AI Workflows for Support Ticket Integration

Julián
Hello! Welcome to Code[ish] the Heroku podcast. I am your host, Julián Duque, Principal Developer Advocate here at Heroku, and with me I have today Keegan Bakker. He’s the CEO of Audada. Hello Keegan, how are you doing?

Keegan
I’m very well. Thank you for having me.

Julián
Of course. Thank you for joining us here at the Code[ish] Podcast. First, tell the audience who you are, what you do, and a little bit more about Audata.

Keegan
So, Audata is a media software business that I started in 2015 when I was a radio presenter. So, I was a broadcaster myself.

Julián
Radio presenter, okay.

Keegan
And yeah, had a very big radio station in Australia, in, in Melbourne. And we were doing a lot of prize giveaways and things, and I… we were still writing it down on paper, and I made this simple app prototype one night to automate that process. So it sends the winners like a text message with a unique link. They provide their details instead of us having to get it over the phone and write it down, and then it handles the fulfillment of their prizes. And then, also unlocks opportunities with that first-party data and things like that. And, I was a hobby developer and wrote that one night. It was a Ruby on Rails application. And I tried a couple of different services to deploy it to get it running, and I remember then I tried Heroku, and it worked immediately, like, because I couldn’t get it to work anywhere else, and it worked the first time, and I was so excited. And, initially, it was just going to be me like using that at work, but it quickly became the industry standard in… it’s used all over Australia and New Zealand, but also now some of the markets internationally, which is very exciting. And still the same app on Heroku today and as well as other products that we now offer to our customers all in the audience engagement space.

Julián
Awesome. I always love hearing stories of people doing their first deployment, and it just works.

Keegan
Yes.

Julián
And the magic behind that is that today it is the same. So it’s the same experience, the developer experience that everybody has a lot that is just I push, and it works, so that’s beautiful.

Keegan
And I… I… I didn’t know much about DevOps, and our customers say we’re a real small business, but our customers are all enterprise size. And so when I then started going out and trying to sell to them, sitting in front of CTOs, and they’re asking me about data security and compliance and all that kind of stuff. So, I mean, we were… I was glad that we were on Heroku as well because we kind of had that all built in. It was a trusted platform as opposed to having to go and create something from scratch myself. And still to this day, we sort of focus on… we love, you know, building new products, building new features, and getting it out into the hands of customers as quickly as possible. Constantly iterating and evolving and trying new things and new ideas. And that’s still sort of in that DNA of Heroku and why we use it, because you can just try something, write some code, build a new thing, push and and it works. Instead of having to spend time investing in DevOps.

Julián
Yeah. And it’s amazing because you’re talking about getting this idea as a hobby developer. So pretty much like a hobby project. And now it is a full business.

Keegan
Yes.

Julián
So how has been that like growing and evolution process for you? Because I know right now you’re using technologies beyond Heroku. So you are also integrating with Salesforce or using Agentforce. So how has been that that whole process of growing this business and application?

Keegan
Yeah. Well, thing about Heroku, it was growing the business is just sliding, dragging, dragging the slider up. But no, so, we did, we early on started with, Salesforce for CPQ or Revenue Cloud as it’s now called. So, I figured out pretty quickly, because of the way our business worked, we couldn’t really do like, a simple monthly tiered subscription model. We had had to do some real complex solutions, bespoke for each customer. And so we had a lot of SKUs in our price book. And… so we went, we went to Salesforce for CPQ so that we could handle quoting and putting together those quotes, and and also upselling and things like that, you know, adding new licenses for a customer, for example. And we went with Salesforce because of Heroku Connect so that we could have the Salesforce data and our Heroku data seamlessly connected together. So for those who don’t know, that’s what basically what Heroku Connect does. It just takes your Salesforce data and makes it available instantly in Heroku and vice versa. As if it’s in your own database, which it is in your own Heroku Postgres database. And so, we used that to make provisioning instant. So as soon as the contract goes off and gets, you know, Docusigned or whatever by the customer, that contract is in effect in Salesforce, and instantly we just in our application, we just have a line of code that just checks our Postgres database to see what… if the customer has that product subscription, then they can use this product, rather than needing to do API calls or anything, it’s immediately available, whatever their… whatever their licenses allow them to to do is, is accessible inside, inside our application, so.

Julián
That’s amazing. I was like showing Heroku Connect to one of our, like, prospects at the booth at Dreamforce. It was just like, okay, let me connect, click just map some fields, and it’s done.

Keegan
And then you… like as a developer, you’re just writing your own SQL or whatever, then accessing that Salesforce data in just the way that you would access your own application data, just directly without having to change anything about your code.

Julián
Yeah. We are in this new era of AI and agentic applications. So, what has been the story of Audata to get into into AI, or how are you implementing AI into your workflows?

Keegan
We started with… we started deploying Agentforce about 12 months ago. And, obviously, the, you know, the evolution in that time has been dramatic. And, and, I mean, there’s so many potential applications. For now, we’ve started with, for, I think, the low hanging fruit, which is service and support, and we’re really at the moment using AI to… to allow us to expand globally. So we’re entering new markets, like Latin America and Europe, where there’s different languages spoken. And the biggest challenge is timezone differences. We really pride ourselves on the level of customer success that we deliver and the support that we deliver. We constantly have, you know, great CSAT scores for our support and we’re very proud of that. And so, you know, it’s always nervous… a nervous time entering a new market if, if I was worried that we might not be able to deliver that same level of service and support. So, instead of just increasing headcount or having new, you know, another support center and another support team working in a different time zone, we can have agents that aren’t just answering questions, but at the moment, keeping the human in the loop, but the agents are currently really key for triaging cases and determining how critical a case is. So, if someone’s got an issue, the agent can… and it’s not just like, you know, a drop down box on a on a customer support form where you say how critical is the issue and the customer always chooses super critical.

Julián
They’re like super high of course.

Keegan
Yeah, because of course if it’s an issue to them then it is super critical. But the agent can kind of reason about is it… is the… is the issue having an immediate business effect? Do you have a workaround in place? Is it, what’s the level of the impact to the customer’s business? And and then determine, for example, if it is critical it can it can maybe page someone, one of our engineers or myself in Australia if… even if it’s the middle of the night if it’s a very serious issue, or otherwise, it might wait until the morning and, you know, put the support case through. So, so that’s given us confidence that we can, even if we’re not, you know, 24/7 always manning the phones, so to speak, we can still have confidence that if there is a problem that needs immediate attention, we’ll be made aware of it. But we’re doing that as well with Heroku Connect, I mentioned before Heroku Connect with CPQ, but we’re also I mean, that’s… we’re still in the early stages, but what’s really exciting is, again, all of our products live on Heroku so our customers’ data is on Heroku Postgres. And so, I’m, you know, imagining we soon will be able to have our agents handling customer support that can see into our customers’ data in Heroku through Heroku Connect, and even take action on it and edit things, maybe fix code, fix bugs, in real-time. So, having that all on that connected ecosystem is really exciting, so that we can have agentic AI that sort of has trusted access to our product.

Julián
It’s great how can you like… now you’re seeing like the like the evolution, like the path to improve by leveraging the technologies you have today. And that’s pretty much how a lot of companies are approaching AI. Let’s start with like an assistant, or a chatbot to solve these specific issues, but then you see all the potential to start, like, do it more. Have you considered any other use cases where AI can augment what you do at Audata?

Keegan
I mean… So many. There’s a… there’s a million. I’m interested… Well, so to go back to what we were talking about quoting, and how, like, we were originally using CPQ because it makes it easy to add a new license for a customer on a complex contract and then provision that. We still have lots of different products in our price book. And for a sales rep, it can be quite complex for them to, like, they might be messaging me on Slack going, hey, which product do I need to put in for this customer? So as you know, having an agent that can be like a sales agent, so the rep can just say, hey, my customer, you know, can message the agent in Slack and say, hey, my customer needs this, this, this. And the agent will generate the quote, and then the quote gets signed, or the contract gets signed, and then it’s instantly provisioned in Heroku. So that’s something which I’m excited about. But I mean, the use cases go on and on. I’m sure… I don’t imagine there’s any part of the business that can’t be or won’t be in some way transformed through agentic AI.

Julián
It’s very, very inspiring to see when you get the problem and how to solve it with this. So, Keegan, thank you so much for sharing Audata’s story, how you are implementing Heroku and Salesforce to solve the problems for your customers, and making, of course, like a satisfaction increase for them. So thank you for… being with us here at the Heroku podcast, and looking forward to seeing how Audata keeps growing.

Keegan
Thank you very much. It’s been great.