confidenceassociated with that prediction.
You, as the business decisionmaker, can validate whether or
not you like that mapping.
And then we know that businessdrives transformation changes.
They say I want you to jointhese two fields, I want you to
concatenate this or split this.
(12:34):
So let's give business a toolto allow them to get that data
change applied but not have todo the technical steps of
writing the syntax.
So again, this is where Zengensoffers an LLM.
We allow the user to writeplain English on what that data
transformation rule should beand they see what syntax is
(12:54):
auto-generated and then theyapprove or disapprove that.
One of the biggest things wasalways thinking that you had to
bring the representative of thedifferent areas to the table,
but instead it really should beto cater to the decision maker,
which is the business, and givethem good tools to make sure
that they can get through thatprocess really fast, right.
Take away the technicalchallenges.
(13:15):
Take away the fact that youdon't always know what's on the
other side, so let AI predictwhat's best on the other side.
Speaker 1 (13:24):
This is great.
This is the classic what wewould call a shift left approach
.
At the end of the day, as yousaid, the business units
anything that has to do withtechnology.
The only reason it exists is sothe business can do something,
make money, get more customers,sell more stuff.
The more you can move theactual doing towards where the
responsibility lies, ie thebusinesses a shift left approach
(13:48):
.
It's less expensive, moreproductive and faster, so all
three benefits.
A classic shift left approachwhere the businesses who own the
data, the business response,the P&L, so to speak, let them
be more productive.
In the old days, we couldn't dothat because they didn't have
(14:08):
the technical capability.
That's why we had to bring inarmies of people and it's why it
took so long.
So what I'm hearing and what issuper exciting about what your
company is doing is now you'reenabling, using English language
, large language models, aiapproaches to help these
businesses do what they need todo without engaging a ton of
(14:30):
other people.
Speaker 2 (14:31):
Yes, I think that was part of our secret sauce really
understanding the problem andmaking sure that we solve the
problem for the right person.
John, and one example that cameto mind this was one of our
customers was that the businessanalysts.
As I said, they ultimately madethe final decision as to this
(14:51):
is what I want it to look likeby the time it shows up in the
production system.
And that teammate was workingwith the DBA, who would write
the SQL query, to extract it andshow it to the business before
it got approved.
Well, the business user wouldsend an email or have a
conversation and say okay, theseare the fields I want and this
is what I want you to do tothose fields.
And that was written in plainEnglish over an email sent off
(15:12):
to the DBA.
The DBA picked it up, sawsomething, interpreted it and,
in this case, as probably halfthe time, interpreted it
somewhat incorrectly, did thequery put in the transformation
rule, sent it back to thebusiness the business doesn't
see the query, just the outputand said that's not what I want.
Can you do this now thatiteration goes back and forth
(15:33):
and back and forth because, as Isaid, data is ultimately the
truth teller.
Business is the decision maker,but then you just always had
that friction because theycouldn't get to the data
directly or they didn't know howto write the right query in
order to get the data to lookthe way they wanted to.
But this second piece that Iwould say is that Data
(15:54):
conversions are alwaysunpredictable.
With AI that's patternrecognition Unfortunately, we
can't go back and make it bepredictable.
I can't go back in time 20, 30years ago and change what was
inserted into the database wayback when, but what we can do is
at least have some tool thatanalyzes it really fast and
allows you to look at it andgives you some sense of this is
(16:16):
what I think it looks like, andthe business user can iterate,
and iteration is key.
So that was the other thingthat I think is really important
is that you just take a shot atit, you run it through, you
take a look at what you'vemapped, you take a look at what
you've converted and then yousay, does that look right?
And start with a small datasample set, because then you can
build rules on top of rules, asopposed to believing that
(16:38):
you've got everything you needand now you can accommodate for
all of the variations oftransformation rules that you
might need to put in there.
So just start with a first dataset, get it through very fast
and, like I said, we're talkingabout generating a load file in
minutes.
Look at that load file and thensay do you like that?
If it looks good, all right,let's continue to add more
(16:59):
upgrades in and continue to seeif your rule applies and if you
need to change it, then put morerules on top of that.
But I'm just saying it's thefact that you just don't know
what is in the data.
So give you a tool that allowsyou to get through iterations
really fast and allows you toplay around with getting it
perfect.
Speaker 1 (17:17):
That's tremendous advice.
We've covered really essentialground on modernizing and data
migration.
I heard 60% of all digitaltransformation initiatives fail.
They don't understand whatthey're getting into or the
level of effort.
What you've discussed is a hugebenefit for digital
(17:38):
transformation.
Wow, Thanks for sharing allthat.
Why don't we wrap this up?
I appreciate your insights andexperiences.
My last question is is thereanything I haven't asked you
that you'd like to share beforewe wrap the show up?
Speaker 2 (17:51):
When we think of data migration, john, a major aha
moment for me was we need toempower the business analysts
because, again, I believe thatit's the data.
And we need to empower thebusiness analysts because, again
, I believe that it's the dataand the decision belongs to the
business.
And I think all the yearsbefore one, it was that there
was always a need to havemultiple translation points.
(18:11):
Right, oh, business only knowsone side of the transformation
the conversion.
They knew their system but theydidn't know the target state
system, and then also, businesswasn't as technically adept as
some of the SMEs.
So let's unpack that and solvefor the business.
So I just think that that wasone aspect that I think is
really important because,compared to a lot of the other
(18:34):
products in the marketplacetoday, data mapping, data
integration that's a need.
In some cases that's an internaltalk, the data pipes and that's
a technology problem.
I'm trying to solve theonboarding, the new
implementation, the post-mergerintegration, where you want that
data to truly be right.
So that's one thing that Iwould say on the data conversion
, data migration side.
(18:54):
So this is a little bit of theproduct mindset I think
organizations and teams canimagine and design and design
and talk about it for a reallylong time and I say you'll learn
and you will get so much morevalue just by doing.
You might make mistakes becauseyou didn't plan for all of it,
but you're at least alreadystarted and not waiting to
(19:16):
believe that you'll discover allof it through planning.
Speaker 1 (19:20):
Great advice, I agree .
How do you eat the elephant?
One?
Speaker 2 (19:22):
bite at a time.
Speaker 1 (19:24):
Yes.
How can people learn more aboutyou and your services?
What are your socials?
What's the best way?
We'll include all of these inthe show notes.
Speaker 2 (19:31):
Thank you, john.
So Zengens' website is wwwaiand that's Zengens with a z or a
z in front of the word engines,plural and, by the way, our
domain.
We had AI from the verybeginning, so way before a lot
of these other companies startedto change to a ai on the domain
name and I'm on LinkedIn, sofolks can always connect with me
(19:56):
or send me a direct message andwe have information as to how
to reach out to our team, to geta demo and work with us, to go
through a POC or a trial.
In this world, everyone likesto take a look and see how it
works and we would be delightedto have people give it a shot.
I really believe, when it comesto data conversions, it doesn't
(20:16):
take an army.
I really believe everyone cando it themselves.
As I said, it's to the businessuser.
So if we start to change themindset and say, hey, this can
be self-service, that's what I'mhoping to move to, so that
everyone can change faster.
Speaker 1 (20:30):
Excellent.
This has been a great interview.
You're so energetic, soknowledgeable.
It's a pleasure to have you asa guest on the Disruptor
Everybody listening.
Please don't forget to connectwith Caitlin on LinkedIn.
Check out their website.
And, caitlin, before I wrap up,I always give the guest the
last word.
Then we'll say goodbye.
Speaker 2 (20:51):
Thanks, john.
Well, first I wanted to saythank you.
We had so much fun and I thinkI capitalized on the talk time
here, but your listeners shouldknow just what an interesting
and knowledgeable person you areand I had so much fun learning
from you in our conversations.
And then also, just to wrap, Isay that it's all about
gratitude.
I appreciate so much theopportunity to spend time with
(21:14):
you, that you've made time andspace for me, and that this
experience of building a companyto deliver value to people.
I believe data is solvable.
Let's help people focus onother things and let's make the
data be not something that theyhave to worry about.
Speaker 1 (21:30):
We're going to come back with part two of our
podcast and when we return we'regoing to shift gears and we're
going to explore the challengesthat many enterprises face on
mainframe modernization.
Stick with us and we'll be backAll righty everybody.
So thanks, caitlin, for beingon our show I'm John Kuntz and
(21:50):
thanks for joining this editionof the Disruptor Podcast.
Have a great day, take care.
Speaker 2 (21:56):
Thanks, John.
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.
Boysober
Have you ever wondered what life might be like if you stopped worrying about being wanted, and focused on understanding what you actually want? That was the question Hope Woodard asked herself after a string of situationships inspired her to take a break from sex and dating. She went "boysober," a personal concept that sparked a global movement among women looking to prioritize themselves over men. Now, Hope is looking to expand the ways we explore our relationship to relationships. Taking a bold, unfiltered look into modern love, romance, and self-discovery, Boysober will dive into messy stories about dating, sex, love, friendship, and breaking generational patterns—all with humor, vulnerability, and a fresh perspective.
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