RAISE Summit 2025 | Betty Junod, Heroku

Clips
News
More from RAISE Summit 2025

Betty Junod

CMO

Heroku

12 signs AI infrastructure is reshaping the IPO market: theCUBE’s RAISE Summit insights

Capital markets are being reshaped by artificial intelligence, with AI infrastructure emerging as a critical pillar of how companies scale, compete and go public. What was once considered a back-end concern is now front and center in IPO roadshows and investor decks.TheCUBE’s John Furrier talks about AI infrastructure at RAISE Summit.The shift is about more than hype. Investors are scrutinizing how companies embed AI into their business models — not just to drive innovation, but to ensure long-term operational resilience. It’s no longer enough to show top-line growth; firms need to demonstrate that their AI capabilities can sustain profitability and differentiate them in crowded markets

play_circle_outline Connecting Silicon Valley and Wall Street: Salesforce's Evolution into 'Agentforce' and Global Technology Trends from Paris

play_circle_outline Heroku’s role in modern application development and data integration strategies.

play_circle_outline From App Development to Building: Heroku's User-Friendly Infrastructure for Seamless Deployment and Scaling for Broader Audiences

play_circle_outline Effortless Application Scaling at Enrollment Peaks: Enhancing Customer Happiness with Heroku's Reliable Service Offerings

Info
Transcript

Betty Junod, Heroku

Betty Junod

CMO Heroku

In this interview at RAISE Summit 2025, Betty Junod, chief marketing officer at Heroku, joins theCUBE’s John Furrier to share how Salesforce’s developer platform is empowering a new generation of builders. Junod explains why English prompts are now a first-class programming interface, how Agentforce builds on Data Cloud to unite customer data with AI, and why Heroku’s managed platform remains the fastest route from idea to live production.

The conversation explores the rise of agentic applications, the industry’s rapid alignment around standards like M... Read more

explore Keep Exploring

What is the current focus and direction of Heroku within the context of Salesforce and the app development market? add

What is the relationship between cloud-native application development, data, and artificial intelligence in the context of a company's portfolio? add

What are some examples of customer success and use cases for Heroku deployments? add

What are some examples of how Heroku enhances customer experiences during high-traffic events? add

bolt Powered by CUBE AI

Betty Junod, Heroku

search

>> Welcome back everyone to theCUBE, live here in Paris, France. I'm John Furrier, host of theCUBE. Two days of coverage here as part of theCUBE, of course, our partnership with the NYSE and Brian Baumann and the team there. expand the coverage connecting Silicon Valley to Wall Street. Now we're going global. We're in Paris. As you notice here, CMO of Heroku, friend of theCUBE, got a great view of the app dev market as well as what's going on with builders. Betty, great to see you.

Betty Junod

>> Great to see you as always. - Thanks for

>> coming back on theCUBE. What I love about what your role is, in Salesforce, is Salesforce is basically going to turn into Agentforce. The name should change. In fact, agents are hot. Marc Benioff came on theCUBE. Was a very famous interviewer. It was a deep dive, very technical, but he basically laid out a vision of essentially the data cloud meets Salesforce. It's not just silos of apps within the company. Obviously, Salesforce is their killer product, continues to have such great success. But as abstraction comes in, agents can abstract away essentially integration opportunities. So, you're on the more the developer side. Give us an update on Heroku, what you guys are doing, because this is where the action's going, right? You got the cloud meets data. Okay, so Kubernetes is doing its thing, making things simpler, platform engineering, check, check, check. But now you got data engineering, platform engineering going into the intersection of an analytics market meets apps.

Betty Junod

>> Yup. - So, it's almost like, it's not as easy

>> as saying App dev as just cloud native. It's cloud native with data equals-

Betty Junod

>> All the things. - ... the new thing.

>> Yes. Yes. And I think when you think about our entire

Betty Junod

>> portfolio, you have things like Data Cloud, Heroku, MuleSoft. Those are really horizontal, right? New foundational underpinnings, right? Because data is what feeds AI, right? And Agentforce is something that is now able to sit on top of all the data that we have, that the customers have been entrusting sales force with, and then be able to express through sales, marketing, service. These are really the functional departments in which the jobs to be done, and how we do that work is going to be transformed with AI. So, I think that's where it's, it's really, yeah, the data and the integration layer and the code, code layer is really going to, it's the silo-busting across. You used a great word, builder. For years, we've been talking about developers, developers, developers, developers. But the kind of stuff that we've been seeing in the last six months and the kind of companies that we see here at this conference, English is now the way you can build things. You can prototype software. What does that mean? Who is a builder? We're unlocking this whole new era of anybody, folks with ideas, folks that understand various business processes that now want to create apps. Now, there's going to be a flood of new apps being written. That doesn't solve the problem of, okay, the app is not real until it's deployed and it's running, and it's accepting traffic, accepting...

>> And scaling up. - Exactly. And scaling and

Betty Junod

>> performing well. And that's really the sweet spot for Heroku. We've always been here to help take away those concerns.

>> I saw a LinkedIn post by Adrian Cockroff, a good friend of ours, talking about they're coding a lot, swarm coding. He called. I did some vibe coding a couple of weekends ago on a Saturday. It's kind of what I do these days. But I didn't write a single line of code. I wrote four apps. Okay? Now, I was playing with Replit and some of the tools just to see how they would do, and these are things that I wanted to do for years.

Betty Junod

>> Yes. - But didn't have the money or time to go sit down

>> and write a doc and go to engineering. I wrote, and I got them up and running, provision the data. Now, the hooks weren't in there. I could have spent more time, but for the most part, it was heavy lifting was done. And what you're seeing is now, that's the mindset of the builder of the magic wand. We always ask theCUBE, if you had a magic wand, Betty, what would you do? Well, actually that's possible. The art of the possible is here now, and all that work under the covers that we used to talk about, service meshes, data layers, all kind of taken care of.

Betty Junod

>> It's still there. It's just like, we're making that not the concern of the person who wants to build. And before it used to be, I feel like what we've done is kind of lowered the barrier to entry, right? You talk about, I had to write docs, I had to, then you have to go convince some of the technical folks that this is an important project. Please build it.

>> Begging. - Begging and pleading for time.

Betty Junod

>> Then you get scheduled, and it's a whole thing, right?

>> Yeah. - But now it's like, I have an idea,

Betty Junod

>> let me prototype a few things. Maybe it'll work enough for you to play around yourself. The questions around, is it still the right architecture to scale? Does it have all the dependencies sorted so that it'll deploy correctly? I mean, those are things where a lot of the technical specialists can still come in and help for deployment, but it's going to be an exciting time of, what are you going to build? Because now the currency is, how creative are you? What are your ideas?

>> Yeah. And I think you put a good point. I think I love vibe coding, but there was a quote on the queue a couple events ago, vibe coding was so 2024. I'm like, well, what kind of thing is 2025 too. But the point was is that engineering it into production is a whole 'nother ball game. So, you mentioned prototyping, which I think is the killer app for vibe coding, because what I did, what our people are doing is, they're seeing the UX side of the functionality and they go, "Oh, I get it. " Now, it's like, okay, that's a go, no, go decision. Just kind of playing around.

Betty Junod

>> Yep. - And I think that's where we're going to start

>> to see that creativity. So, now, this is your world. Now, you're like, okay, your company does this. You move stuff into production. And so, this is where the action is going to, I think going to be a Cambrian explosion. What's your thoughts? Do you agree with that? And what evidence or what data can you share to kind of highlight that, yeah, a 15-year-old to the boardroom is doing, from the elementary school to the boardroom?

Betty Junod

>> Yeah, and this is really where we're working to support all the vibe coding tools, right? Those tools, so that the process from going from, I've built it here, it works on my machine to I want to deploy is smooth. We want to remove as much friction as possible there. And then also, from a technology standpoint, doing things like supporting MCP server, having our own MCP server, also having toolkits so you can build and run any custom MCP server because that is now, that's the handshake, right? In order for a lot of these new apps, which many of them will be agentic, they need to access different tools, different data, different templates in order to start doing the job that's being described of them. So, we're making it a point to, again, do the heavy lifting for the customer, bringing in managed inference, but focusing on what is a great experience for the builder, right?

>> Yeah. So, I have to ask you because I am not going to say the school, but they're a prominent elite computer science school. I personally don't think it's as good as Northeastern where I went to got my computer science degree. Okay, it's MIT. Okay. An MIT guy said, "Ah, MCP." A2A's out there. We had Google on earlier, Allison Wagenfeld was on. So, you got the MCP and A2A, it's kind of an organic kind of consensus, de facto rallying mode. How are people doing these standards? How are you seeing the community react to these things? Because I think it's a good thing. I think it's the biggest surprise of the year for me is the MCP momentum.

Betty Junod

>> 14 weeks? Yeah.

>> It shows momentum around, it kind of feels like Kubernetes at one point when it kind of was agreed upon.

Betty Junod

>> Yeah. - I don't know the exact time that was,

>> but it felt like it was a fast version of, we got to rally around something. What's your thoughts on this community-driven, organic?

Betty Junod

>> I think it's really interesting, and I was just literally just having this conversation last week about how Docker took about 14 months to really catch on, and people were like, now we would love to have some sort of standards. At that conversation, MCP was 14 weeks. So, the time is just compressed. I think A2A, between A2A and MCP, both are trying to do very similar things. As a whole, the market is early. I think it's great that Google and Salesforce is also participating in having A2A be donated.

>> Yeah. I'm sorry, AWS just announced this week their support in getting behind it.

>> So, I think- - Jim Zemlin said the Linux Foundation

>> is creating a

Betty Junod

>> project. - ...

Betty Junod

>> I mean, in the Linux Foundation and CNCF, you and I have been very active with all these years, I think it's a great place that we have seen kind of maintaining that independence, helping mature these projects. And also, it's a place, it's something that industry, the enterprises now look towards as a way of like, hey, if I'm going to make larger long-term strategic bets, I do want to know that things are standard. And I do want to know that enough is open, right? Because that's what we're betting our architecture on.

>> Yeah. One of the things I was asking George Kurian,

>> who was in earlier at NetApp about some of the successes around how data platforms are going into there, and you have all these AI competency centers, a little bit more enterprise kind of vibe there. But his comment was interesting, and I want to pivot kind of to your world, which is, he said, "In competition, understand your tool chains. Understand how things are done, and then rally around that. " So, how are some of the tool chains emerging and how does that play into, or is it relevant as automation comes in? What's your view on that? Because tool chains have been a great conversation in our world, but is that going to be a factor? Should people have to agree on certain tool chains and roles, or is it not a factor?

Betty Junod

>> I mean, I think it's still evolving pretty rapidly. I think that it's interesting. The big thing is, we'll see where the standards are going. I think the idea that the market is dictating kind of open ecosystems right off the bat is interesting. I think we're going to need tool chains. It remains to be seen if the tool chain for agents is the same as we've had for software, just regular software.

>> Yeah, and obviously the agent hype is-

Betty Junod

>> Everything - ...

>> hyped up beyond all recognition, as I would say. What's real in your mind? I mean, obviously I love agents. I think the future of agents is something that I could see. I can see the dots connecting in my mind, so it's pretty obvious to me. But if people are, I won't say super skeptical, but it's like, okay, I'm enthusiastic but not truly confident. Oh, there is agents happening. What's your take on where we are on the evolution of agents knowing what you know, because Salesforce is a Petri dish of agent development because, one, the company leaders behind it, two, you're perfectly set up for agents. You got the data, you got the engineers, you got the data cloud, you got Heroku. So, all of the things are in place for Salesforce. So, what's your take on?

Betty Junod

>> Yeah, I think in many ways, this feels like where some of the cloud-native ecosystem was like a decade ago. So much innovation, so many new companies, so much new tech, right? And I think in that sense, it's like the market is very frothy. But what's real? I think a lot of companies look at this and say, "Do I geek out on individual pieces of technology or do I start to think about true application of this? Where will I use it?" And I think this is where we've really excelled at Salesforce. Because we're anchored in the 360 view of the customer data, we can say, here are crisp and clear defined use cases so that practically, otherwise you're in this endless exploration, and then your ideas have tons of scope creep. But what we've done is say, "Here are some defined agents to start with. " And we're also an organization that's like, we are deploying this internally so that when we go to customers, it's not just, "Here, we built this software for you, but here's how we use it internally. Here's how our support team uses it. Here's how our engineering team uses it, marketing teams, everyone."

>> Yeah.

>> So, the best model.

Betty Junod

>> It reminds me... That reminds me of the couple comments I'll share with you and we'll kind of tie it together.

Betty Junod

>> Yeah. - So, Naveen Rao from Databricks,

>> who's a legendary entrepreneur, we'll see how long he stays at Databricks. He's got a great job there, sold for billions. But he's an entrepreneur. We also had Jensen Wong earlier at CES say, "The agents will become the IT/HR department, " which I thought was a clever meme. But he was taken out of context. After the Databricks event and after the Snowflake event, I now get kind of what he meant. I think I do, but I'll share with you for your reaction. The thing that Naveen said was, "If you're going to do agents, think about the outcome you want to do and evaluate it. " He didn't say like an employee, that's my mind. But he says, "You want to give it a grade, 100% or zero and anything in between. " So, this idea of evaluation is also maths and probability factored in. There's algorithms now on agent effectiveness, and the script is flipping. So, my point is, eventually there is an HR angle here in the sense, of did you do your job? Did you work well with others? Did you make good judgments? Are the customers satisfied? I mean, you could go down to an employee review. Are you proficient at your job? I mean, these are check boxes that we all been through in our evaluations in our careers. This is kind of what agents are kind of going through right now. What's your reaction to that? Is it a little bit too Pollyanna? Is that on point?

Betty Junod

>> No, I mean, I think that I agree with you in that you'll have to evaluate the agent. So, you're going to build an agent for a certain job to be done, and you want that job to be done with a certain level of quality, right? And so, in that sense, what are the metrics that define success? It's not just a case deflection rate, but is it case deflection with good CSAT, as an example? And how accurate are your answers? I do think we'll have to have that because otherwise, hey, is my agent configuration wrong? Is the data that we've set up for it wrong? Or what's not working?

>> Yeah. One of the things I love about your job, you're kind of in the transition on both the app dev, cloud native, now we've got the agent world. I was talking with Frank Fay and Paul Nashawaty the other day and theCUBE team, and we were saying that it shouldn't be called app dev. It's the category that's kind of, the analysts kind of put in a bucket. It's app building, to your point. So, as app dev, the application development kind of, the developer market moves to builders, which basically puts everybody in that population, what's your vision and what are you guys optimizing for now? Because now, we're moving to a much bigger aperture of ease of use, scale, prototyping, UX. It feels like it's all rolled into one, so it's not your classic app dev or is it? I mean, what's your view on the category?

Betty Junod

>> Yes, I love it as more as an app build perspective. And from a vision standpoint, selfishly for Heroku, we want to be the place or the default place where people think of when I have to bring an app to life, because we just make it easy, right? And we get easy, safe, scale, right, for every need. So, that's really where we're trying to take things.

>> Give some examples of customer success you've had.

>> What's in a use case? What are happy customers look like? What's the workflow? What's a day in the life of a Heroku deployment?

Betty Junod

>> Yeah, I mean, we have customers that are doing things like, things like the insurance as an example, like open enrollment. There's massive country- wide scale open enrollment that happens on Heroku. And when you think about what makes them happy is, they don't have to prepare for those spikes. We take care of that for them automatically. Those are the kinds of things that make people happy. So now agents, when you have that with agents and say it's some agent helping out retail, imagine during Black Friday or Singles Day, if you don't have to think about that and the system can handle it for you.

>> Yeah. - We have this also with customers

Betty Junod

>> who are using Heroku with Salesforce, to be an extension and customization of, they know that they are getting kind of the scale and reliability and performance and enhancements to their Salesforce environment, whether that be for, they've just paid for a very fancy, expensive car and they get to follow it through. It's two years of being built. Or it can be anything like handling their support cases for an HR provider.

>> So really, the app plumbing, the app infrastructure, the app depth.

Betty Junod

>> Yeah, . - All right.

>> So, what's on your agenda? First of all, what do you think about the event here?

Betty Junod

>> Yeah. - It's pretty crowded.

Betty Junod

>> Very crowded. It's been exciting to see all of the ideas

>> and innovation happening within the startup ecosystem that's here. So, there's a lot going on here. A lot of folks on the inference side, so have been talking to a lot of folks who could be potential partners for us, providing customer choice, but then bringing that into the experience. And earlier, got to do a panel on what does AI agents mean for the enterprise and that enterprise transformation.

>> What was the summary of that panel? What was the takeaway?

Betty Junod

>> Takeaway is that, think about it, making sure, number one, that you are prepared and ready from a data perspective. Data is what feeds what your AI is going to do. And then, thinking about it from your increasing amount of risk profile. And you don't have to start with the hardest thing first. Start with something that is going to give you immediate ROI and use that to continually better train your teams and then increase your level of risk and agent scope.

>> Betty, it's great to have you on theCUBE. Great to see you. Second half of the year is coming. What's your plans? What's on your agenda? What are you optimizing for? What's your focus?

>> I mean, October 14th,

Betty Junod

>> everyone should be in San Francisco for Dreamforce.

>> Yep. - Yeah.

>> CUBE will be there. - Yes. That's fantastic.

Betty Junod

>> I mean, we're going to have, I mean,

Betty Junod

>> that's our big show for the year. We'll have, I'm sure we'll have tons of announcements across all the products at the company. And then, the Heroku team will continue on ahead to CubeCon in Atlanta and then Vegas for re:Invent.

>> Yeah, we'll see you CubeCon. Great to have you on the Cube. Good to see you.

>> Thanks. - You look great. And we're in Paris.

>> From Silicon Valley, it's like a lot of Americans here,

Betty Junod

>> a lot of Europeans. This is a global event. This is an AI melting pot of innovation. I'm John Furrier of theCUBE, bringing you live coverage for two days. Thanks for watching.