Good Vibes x Domo.AI & Apps

Hi everyone. We're
so glad you're here.
I'm Jason from Domo, and this is good vibes.
Our conversation about data apps
and how you can use AI to go further and faster.
Turning ideas into apps that solve real business problems.
We're gonna be talking about the future of ai, talking
to some real product experts that I've worked with, with
for Forever about what they are doing
to help customers take their data
and use it to build purpose-built applications
to help them solve real problems.
But here to kick us off is RJ Tracy,
our Chief Revenue Officer, who's gonna show you
what we've been working on and give you a glimpse
into app catalysts.
I talk with customers every day,
and a couple of themes come up over and over again.
And one of those is, how does AI apply to me in my role,
and how do I put AI into different workflows
or work streams that I'm part of?
Another theme comes up, which is, I know what to do with ai.
I know what I want, but how do I take my vision
and put it into production?
And almost take in a very deliberate approach about
how we help our customers
to take AI throughout their companies by enabling them
to build and put into production solutions
that have a big impact on their business.
So introducing App Catalyst for App Catalyst.
It starts with a simple prompt.
I'm gonna go to chat GPT to help me for my prompt.
So I'm gonna go ahead and paste this prompt in.
I'm running a large retail location retail chain,
and so I'm putting in a prompt in for AI to help me
to manage all of my locations as well
as my e-commerce store chat.
GPT will build me a prompt, and again, this could be Claude
or or whatever else you want to use for your prompts,
and I'm gonna go ahead and copy this prompt.
As you can see, it's getting pretty detailed,
but the more specific you can be,
the more specific the solution will be.
And then I'm gonna move back over to App Catalyst
and I'm gonna paste it.
I may want to come over here
and select the data sets
that I actually already have connected up,
and that could be connected into, you know, snowflake,
Databricks, Google, wherever your data lives.
And if I select data sets, then the app is gonna be confined
to the data that lives within those data sets.
But in my case, I'm not sure what I should be doing
to manage my locations.
So I want Domo to be creative,
so I'm gonna not select data sets,
and I'm gonna go ahead and just click run.
Now this runs going to take, you know, 10 to 15 seconds
or so to complete.
And as it completes, it will now take all of the feed
that I put into the prompt
and it will build me an Agent Agentic solution.
And as you can see in a short amount of time,
app Catalyst has taken one of my ideas
and put it into production for Domo.
So I can come in here now and we're gonna go ahead
and explore this app together.
So it looks like at first I've got an overview section.
I can see that there's already some active alerts,
so it's alerting me of maybe inventory, stock outages, um,
supply chain delays, maybe unusual customer churn.
So maybe we're not seeing the right level
of loyalty customers coming back to our site
or into one of our locations.
And you can see it's gonna be some high level metrics.
And as I scroll down,
you'll also see it's given me some recommendations.
So it's got this AI orchestrator built in
and it's helping me to make decisions
for my retail location.
So you need to increase inventory for seasonal items.
Gives me the confidence score, what kind
of impact it thinks it's gonna have, um,
across either my location or all locations.
Uh, looks like maybe launch a flash sale.
So we could have underperforming SKUs
or it could be seasonal SKUs
and we're moving into a new season.
And then it's also looking at optimizing staffing, um,
for peak hours for high traffic stores.
So already some really cool things that are happening, ally.
So it's putting in some agents
that can help me to make better decisions.
It also has a demand forecasting tab.
And you can see in here it's looking at, uh, forecasting all
of my different locations and my stores showing me the
accuracy of the forecast, um, SKUs
that are maybe at high risk of us not hitting our forecast.
And then down here it looks like we can build also a
what if scenario for planning.
It's got inventory, inventory and supply chain.
So what's our inventory value, our turnover rate, all things
that that matter, um, in the retail sector.
And then looks also like I can see all
of the shipments that are arriving.
Are they on time? Are they delayed?
What's happening with my inventory?
Critical actions that need to happen so I can reorder.
And you can see these are links.
So I could click on and reorder this inventory from here.
You can see when I click on it, order's been placed,
obviously this app needs to be wired up,
but a lot of these solutions can be wired up in hours, um,
using the dommo platform.
Next is store operations.
And this is looking at staffing,
customer experience, et cetera.
So I can see foot traffic,
conversion, grade of that foot traffic.
So based on how many people come into my locations,
how many buy or how many hit my website
or go into a cart, what's the, the percentage
that checked out, staff utilization.
And then I can see peak hours at
different locations as well.
And then of course, how to optimize my staff on schedule so
that I can manage labor effectively.
And then it looks like e-commerce and digital.
So cart abandonment rate, mobile traffic,
website conversion, all my different funnels,
traffic sources, et cetera.
This is pretty impactful.
And in Domo, we're the only company that has all
of the technology available right here
to go put this into production in hours instead of months
to put something like this into production.
Alright, now let's go through another example.
I'm gonna go back to Chachi pt.
In this case, I'm gonna assume I work in a legal department
and I need help with reviews of red lines
to legal documents.
So I'm gonna go ahead, put my prompt in.
And Chachi PT is now going to write me a longer prompt
that we will have detailed information about
what the solution needs to have.
I'll go ahead and copy the prompt
and we will paste that directly into App Catalyst.
And in the second app you can see
that is now built out a solution to do legal reviews.
First screen here is the document upload screen
where I can come in and upload any of my documents
that have been redlined any of my contracts.
It's also given me some analytics about documents,
process changes detected average processing time.
If I come over here, it's given me a red line analysis,
so it's going to show me all of the different red lines
that have come across on this particular document,
and then how many of them are high, medium, or low.
And of course, recent changes that were detected
to the contract, the back
and forth between legal departments
and how many rounds we're on and what's resolved
and what's unresolved, et cetera.
You can see here we're also looking at clause reviews.
So clause by clause analysis shows every single clause.
It's also showing our proposed response.
So this is what it's recommending based on prior contracts
that we negotiated and a contract history intelligence
to show times that we've maybe negotiated these changes
and what we settled on.
And then it's got suggestions as well.
And in this case, it's actually given us not only a primary
suggestion for how we should respond to the terms,
but it's also giving us a couple
of fallback options as well.
And so I can come in here
and accept the changes right from the application
that will update the changes in the document
and then allow us to, to send
that back over for to your review.
Lastly, an executive summary to give me an idea of
what the risk score is, how many critical issues,
negotiations, and then key findings across the agreement.
And then you can also see here that this is showing a lot
of the different terms that were being negotiated across all
of our contracts so that we can update overall terms
if we need to make adjustments.
Now with both of these apps,
once I've got the app in this state,
I can go ahead and sign out.
Once I'm done, I can go ahead and publish this out.
I'm going to create a new application and I'll save this,
and this is gonna turn it into an application in Domo
that I can now disseminate out to to my team.
And of course, like all apps in Domo,
it's very easy to go ahead and edit.
And as I'm editing this app, I can even get down
to the pro code version where I have access to all
of the HGML, the JavaScript, the CSS.
And then of course we create a manifest where you can come
and wire up all of your data flows,
you can wire up your data sets, you can wire up workflows,
uh, so you have workflow automation built into this.
And then with our AI service layer, you can go ahead
and power that up with a LM, um,
no matter where it's hosted.
And if you're, if you have a team that's more technical
and that team wants to use other tools like Cursor, you
of course can start in Domo, take the code into cursor,
update it the way you want,
or alternatively, you can even start this directly in cursor
or cloud code
and you can bring that code directly over to Domo.
We also support things like React, so it's pretty flexible
so that you can get the application exactly how you want it
so gives you an ability
to take these apps from idea all the way
to production in a short amount of time.
The gap between idea and impact has never been smaller,
and we're excited to see how you use App Catalyst,
What an incredible demo app.
Catalyst is the way to show people how you can take data
and visualizations and dashboarding
and turn it into apps that help them do their jobs.
That's really what we're here to do,
and AI is what makes it all possible.
So that's why we brought Chris Willis,
chief Design Officer at Domo to talk about the future
of ai, where we're headed next.
Hi everyone, I'm Chris Willis,
chief Design Officer at Domo,
and I am excited to welcome to you, uh, a conversation
where AI is headed
and what it really means for teams building
modern data products.
Today I'm joined by Jim Fairweather, head
of AI's Go Do Market at Google Cloud.
Jim has spent years helping enterprises and ISVs design data
and AI strategies that scale,
and now he's at the center of
how organizations are taking AI from experimentation
to real business impact.
Welcome, Jim. Good to have you.
Thanks Chris. Yeah, thanks for having me.
Looking forward to this,
Jim, you've seen many teams start
with fast AI prototypes in early generative builds.
Um, where do these projects typically start
to break down once things like data quality governance
and day-to-day operations come into play?
Yeah, it's funny, we've seen this kind of evolution
of the POC or the pilot, right?
And I think the thing that a lot
of people learn in 2024 is there was this mad rush
to do something and
so everybody changed their domain name to.ai
and, uh, tried to overhaul the entire company all at once.
And, uh, I think what we saw is that failed often.
And so I, I think what we saw in 2025 was where people got,
uh, way more specific and they started small
and really focused and, um, they said no to a lot more.
And I think that's where we saw, uh, people gain a lot
of traction and just start with those simple, you know,
use cases where you can eliminate friction somewhere on
these like repeatable tasks that nobody likes doing.
They don't add a ton of value, but they have to be done.
And I think that's where we saw people kind of take aim, um,
and, and set their sights on,
on unleashing AI on some of those problems.
And where we started to see some, some ROI for folks.
So, you know, I I think to net it out, uh, you know,
when people started to big
and broad, uh, that's
where there was a lot of a lot of failure.
Oh, that's, that's good advice.
Um, AI makes software more personal
and much more context aware.
Yeah. How can organizations enable teams
to build locally without fragmenting data or logic
or governance across the business?
Yeah, you know, it's interesting.
I think something kind of funny happened over the,
over the holiday break is that everybody had a little bit
of downtime, um,
and got to play with all these new and cool tools.
You know, you think about lovable, you think about Repla,
you think about Cursor, uh, you know,
we have AI studio there, there's so many
of these different tools to kind of vibe code.
And I think what people are starting
to realize just this year is that, hey,
these things are actually pretty good.
And I, I think there's gonna be this next evolution
of personalized software.
Um, and that will challenge some of, uh, the SaaS companies
and, and what people are doing
because, you know, people are able to build kind
of their own experiences.
But, um, I think to, to make that stuff kind
of enterprise production ready, security, um, all of
that needs kind of a, a platform and a backbone.
So, um, you know, we've, you know, done
and come out with Gemini Enterprise, which we hope to be
that, uh, which it makes, you know, context aware AI
for you, personalized to you with all
of the different data streams and,
and applications that you use normally.
Um, but bringing them all into one.
And I know, you know, our companies have worked together on,
on doing that and trying to make more available, uh, to our,
our joint end users and joint end customers, uh,
to make life easier and more helpful.
And I think that's a really exciting phase
of this is like this next new surface of agents, um,
can really change kind of the, the customer experience.
Uh, we provide people together.
Yeah, it really feels like a,
a big shift in software is coming, uh,
the paradigm is is moving, it's moving very, um, speaking.
Yeah. Speaking of, uh, changing paradigms, um, you know,
it seems like multi-mobile AI is becoming table stakes, uh,
you know, moving just beyond, you know, strict language.
So what technical teams, uh,
what did technical teams most underestimate about making
those experiences reliable and scalable
and also maintainable, uh,
when they get put into the real world?
Yeah, that's a good one. You know, I, I think about this
and I, I always think about my son in this
and this, this may be speak more
to my parenting than it does anything else.
Uh, but you know, my son doesn't type, you know,
he is 10 years old, he doesn't type,
and yes, we need to work on that,
but, uh, he talks to his devices
and I think that's just become, uh, a very normal thing for,
for kind of these next generations coming up.
And so I think in 2026, we're gonna see a ton of audio.
Um, and you know, the, the thing when technical teams try,
try to take that stuff in, I mean, you have to be more aware
of the context window, and so how many tokens you can kind
of allow, because can you do the, the three hour podcast
and just put it in and start to work on it?
Or do you have to parse it out?
And so I think as people take on these new mediums of,
of audio and video
and start to infuse that into what they do
to get out better answers, I mean, it opens up a ton
of, of capabilities.
Um, and that's one thing we, we love about Gemini is this,
this monster context window where you can throw a bunch
of stuff in and get insights out of.
But, um, you know, I,
I do think it's gonna become table stakes.
I think the consumer experiences,
and you, you look at open ai, you look at Anthropic,
you look at us and all these other different, you know,
you look at Suno with music
and it's like, um, the table stakes or,
or the expectations from the consumer are now so high
and enterprise needs to start to meet that.
And, and a lot of these startups are providing these really
seamless, frictionless experiences that,
that have multimodal into it.
And, and I think it should be,
that should be how software evolves.
I do want to be able to just talk instead of type
or instead of like drag and click around.
Um, I, I think that it can just be more effective,
um, in how we're using this.
I think about Shopify in this example to
where they've used our, our Gemini live API to
where now when you set up your Shopify account,
you can share your screen and Gemini can see it,
and you can talk to it, and it's like, no click,
don't click the red button, click
the green button over here.
And part of that also for the technical folks, gives them
insights into how people are using, um, their user interface
and maybe things that they need to adjust or change.
So, I don't know, I I think it's gonna be really fun.
Um, and we're all gonna change a little bit of the way we,
we interact with software, which I think is a good thing.
Yeah, I was just thinking as you were speaking about that,
that, you know, as a designer, I kind of feel that, um,
if I'm really honest, all the work I've mostly done has been
a compromise because it's usually meant one kind of sort
of visual interface for everybody.
It has to work for everybody, right,
right. But that's shifting.
That's shifting. Yeah. And we, we can provide options.
And I, and I, and I do think that's the thing you can have,
you can have defaults
and you know, how you want this to be consumed,
but I think we, we do need to provide the end users options,
uh, for how they want to consume software.
Yeah, that's fantastic.
Uh, speaking of that, you know, when you're,
when you're sort of creating more powerful tools for more,
um, creative and, and maybe more exploratory experiences,
and you're putting tools with say, no-code
or low-code, in the hands of analysts and ops teams
and others, uh, what guardrails do you see as essential
to accelerating insights without harming data quality
compliance or, or model health?
I guess what we're really talking about here is trust, is,
is how do you think about
that in the systems that we're building?
Yeah, look, trust is, trust is the foundation
to all of this stuff.
Without trust, you can't use any of it.
And, and there's so many cool,
like flashy things that people can do.
And here I'm talking to my data and, and all of this,
but like, you, trust is found fundamental to that.
Um, and I, I know you share this ethos,
but it's like you, you need to balance this like boldness
with responsibility.
And so we want to be bold in the things we can try
and use and do.
Uh, but you, you have to protect the data.
You have to protect the company.
You have to protect the person, uh,
and the individual who is gonna be using these tools.
And so I think that, um, that's gonna be a,
a really critical piece for
how people should evaluate the partners they work
with in this next era of AI and agents.
And, you know, we, we think about, um, you know,
how these different companies talk to each other
and share data, but I, you know,
I think it's incumbent on everybody to, to think about that,
um, yes, the IT teams and to be able to have guardrails
and to make sure that only the people who can
and need to access the data, uh, you know, have permissions.
Um, but also, you know, everybody needs
to personally think about, you know,
how they're using data in a responsible way, uh,
to make sure that they're not, you know, sharing too much
or, you know, putting the wrong data into a place, uh,
where it shouldn't be or, or make it publicly available.
And I, you know, we see this too much, I would say, and,
and it's, this happens with any new technology, is
that you have this kind of like shadow IT
or AI to where, hey, the company hasn't approved
that new latest AI tool or function,
but I use it at home.
And so maybe I can just use my personal account
to take this, you know, company data, put it over here,
get the insights, and then put it back,
which I think we all know is a no-no.
Uh, but I, I don't think anybody's surprised
to see that that happens.
And so making sure the security teams are set up to where,
no, that, that data can't go from here to there.
Um, making sure companies are aware,
but also I think companies really need to look at, you know,
when you think about attracting, uh, retaining, growing,
developing top talent, you have to give 'em top tools
and people aren't gonna, when they can do their job in a
10th of the time using ai, um,
let's not make 'em do it the old way.
Um, and so I think that we've seen
that this like rapid adoption, uh, you know, for us, like
with Gemini Enterprise
and things like that, that are, that are safe, secure tools,
uh, that the folks, you know,
folks in the enterprise can use.
But, um, you're right. Like it all comes back to, to trust.
Yeah, you, you brought up a lot of points in there,
and I, I think one is, um, and,
and we've seen this a lot too, which is, uh, you know,
convenience is usually something
that wins out for a lot of people.
So when you're talking about, you know, security
for example, that's always a, a battle,
which is like, well, it's easy.
I can just sort of put it here. And it,
it doesn't seem like there's gonna be much of a,
of a consequence to it.
But on top of that, you know, what you're using
and how you're sort of creating, um, you know,
these insights, I think people have to take a,
a more active stance
and be more, more critical in their thinking about,
you know, what's being generated.
So bringing their judgment into the process in maybe a way
that they haven't had to do before.
Um, but yeah, that, that, that's fantastic insights.
Um, you know, sort of speaking about, uh, you know, the,
the big elephant in the room for many organizations is,
is the ROI on ai, right?
A lot of people jumped right in
and they're like, Hey, look, we're using all the tools.
Um, we're the big cost savings.
Uh, so, so with that, from what you've seen, what,
what is the clearest signal that an organization has turned
AI into a durable competitive advantage rather than,
you know, just a set of isolated experiments?
Yeah, look, this is a good one
and it's a tricky one, um, Because it's like, you know,
I think you have to ask the question first of like,
how do you define ROI?
Is that time? Is that money, is that, is that, you know,
just efficiency of people of headcount?
Um, so I think there's a zillion
different ways to measure it.
You know, we did this, um, an ROI
of AI study in 2025.
Uh, I'm sure we'll be coming out
with our 2026 study before too long.
But, you know, we surveyed 3,400, uh, business leaders,
uh, across the world.
And not just using Google tools, but, but all AI tools.
Um, and so one of the things that was interesting
that we found was that 88% of the folks who were kind
of early adopters, um, saw some sort
of ROI from what they were building with agents.
And so, you know, again, people define that in,
in different ways, but I think the, the thing to learn from
that is like, fortune favors the bold.
And so I think the a big you we're not gonna get
any ROI if you do nothing.
Um, so waiting for that perfect use case
or the massive use case, kind
of the point we were talking about earlier, I think the,
the biggest thing is to try and just start.
And that's where you'll start to see the ROI,
and I think you'll learn from that.
And that's, we've seen a bunch of companies, you know, and,
and clients that we've worked with learn from starting small
again, reducing some friction, um,
that you can give people time back.
And I, and I think the thing is like,
as you think about a company's competitive advantage,
is like, how do I translate the AI
that I have infused into my product or service?
And what's the best way I can quantify that?
And I think those are the folks that are gonna be able
to win, certainly in this like, new agent economy.
'cause it's like, what do I even charge for my new agent?
Do I charge anything at all?
Um, you know, what's the, people are still figuring out kind
of the, the cost on that, you know,
to, to the different companies.
And so, you know,
I think when you think about making it a competitive
advantage, I think it comes down to like,
how does your end user define value?
Um, and then trying to optimize for that.
'cause there's a lot of different ways to do that.
There's gonna be, 2026 will be kind
of the year of measuring that.
Um, and there's a lot of different pieces
and parts in, in order to how to do that.
And, and I'm not trying to like skirt the answer.
I, I think there's a lot of different ways to do it.
Um, and rightfully so.
I think they'll be scrutinized this year.
Um, but I guess the encouragement I would say is like,
just don't, don't pick one metric, uh, for all.
I think you need to look at multiple metrics across multiple
things, and, you know, um, how do you kind of define, uh,
the metrics that matter
for your organization and your end users?
Uh, and start there. Uh,
because just your ROI might not be the same as mine.
It's interesting. It feels like, uh, for many, uh,
people feel like they're behind,
but in many ways, we're just starting, right?
I mean, we're learning on the job in many ways,
and I'm sure there's, uh, I think to your point, um,
the companies that are addressing this, uh,
and jumping in on it sooner rather than later,
we will, we'll stay ahead.
The only way you're gonna be able to do
this is, is by learning.
So I think that's, that's a, that's a great point.
Um, Jim, I really appreciate the conversation
and your time and your insights.
Uh, this was super helpful.
As a final thought, and feel free to add in any, uh,
lessons you've learned as a budding triathlete.
If you had to give, um, our audience one piece of advice,
uh, to think about AI in their
organizations, what would you tell them?
Yeah, but I just allow your budding triathlete.
I'm trying, man. It's been a, it's been busy.
Um, but look, I, I think my main advice, and I,
and I, I take this from, uh, from a group,
this conscious leadership group.
Um, but I, I talk about, you know, they talk about being
above the line and, you know, you understand
where you are mentally, whether you're
above the line or below the line.
Above the line is open, curious, committed to learning.
Uh, below the vine is defensive, uh, closed,
committed to being, right.
And so I think throughout our day, we're gonna go up
and down in this, but I think as it comes to approaching ai,
it's, it's really what you just said, Chris, it's like
taking a personal,
but also a corporate mindset to be above the line
and say, I'm gonna be open, I'm gonna be curious,
and I'm gonna be committed to learning.
I think that's the way we can all get the
most out of this moment.
Um, but for the folks who are gonna be closed, defensive,
committed to being right, you're
gonna be stuck standing still.
There's so much to learn.
I, there's, you know, it's kind of like anything.
It's like the more you learn,
the more you realize you don't know.
That's, that's what we're in, in this moment.
And so I think put all the ego to the side, uh, and just,
and just get curious, uh,
get open, get committed to learning.
I think that's how people are gonna, uh,
really thrive in this environment.
At least that's what I've seen so far.
I love that. Stay above the line.
All right, well, you heard it here.
Thanks to Jim Fairweather, head of AI's, uh, go
to market at Google Cloud.
Uh, Jim, thanks for your time today.
Thank you, Chris. It's time for my favorite part.
You're excited about App Catalyst,
but you're not quite sure where to get started.
That's why I've invited three of my favorite people to talk
to us about how to take your apps
and conceptualize your ideas, build them using ai,
and then deploy them into Domo using code.
These people have been building apps for years at Domo since
before it was cool and before AI made it easy.
So if you've got questions, drop 'em in the chat.
I'll have there answering live.
But to get started, let's welcome Terry,
our senior UX designer over the app team to get us started.
So Terry, welcome to the launch. Thank you.
Happy to be here. You've designed hundreds of apps
for customers and I know that you have learned
so much during that time.
What is different about your process from the early days
of working with those first few customers
to what you're doing now?
Every en engagement that we've had with the,
with our customers is they all start with an idea, right?
It's some sort of a department within, within their company
that wants to improve their process, right?
They have that idea and that's great. And
that's why they bring us out there.
And in that process we go and talk about pain points
and what struggles they deal with currently day to day.
We are so strategic in, in the way we communicate with them,
and we, where we talk about the whole process
and we get those components right out the gate.
And I feel like because of that and
because of the team that's kind of working
with me on getting those requirements, we're able
to get a really nice solid foundation
to build the app right the first time
and not have to come up with their solution
with the enhancements and improvements. So
Tell us about some of the pitfalls
of if people aren't doing ideation at the beginning, right?
What are some of the
challenges they're gonna face downstream? I thought,
I think about this a lot.
Say if you, if you're like, oh yeah, I wanna go into
any type of a AI
and I wanna say, I want you to design me a house
and it's going to build a house,
that's gonna look amazing from the outside.
But I think what's missing with that is you're not prompting
the AI to think about the foundation.
You're not thinking about the plumbing,
you're not thinking about the electrical,
you're not thinking about all the HVAC
and everything that's gonna go into making sure you're
comfortable inside that house, right?
And so by doing all the steps beforehand
and then asking and,
and trying to prompt engineer the best possible solution,
you're not gonna get that blueprint
that you're, that you're looking for.
Yeah, you might get something that looks cool
and that will be a, that'll be a talking point.
But it's, it's a lot different when you say,
I want you to build this app for me.
And it's missing all those components.
And that's why I feel like the app ideation process is
what's going to really help your, your app happen.
So when we're talking to a customer,
you're coming in from the outside, you're trying
to understand their business
and what are the things that they should be doing?
'cause they're not gonna have you necessarily in the room
with them asking these questions
so they know more about their business.
So how can they start capturing some of those data points
that they're then going to use to inform the AI
to help them build a solution?
Just as I have a great team that goes
and helps with drilling into what exactly we need to do
with the requirements, I think they need to also think
outside, you know, their head
and start talking to other people and what's,
and what ways could we make this better that
ultimately will help with prompting.
So show us a little bit about what it is that you do.
Some of the tools or the frameworks that you have,
they've assembled the team,
they're asking the right questions,
how do they kinda document this so
that they can get it into ai?
You, you break all those, those points
that we talked about a little earlier, such as you know,
what your pain points are and the solution, uh, components.
And then you also looking at the, uh, current state
of your process and then the future state.
And I, I put those as a, as a, as a column header, each one
of those items as a column header.
And then basically, as much as we need to do, uh,
vertically, I'm running answering each question of
as a group and as a co uh, collaborative effort,
we're answering those questions together.
And ultimately we're gonna build a, a giant list of,
of takeaways in that future state column.
So every app that you build, regardless of its size,
starts with those same four questions of
what are we trying to solve?
What are the components of the app that make that solution?
What are people doing today?
And then what do we want people to be doing,
or how do we wanna solve that problem?
And then just replicating that for all of the opportunities,
challenges, and pieces of the app.
And that's what you use to create all
of your amazing designs and prototypes and,
and really solve those customer problems.
Yeah, like, so up to even six months ago,
I wasn't following this same model,
but the last two solutions
or strategic solutions sprints that I went on,
I followed this model and I feel like we were able
to hit the mark more closely with our customers,
both, both of those times.
If I follow that model, it works.
Hmm. I think everyone, uh,
listening today has this opportunity to learn from someone
who has just spent the last six years trying
to figure out the process to go through on this.
And so lucky to say, okay, if we do this,
you're gonna get better apps.
You're gonna solve your problems faster.
You are not gonna skip the ideation phase
and build the wrong thing.
Um, that's really powerful.
Any other suggestions, final takeaways, things
that you recommend to everyone on the call?
No, I think once you, once you have that,
just once you have the discovery laid out, use that LLM
and go in and say, I'm looking
to build an app that does this.
But then you also say, here's what my current situation is
and here's how I want to improve my current workflow.
And I think that type of an equation, if you think of it
as an equation, this plus this,
or you think about the pain points plus the
solution components, you know, plus the current state
plus the, the future state.
As you think about that in your prompting
and you think about your prompt to your LLM,
that will ultimately leverage all the information,
all the findings that you found in those hours
and hours of discussions,
and then that will be what's delivered in the ai.
Yeah, you're still gonna need to pivot,
you're still gonna need to shift, right?
That's all. It's, you can't build it in one pass,
it's never gonna happen that way. You're gonna
Learn things, You have to think about the details
that are, that are inside the app
and what's going to make that person connect with your app.
Very cool. Yeah.
Terry, thank you so much.
Really excited to see the apps that people build, uh,
using this framework and where you all go next.
And if you want to talk to Terry
and meet with his team, always available doing solution
sprints for customers and, uh, pushing the envelope further.
So thank you Terry for being with us.
Thank you for having me. If you remember nothing else,
remember that AI like business is a team sport.
If you get the right people together early, go through
that ideation process, you're gonna get better results.
So now that you've done your ideation
and you've planned out your app, now you've gotta build it
and make changes and ideate using AI and App Catalyst.
Let's talk about it.
Flo, welcome to the launch. Thank
You so much, Flo.
One of the things I love about your work is you sit right in
between the customers and what they're trying to do with AI
and data science and the actual dev teams
that are building these capabilities so
that customers can use 'em.
And so you've got a really unique place
that you sit right in between these two,
which is perfect today for App Catalyst
and the launch we're talking about.
Yeah, it really is an amazing place to be.
We, we get to really interact with customers
and find out what the, what the pain points are,
where we have opportunities for growth,
and then we are able to really quickly hand that off
to the development team and, and start working their magic.
So I wanna start off talking a little bit about
App Catalyst, how it works from your perspective
as a data scientist and also as a data person that has a lot
of data that you're trying to talk about
and share with customers
and finding the best ways to do that work.
Uh, so tell us a little bit about your usage
and what you've seen as we've gone throughout this process.
Yeah, so, so you all just heard about, uh,
UX design principles right from Terry And Terry can be
so creative when it comes to dashboarding and app creation.
And I, I am not that person.
I am only creative in the space of data engineering
and statistical modeling.
And so when it comes to having to create these apps for
so many different customers
and the apps are generally pretty similar,
they follow the same configuration,
there's the same purpose.
And so having
to build these apps from scratch every single time can
get really time consuming.
So App Catalyst can, can really save me some time
once I have figured out exactly what it is
that I wanted to produce.
So I know you brought some things
you wanted to share with us today.
Tell us a little bit about the experience
and where you've seen this evolution
and what's cool now coming into App Catalyst
as we've been working on it for a while.
What's the newest, what's the latest and greatest? Yeah,
Yeah. So the development
team has, uh, created this, this tool
that has really been fine tuned over the last few weeks and,
and months to really match the needs of, um, a variety
of people including data people like me.
So if I think about all the solutions that we build
for our customers, whether they're data science,
machine learning, ai,
there's always a model performance component involved.
We always have to evaluate how the solution is doing.
And so for every single solution, regardless
of the use case, we have have
to put together a model performance app or dashboard,
and that requires possibly 10 20 data sets
of inputs to put together.
And it could be really time consuming to create all
of these components.
It, it can, it can consume a lot of time for me.
And so with App Catalyst, I can tell it exactly
what I want with the data sets that are relevant for
that customer and I can just kind of rinse
and repeat this template that would work at least 95%
of the way for a majority of use cases.
So maybe like a traditional prompt
where you're just uploading a spreadsheet
or taking a small sample of the data
and saying build something around this, you're actually able
to give it multiple tables coming from your source systems
and use that to power the
Apps. Yeah, that is exactly
it
and I can tell it, I want this visualization to be powered
by this particular data set
and this is how you should use the columns
and I want this other visualization next to it
to be powered by something different.
Cool. It's a very, very powerful
Tool. So this may be
as close as we get to like machines,
building machines, uh, as we have, right,
where you have AI data being visualized by an AI system,
uh, and distributed in that way.
Tell me a little bit about the technology here
and um, you know, when we saw the demo earlier with rj,
what are some of the things that customers should be looking
at as why this is so cool and,
and what they're able to do with App Catalyst?
Yeah, one of my favorite things about App Catalyst is
that it really works with what you give it.
So I'm gonna walk you
through an example in which I am providing a very structured
template because it's what I have built time
and time again, so I know exactly what I want,
but it can also handle an interaction in which you just ask
it to create an app based on these data sets.
And you may not have, you know,
a specific configuration in mind.
You may not have a specific narrative
or story that you wanna tell with it, you just wanna see
what the AI comes up with and it can handle
that very well as well.
So show us a little bit what you've got here. Yeah.
And let's take a look at,
Alright, so one of the most common use cases
that we do is multivariate forecasting.
And it's a, it's a very common,
generally pretty straightforward,
but it does require a lot of data sets to,
to think about model performance
because we want to think about the model performance at the
time in which we're training a model
and we're deciding which model to deploy.
And then once we've chosen the model, we need
to track model performance
and model decay over time to see if the model, you know,
in six months is not doing well.
So already there's possibly 1520 data sets
that we're gonna be leveraging for this.
And this is all you as a data person,
this is the problem you have is having so much data
how I communicate about it Yeah.
To the consumers who maybe don't understand all
of the data science Exactly. That's going into it. Okay.
And, and because we've built so many of these,
we've already come up with a recipe of what needs
to be communicated for something like model performance
for a non-data science user.
And so we already know what the narrative should be,
and that's the prompt that we've put together.
If I didn't know what I wanted, um, the app to look like,
I could just ask AI to do it
and I could make all the adjustments
that I wanted As you engage with the AI in natural language.
So based on the model performance deliverables that we like
to give customers, I've created this almost annoyingly
detailed prompt that specifies each visualization, the title
and description, its purpose and the data set it should use.
I'm creating one page for model selection metrics
and another one for life model performance evaluation.
So this automatically is creating two different narratives
that I need the customer to have access to.
And you'll see the app Catalyst handles these instructions
seamlessly creating a custom pro code app based
on the prompt that I just provided.
You can see how the app includes the visualizations I
described, and
for other use cases I may even need more than 20 data sets.
Now that's a pretty advanced dashboard,
but there are times in which I, as a data scientist need
to stitch together all these different components
because that's just the way
that the data comes out in Jupyter, right?
And what's also great about the product
that our team built is that I have the ability to edit
after I have already created my app, right?
So I can continue to engage via natural language
with revisions because I'm sure, I don't know,
I could ask Terry, but I'm sure Terry still has
to make revisions to his work for sure.
Um, maybe I just don't like the color scheme.
I can give it exact hex codes
and then it will adjust my app to match my request.
Or maybe I changed my mind
and I don't want squares in my line graph
and I want little dots.
I can adjust that really easily as well.
So it's, it's pretty cool what it can do. See. Yeah,
I love that you have so many times creating this,
like why build it one more time when AI can do that for me?
Yeah. Uh, again, you've done all of the work
to build the prompt and understand and,
and really get the data into this state.
And I think so many of our customers out there,
that's exactly how they are.
They've got this data, they've found something interesting,
and now they're trying to get that into a, a, a distribution
art artifact that people can consume
and say, oh, I understand this information
and they're maybe not visually inclined
or, uh, they don't have the time to go
and build, you know, all the different permutations
of this pages or, uh, things like that.
And so this gives them that, uh, escape patch
or this like cheat code to go
and build something really cool
and then tweak it, make those modifications,
really make it personal to them,
not just whatever the AI's first guess is.
Yeah. And it can be like an iterative process, right?
You can share the current version of the app with someone
and get feedback and then go in
and still edit, um, through this natural language.
So there's still this opportunity for feedback, um,
and continual improvement on the app.
And I know we talked a little bit about this.
How do we, how do we get the AI to be more performant
and, and give better apps?
Yeah, and I love that Terry talked about that
because it's something that we have to do
as data science consultants as well.
Customers always think
that they've given us all the information we need
to build a solution that they want,
and we come back with a million questions every single time
because we don't understand a concept
or be we're seeing a discrepancy.
And that is the type of feedback loop that we've built
into App Catalyst as well.
And so App Catalyst has the ability to elicit information,
which is really just the process of AI asking
for information that wasn't given.
We also have confirmation built in.
So, you know, sometimes AI just needs to make sure
that they understand something correctly.
And then there's this idea of planning
that was also built into it.
And that is just, you know, if you give AI a broad request
of build me an app using these three data sets,
it will come back with a proposal of what it's gonna do
and it will tell you, this is my approach
to the visualizations.
This is what I plan on doing
with the configuration or the structure.
This is what the narrative is gonna be like.
And you can give feedback on that planning
before it even creates the app.
And so it really gives you this opportunity to fine tune
the process that AI is going to follow
before it even creates the app.
Very cool. So what else do you have
to show us? Anything else that you wanna show us?
I would actually love to show the elicitation
because I, so for modern performance dashboards,
I know exactly what I want, but there are
occasions in which I have to create a dashboard for, uh,
a concept that I haven't done before.
What if you don't know exactly what you want to build
and you want to explore the possibilities?
Um, and if you've used generative AI
before, you've probably been asked questions from ai, right?
Like, uh, a few months ago I was planning a trip
to Costa Rica and I asked Chad GPT to think about the fact
that I was going with my, with my child.
And the first question that AI asked
was, how old is your child?
And I was like, how did I not think
to give that context, right?
It matters if my child's three
or if they're 10, makes a difference.
It makes a huge difference For this schedule,
I had given all this other information, I had asked it to,
you know, plan two activities a day
and minimize like the driving time during these days
and I missed a really important fact.
So AI and App Catalyst will also, um,
provide this elicitation experience.
So here I'm asking for a generic app
that uses those same data sets,
but it's less granular instructions.
And you can see how AI is giving me back a proposal
that is broken out into sections that outline how it plans
to tackle my request.
And if I don't like something, I can request revisions
until the plan feels right
and once it's finalized, AI executes the strategy
and produces the app that I requested.
So just the coolest thing that they built.
Very cool. Yeah. Flo, what can we look forward to?
Is this done? What's next on the roadmap?
I know you can't give away too much.
We've got Domopalooza right around the corner.
What are you excited to see? Yeah,
I'm excited to see how customers use App Catalyst
and the feedback that they provide.
And, uh, for everyone who's on the call, if you have ideas
about, or on the podcast, I guess, um, if you have ideas of
what you want to incorporate into App Catalyst, you know,
added to the comments because we have the people
who are developing this product on the backend here,
looking at the questions and ready to answer
anything you have, um, for them.
So just, just tell us what your ideas are
and where you, how you plan to use this. Cool.
Yeah. Awesome. Well, thank you Flow.
It's been awesome to have you.
Excited to see where this goes and, uh, wow.
Where you take App Catalyst next.
Maybe give us one more thought would be, what would you say
to people who have not tried apps inside of Domo yet?
You are missing out. Wow.
If you remember nothing else,
remember the app Catalyst is here to help experts be experts
and not have to redo all of that work
that's already been done somewhere else.
Now you've built your app, you've iterated through it,
now it's time to deploy it and get it out to people.
So let's talk about that.
So Emmy, welcome to the launch.
Yeah, I'm happy to be here.
Thank you for coming. I know that one of the things
that's really cool about the work that you do
is you are using AI and building apps every day.
Yes. Trying out new things, talking to customers.
So tell us a little bit about your experiences
and why you as a developer choose to build apps on Domo, uh,
when you have the skills
and abilities to build them using any tech stack.
Yeah, so, um, I'll start just by saying that, you know,
all of your data lives in Domo obviously,
and then you have a lot of features in Domo that you love
to use, um, workflows
and using different kinds of charts and analyzer
and being able to create cards and dashboards is great.
But then if you wanna go a step further
and have something that's more interactive, it's hard to do
that unless you create an app of some sort.
And Domo makes it really easy
to create all different kinds of apps.
Um, I've been working in the custom app space the whole time
that I've been here, so just writing code
and doing everything completely custom.
And, um, I love to, to write all of my code outside
of Domo and then push it up.
That work is what works really well for me.
But Domo has different ways to get apps running in Domo.
Um, hopefully a lot of you guys are already familiar
with App Studio that's been around for a while
and that's great for getting different kinds
of visualizations in the same place.
Um, we also have what's called a Pro code editor,
which is just a in, in IDE that's inside
of Domo where you can write code.
Um, and it has a really nice integration with a manifest
file GUI in there.
And I'll talk a little bit more about Manifest file later
'cause that's really important, but it makes it easy for you
to connect to your app, to, to data into app DB
collections if you need to save data.
Um, but then once you're done in the Proco editor,
it's really easy just to hit publish
and then to have that app
available throughout your instance.
Um, and then of course we have
what we're talking about today, app Catalyst, which is a way
to integrate AI into the experience
and have AI help you actually write the code.
Um, perfect way to get started if you're unfamiliar
with building apps or writing code yourself.
So that's kind of integrated with the Pro Code editor
and brings everything full circle.
Awesome. I mean, tell me the people we have today, not all
of them are gonna be developers, um,
but we're giving them this ability to write their own code
and, and explore this world of apps
and purpose-built solutions, uh, to turn data into action.
Well, what are some of the benefits
that you get from building apps inside of Domo?
Yeah, so it just makes it really easy to take advantage of
what you have inside of Demo already.
So like I said before, you have your data sets in there.
Um, Domo's app platform has a lot
of really great features built in
and a really rich API for connecting to your data sets.
Um, saving data to a database
and running features that exist,
like workflows and other things.
Um, so if you use the app platform, um, you can connect
to all of that really seamlessly with our APIs.
Um, and then also as part of that, you have a lot
of the permissioning and controls and,
and user authentication that's part
of Domo baked in as well.
So you don't have to worry about a lot of that.
And if you're, if you're building an app by yourself outside
of Domo, those are some
of the hardest parts of getting it working.
You know, it's not writing the code.
Um, if you know anything about AI generated code nowadays
or vibe coding, a lot of people are able
to create some pretty cool things.
But then once it comes time to put into production
and have users actually use it,
you run into authentication issues or permissioning
or security issues.
And when you build on the app platform, a lot
of those are already taken care of.
That's amazing. I like to joke sometimes with people that,
uh, the difference at Domo is I can build an app
and I can deploy it to our CEO
or to our CMOs phone right now.
No waiting, no code pushes, right?
It's, it's just available today for them
and that usually gets some heads to turn.
Uh, what makes Domo unique about our methods
for deploying apps and why people would do that?
Yeah, well you create, um,
like a reusable asset on the instance,
and then once you have that you can connect it
to any number of cards.
So you could have a card
that is on a dashboard on your CEO's phone, for example,
and maybe that's just his version.
Um, and then you could have a card
that's on an app studio page that is using the same asset
or app underneath the hood,
but is actually connected to different data
or is using a different database.
And so it makes it more personalized to the users
that it's been shared with.
Um, so if we talk a little bit more about the manifest file
that I mentioned earlier, that is the configuration file
within your apps that determine what data sets it has access
to, to query from, um,
and also what app DB collections, if any,
that it should be connected to as well.
And that's where, um,
if you need like editing within the app so
that data is saved somewhere, that's where it would go.
And you can configure that.
And then along with that you can configure the permissions
for the app DB collections and other things or,
or the scope of who can save things, who can view them,
who can delete or, or update things.
And, and being able to have
that fine grain control is really important
for a lot of our customers.
What are some of the, the challenges
or the things that people should consider?
Uh, you mentioned the manifest file earlier, like explain
to some of our technical users, right?
What are some of the things that they should be looking out
for where again, like all of ai, there are benefits
and then there are also challenges that come
with any new technology or any opportunities.
So what are the things that you get these awesome advantages
if they just take care
and make a couple of those things
set up at, at the beginning?
Yeah, I, I would say
that having a plan ahead of time is critical.
Um, I'm, I know you talked
to Terry about it a little bit earlier,
but actually knowing like what the app is going to do
and who you want it to be shared with, um, where it's going
to live, knowing all of that ahead of time is, is crucial to
how you actually build it.
Um, I would also say that when it comes
to AI generated code, um,
it can generate anything, right?
Like it's pretty much limitless,
but the more that you're able to guide it
and instruct it on what you want, the, the better it's going
to perform or it's gonna give you stuff
that you actually can use.
So what we do, um,
on our team is we'll have maybe like a pre-configured set
of instructions that we'll give
to whatever AI tool that we're using.
So it has the context that it needs to generate apps that we
like working with, right?
For lack of a better way to say it.
Um, you know, if you go to Lovable for example,
and you have it generated an app, um, it might use Tailwind,
which is a CSS library that you may
or may not know, um, maybe you instruct it
to use React, maybe you don't.
Just vanilla, JavaScript
or Angular, any other, other, um,
front end framework that you want.
Um, or you could have it use specific
libraries for charting.
A lot of people wanna build apps with charts in them,
and there's so many different charting
libraries out there that you can use.
If you don't specify, then
AI is just gonna pick one at random.
Okay. Um, and just like syntax
and how it decides to write the code and,
and how well it's commented
or how it's formatted, if you don't specify anything,
then it's just gonna do what it wants.
And so if your team has a specific way that you like code
to be written or you have certain libraries,
or you have like a certain branding for your company, like
that's important information
that AI should know ahead of time.
And the more that you give it,
the the more likely it's gonna give you something you like.
So talk to me about the advice that we wanna give
to our non-technical users who are going
to be generating apps, but then
need some technical expertise to get it across the line
or to make those final polish items or, or close it off.
What should the, what are the questions they should be
asking to their technical teams to make sure
that the apps they're generating are ones
that their team want to come in and interact with and,
and we kind of remove that whole vibe?
Coding stigma? Mm-hmm.
Yeah, that's a a great question.
I mean, first of all, you can ask them, you know,
what libraries, frameworks cetera, do you like to use,
you know, based on everything that I was just saying.
Um, also it's, it's not a bad idea
to just have a conversation with AI ahead of time
and like kind of scaffold the project
and be like, this is what I want it to do, this is
what I want it to look like.
Um, these are the kinds of interactions
that will exist once I'm done figuring it out.
But kind of create a plan with AI ahead of time
and then you can even ask AI to give you the prompt back
that you can put into App Catalyst to generate it so
that you don't have to do that yourself.
A lot of the time, even just doing that will generate
a better first version of the app.
Yeah. So
how does using App Catalyst make the transition from ideas
to apps more seamless and quicker?
Yeah, I mean, so when you start, you know,
you give App Catalyst an idea, it's gonna respond back
to you in a similar way that Chachi BT does,
where it might ask another question
to get more clarification on something,
or it could suggest, Hey,
maybe you should connect a data set to this
because it's seeing what you're trying to do.
It might make that connection on its own.
Um, of course our developers are constantly improving it,
so there'll be more futures to come,
but there are some things that it's already really good at.
I know we showed some demos earlier,
but, um, you can take advantage of those.
You don't have to, you know, it's all optional,
but it's great to have the ability to do it when,
especially when you're not technical.
Um, it'll just help you do it on your own.
And then once you're in a state where you feel like you like
how it looks and it's functioning well,
it makes it really easy to just save it, publish it,
and then you can create an app,
studio app right within the future
and it will take you straight there
and you can start using it right away.
Awesome. What advice Emmy would you give to teams
that are just starting out on their app journey
and moving from this world of data
and dashboards to purpose-built applications?
Yeah, I mean, I would say go to developer.domo.com.
That's where we have all
of our documentation on doing stuff like this,
and we've spent a lot of time and effort on improving it.
Um, there's a section in there called Build Apps
where we have a page on App Catalyst
and we have a page on the app platform in general.
And there is so much good information in there about
things like the manifest file, um, the app platform APIs
that you can use, how you're going to connect to data,
how you can save data in, in a database, um, all the stuff
that we've been talking out about, it's all in there.
And so if you just go in there
and familiar set, familiarize yourself with it,
that's really gonna go a long way.
Um, if you don't wanna read through everything,
at the very least, take it
and give it to AI to digest for you
and then summarize it so that you can get the information
that you need before you start.
Cool. Emmy, I love talking
because I feel like every time we chat I learn
so many more things for the people
who are going on this journey, they're trying to learn
how can they reach out, how can they, you know, talk
to this community and, and learn more, uh, from real people.
Yeah, I mean, demo has our just general community
that you can find when you go onto our website
or you're looking through the knowledge base.
And that's a great place to ask questions and get answers.
Um, I would also say that if you're needing extra help, um,
Domo has some great, we have like the dev
portal that I mentioned before.
There's a lot of good resources on there,
but if you're wanting more like one-on-one help,
we have our ACE team, which is advanced customer enablement
and they're really good about helping you with app ideas
or just getting apps
integrated into Domo instance or in your team.
Um, and then we also have our engineering services team at
Domo and we live and breathe custom apps
and we build them for our customers every day.
So, um, if you want help building an app
and it's a little more intense than App Catalyst can handle,
then come to us and we'll help you out with that. Awesome.
Thank you Emmy, so much
for this conversation, for everything you've shared.
Cannot wait to see what people are going to build
and what we're gonna see next outta this team.
So thank you em, me too. Last one.
If you remember nothing else, a successful app starts
before any lines of code have been written.
And Domo helps teams build with production in mind
before they've written any code
and helps them deploy them faster.
Thank you for being here with us today.
We're so glad you could make it
and that we could talk a little bit about apps.
Hopefully you've taken away something that you wanna build
with App Catalyst.
The good news is App Catalyst is available now
to try in beta
and in free trials if you don't already have a Domo account.
We can't wait to see what you're going to build,
the solutions you're going to create
and the change that's gonna happen in your business as use.
Use AI to develop purpose-built software
that helps you drive outcomes.
Keep in touch.
Let us know if you've got something, questions, comments,
or just wanna learn more.
Reach out in the chat on social or in the community.
We'll talk soon. Thanks.
‍