In this episode
Jon Hyman, Co-Founder and CTO of Braze, joins Greg Kihlström live at Braze Forge 2026 at the Fontainebleau Las Vegas to explain which customer engagement decisions Braze now hands to AI agents and which it keeps with people. Hyman covers three products discussed at Forge: Decisioning Studio Go, a self-service route into reinforcement-learning-driven 1:1 journeys; Agentic Standards, which turns the pre-send QA checklist into an automated agent check; and Operator Connect, which lets teams invoke the Braze AI Operator from tools like ChatGPT, Microsoft Copilot, Claude and Slack. The line he draws is specific. Operator can create, edit, test and fix anything in the Braze dashboard, but it cannot launch a live campaign into production without human approval, and he argues creativity and strategy remain the marketer’s job.
Key takeaways
- Decisioning Studio replaces static journey logic with reinforcement learning. Instead of building “send an email, wait two days, then branch” flows and A/B tests by hand, a marketer defines a goal, supplies content and sets guardrails, and Braze sends each customer an individual sequence of channels, timing and offers.
- Decisioning Studio Go removes the data science requirement. Decisioning Studio Pro comes with a forward-deployed data science team that tunes the model to the customer’s use case. Go is self-service: target a segment, provide creative assets, set basic guardrails such as time of day and sends per week, and Braze optimizes toward the goal.
- Agentic Standards automates the pre-send QA spreadsheet. Marketers write freeform rules, for example that every promotion must include “limited time offer” or “while supplies last,” that unsubscribed users can only receive transactional email, or that SMS goes only to the double-opted-in group, and an agent checks each campaign against them.
- Campaigns can be checked against the creative brief. Teams can attach the brief from their creative team or agency to a campaign, and the agent verifies the subject line, assets and layout against what the brief specified.
- Operator covers 100% of the Braze dashboard, with deliberate exceptions. Hyman says anything a marketer can do in the dashboard, Operator can now do, but Braze does not let Operator launch a live campaign into production. That step requires human approval.
- Operator Connect exposes an agent, not an API. Hyman argues that putting a semantic layer over REST APIs and calling it MCP support leaves the outside model reading developer documentation and guessing at API calls. Operator Connect instead launches the same Operator agent that runs in the dashboard, with the customer’s brand guidelines, data and performance history as context.
- Governance matters more as people build less by hand. Hyman expects that within about nine months most marketers will not be building campaigns themselves. Because people have less direct knowledge of something they did not build, automated QA, previews and accessibility testing carry more weight.
- Agent work is becoming multiplayer and proactive. Bringing Operator into Slack lets a whole team direct the agent together, and running it as a background agent lets it flag a flat conversion rate or an underperforming subject line and propose a change.
- Creativity and strategy are what AI can’t substitute. Citing Braze CEO Bill Magnuson’s “wind tunnel of AI” point, Hyman says optimization alone pushes brands toward sameness, and only people can push back against that.
Chapters
- 00:00 — Which decisions should still belong to people when AI runs the campaign
- 01:16 — What Braze does and Jon Hyman’s role as Co-Founder and CTO
- 02:20 — Decisioning Studio: reinforcement learning in place of hand-built A/B tests
- 04:36 — Decisioning Studio Go: self-service 1:1 decisioning without a data science team
- 05:48 — Keeping automated output on brand: Braze AI Operator and Agentic Standards
- 07:37 — Agentic Standards replaces the pre-send QA spreadsheet
- 09:00 — Checking campaigns against compliance rules, cultural dates and the creative brief
- 09:37 — Operator Connect: Braze inside ChatGPT and Microsoft Copilot
- 10:16 — Operator can do everything in the dashboard except launch a live campaign
- 11:43 — Why an API behind MCP isn’t enough for customer engagement
- 14:37 — Multiplayer agents, Operator in Slack, and what faster building means for governance
- 16:58 — Proactive agents: Operator as a background agent that flags underperformance
- 18:56 — How Braze anticipates what enterprises will want in one to two years
- 20:50 — What stays with human marketers: creativity and strategy
- 22:27 — Highlight from Braze Forge 2026
- 23:07 — How Jon Hyman stays agile: a CTO who still writes code and takes on-call
What does Decisioning Studio Go change about 1:1 personalization?
Hyman describes the usual lifecycle campaign as a multi-day, multi-step, often multi-channel flow built on static business logic, with the marketer also designing the A/B tests for cadence, frequency and channel. Decisioning Studio uses reinforcement learning to test many versions of those choices and send each person an individual journey. The marketer sets a goal, such as getting first-time purchasers to a second purchase, supplies a set of emails, push notifications and SMS messages, and defines guardrails on how often and when to communicate. Until now, doing this well required the forward-deployed data science team included with Decisioning Studio Pro. Decisioning Studio Go is the self-service version for customers who want to get started with reinforcement learning on their own.
How Agentic Standards automates campaign QA
Hyman says marketers typically work through a large QA spreadsheet or document before sending, because launching to 5 or 10 million consumers is daunting and people want to check several times over. Agentic Standards lets teams enter those rules in freeform language and has an agent check each campaign against them. He gives examples ranging from required promotional disclaimers and opt-in rules to cultural sensitivities about which days a campaign runs, and says the beta included almost 200 customers. Teams can also attach a creative brief to a campaign so the agent checks the build against what the agency or creative team specified.
Why Operator can do everything in Braze except launch a live campaign
Every demo on the Forge stage was driven by Operator, with no dragging, dropping or clicking by the presenter. Hyman says Operator now has 100% coverage of the dashboard, so anything a marketer can do there, Operator can do. Braze chose a few boundaries deliberately. The one he names is that Operator cannot launch a live campaign into production, because Braze wants human approval on that step. Creating, editing, fixing, tweaking and setting up drafts are all open to the agent.
Why Operator Connect isn’t just an API behind MCP
Hyman’s critique of typical “headless” SaaS is that vendors take their REST APIs, add a semantic layer and call it MCP support, which in practice means an outside model reads the developer documentation and works out which API calls to make. He argues customer engagement needs more context than that: brand and style guidelines, what has and hasn’t worked, the customer’s data and catalogs, and knowledge of which Braze product fits which use case. Operator also runs validation such as email previews across inbox types and screen sizes, spam checks and accessibility testing. So what sits behind MCP in Operator Connect is a call that launches the Operator agent inside Braze, the same one customers use in the dashboard, which can then take inputs passed from tools like Figma, Notion or Google Drive.
Why governance matters more when marketers stop building by hand
Hyman compares marketing to software engineering, where he says teams went from writing code by hand to AI writing it in about nine months, and he expects most marketers will not be building campaigns themselves nine months from now. When more gets built and the person responsible did not build it, they have less direct understanding of what was made. He compares it to handwriting versus typing for recall. His conclusion is that automated standards checks, previews and accessibility tests become more important as speed increases.
What stays with human marketers?
Asked what still belongs to people, Hyman answers creativity and strategy. He relays two examples from Braze CEO Bill Magnuson: shoe brands at the World Cup that each made pink shoes to stand out and ended up looking alike, and a keynote slide of 30 similar-looking cars that illustrates how optimizing everything in the same wind tunnel removes what makes a design distinctive. Hyman says an LLM can help with brainstorming, but marketers will direct the work and agents will carry it out.
FAQ
What is Braze Decisioning Studio Go? Decisioning Studio Go is the self-service version of Braze’s Decisioning Studio. Customers target an audience segment, supply creative assets and set basic guardrails such as time of day and sends per week, and Braze uses reinforcement learning to work toward the goal without the forward-deployed data science team that comes with Decisioning Studio Pro.
What is Braze Agentic Standards? Agentic Standards lets marketers define structured and freeform rules for brand voice, compliance, copy, personalization and campaign setup, and has an AI agent check each campaign against them before it goes out. Jon Hyman describes it as an automated replacement for the QA spreadsheet marketers work through before sending.
Can the Braze AI Operator launch a campaign on its own? No. Jon Hyman says Operator can do anything a marketer can do in the Braze dashboard, including creating, editing and fixing campaigns, but Braze deliberately does not let it launch a live campaign into production without human approval.
How is Operator Connect different from putting an API behind MCP? Hyman says most headless SaaS offerings expose REST APIs through a semantic layer, which leaves the outside model reading documentation and choosing API calls. Operator Connect instead launches the Braze Operator agent itself, which already has the customer’s brand guidelines, data and performance history and can run previews, spam checks and accessibility tests.
What does Jon Hyman say AI cannot replace in marketing? Creativity and strategy. He argues that optimization alone pushes brands toward looking the same, and that marketers will act as directors while agents bring the work to life.
About Jon Hyman
Jon Hyman is the Chief Technology Officer and Cofounder of Braze, and has been at the direct forefront of the company’s growth in the last 15 years, more recently leading the engineering organization through a rapid transformation in how it builds software with AI. He’s become an influential, practitioner-level voice on what it actually takes to move an established engineering organization, not an early-stage startup, into the agentic era.
Jon Hyman on LinkedIn: https://www.linkedin.com/in/jon-hyman/
Resources
Braze: https://www.braze.com
The Agile Brand podcast is brought to you by TEKsystems. Learn more here: https://aglbrnd.co/r/2868abd8085a9703
The hardest part of martech right now is telling which tools are actually earning their keep. That’s the question running through the MarTech Summit Boston agenda on October 27. We’re attending MarTech Summit Boston on October 27 as a media partner. If you want to join, the code AGILE15 saves you 15%. https://aglbrnd.co/r/c70c250b8d52dff3
We’re proud to be a media partner for #MAICON26 – Oct. 13-15! Learn how AI can power your marketing and business and help you grow smarter. Use code AGILE150 to save! https://aglbrnd.co/r/7fe458ced0f04658
Need to evaluate martech, AI, or CX platforms and don’t know where to start? The Agile Brand Guide List is now in beta. We’re building the most useful, free-to-use, and up-to-date way to assess what truly matters when evaluating platforms. See more here: https://aglbrnd.co/r/f76be3a33cc491bb
Reach your customers with Reddit. Spend $500 in ad spend, get $500 back in ad credit! Learn more: https://aglbrnd.co/r/491818c79fb1873f
Chaser is the only Slack-native project management platform that helps teams turn messages into tracked tasks, automate follow-ups, and maintain team-wide visibility, without adopting another tool. Now integrated with Claude and other GenAI tools. Learn more at trychaser.com and use code AGILEBRAND for a 3-month free trial (normal trial is 14 days).
Start building your own apps with Replit and get $20 off. Learn more: https://aglbrnd.co/r/93531742a7625a20
Enjoyed the show? Tell us more and give us a rating so others can find the show: https://aglbrnd.co/r/faaed112fc9887f3
Connect with Greg on LinkedIn: https://www.linkedin.com/in/gregkihlstrom
Don’t miss a thing: get the latest episodes, sign up for our newsletter, and check out Martechipedia, the wiki for marketing technology: https://aglbrnd.co/r/35ded3ccfb6716ba
The Agile Brand is produced by Missing Link—a Latina-owned strategy-driven, creatively fueled production co-op. From ideation to creation, they craft human connections through intelligent, engaging and informative content: https://aglbrnd.co/r/f1782ab12211f765
Transcript
[00:00:00] Greg Kihlström: Hi, I’m Greg Kihlström, host of The Agile Brand, and here’s a question for you. When AI can run the experiments and ship the campaigns, which decisions should still belong to people? We’re recording this episode live at Braze Forge 2026 at the Fontainebleau in Las Vegas, and we’re gonna talk about what changes when AI takes on more of the work of customer engagement. To help me discuss this topic, I’d like to welcome Jon Hyman, co-founder and CTO at Braze. Jon, welcome to the show.
[00:01:09] Jon Hyman: Hi, thank you so much for having me here.
[00:01:11] Greg Kihlström: Yeah, looking forward to diving in here. before we do, though, why don’t you give a little background on yourself and your role at Braze?
[00:01:16] Jon Hyman: So for everyone who may not know what Braze does, we’re a customer engagement platform. So we help consumer brands use their first party data to deliver relevant and memorable customer experiences through personalized messaging like emails, push notifications, text messages, internet messages, and more. And I’m the CTO. We started the company about 15 years ago, and so I’m responsible for our engineering function, overseeing the R&D of the product. And also what I love about my role is I stay at a high enough level, all roles, then enterprise SaaS or sales roles, and I just get to connect with business, marketing, technology, and other leaders at our customers and help, figure out how we can be better partners to them.
[00:01:50] Greg Kihlström: Love it. Love it. So yeah, let’s, let- let’s dive in here. And certainly, I, I love doing these interviews on site at, at events. There’s always a l- lot- lots of announcements, lots, lots going on, and so we’re gonna talk about a few of those today. First, I wanna start with, there was an announcement about Decisioning Studio Go, and for those listening that may not be quite as familiar with Braze Decisioning Studio in general, maybe give us just a little, little background on what it is and, and what Go kinda adds to that.
[00:02:20] Jon Hyman: So essentially, when you think about a marketer building a life cycle marketing campaign, they typically won’t be composing something that’s multi-day, multi-step, often multi-channel, that’s ushering a customer through a part of a customer journey. Perhaps that is an onboarding flow, an upsell flow, or preventing churn or any kind of win back routine there. Now, essentially, what we are doing with our Decisioning Studio product is saying that we can reimagine this workflow entirely and move away from marketers needing to build their own AB tests and experiments and testing out the cadence and frequency and channel and all of those different things here, and essentially use reinforcement learning in order to test out lots of different version of those things and then figure out what’s working for individuals and send every individual user their own personalized journey.
[00:03:05] Jon Hyman: So we can move away from, again, customers defining what I’ll say is static business logic, things like send an email, wait two days, if they open an email, do X, if they don’t open an email, maybe send a push notification, into something more where a marketer can define a goal saying like, “I wanna take my first time purchasers and get them to make that crucial second purchase, give us a bunch of different content.” Maybe they give us five or six emails, four or five push notifications, two, three SMS messages, set some guardrails, how often we can communicate with them, what times, when, any other restrictions on what we might be able to– what decisions we might make or what content to send, and then let us go off to the races and have us say something like, “Look, based on all the content, we will send John two emails this week,” or, “We’re gonna send Greg three emails and a push notification
[00:03:51] Jon Hyman: this week.” And then they get them on different days at different times with different con- content, and there might even be different offers or promotions inside all them ’cause we can use AI in order to do that. And so that’s essentially what the, the Decisioning Studio is all about, is again helping us just automate the journey creation, the one-to-one for every single customer using learning. Now, up until now, to do a high quality job of this, we have needed the assistance of a data science team. So when customers purchase our Decisioning Studio Pro product that comes with a forward deployed data science team, that really fine-tunes and customizes a reinforcement learning model in order to the, to adhere to, to best give results for customer’s use case. Now, we hear from a lot of customers that they wanna get started, kinda dip their toes into
[00:04:36] Jon Hyman: reinforcement learning and, and decisioning, and so we built a self-service offering that can make it really simple for customers to just get up and running and doing this kind of thing themselves. So Decisioning Studio Go is a self-service product where customers can target an audience cohort or a segment, give us a bunch of creative assets, define some basic guardrails around things like time of day, number of times to send a week, and then Braze will just go off and try to help them convert on that goal.
[00:05:03] Greg Kihlström: Love it. Love it. And so that sounds, I mean, that sounds amazing for those marketers out there and, having done e- enough one-to-one [chuckles] pilots and things like that, definitely to your point, it’s, it’s, it takes a lot of, of data science and, and-
[00:05:20] Greg Kihlström: … other integrations to real- truly make that happen in the enterprise, so something like Go sounds, sounds amazing. You, you mentioned the guardrails and I, I think probably some people out there listening say, “Well, this sounds good, but what about, you know… How, how do you keep things in alignment where, let’s say you’ve democratized this, so you have marketers doing this and they’re self-serve, but how do you, how do you make sure that things stay on brand or just, everything kinda running, running in the, in the right way?”
[00:05:48] Jon Hyman: This is really where the assistance of other AI agents comes into play here. So one of the core components of our Braze product is something called the Braze AI Operator, which essentially you can think of as like your agentic assistant inside our dashboard, inside our product, that can help you compose messages. And we’ve built a lot of structured data inputs into our dashboard to allow us to collect things like your brand voice, brand guidelines, email style guidelines. We also know what works, what, what doesn’t work from the messages you’ve sent- Today we made another announcement of something called Agentic Standards, which allows you to create structured rules for us to ensure that we’re abiding by when we have agents creating things around brand voice, compliance, copy, personalization, campaign setup, et cetera. So we have all that information
[00:06:33] Jon Hyman: that, that you input so that we can then check all that stuff. So now when you compose an email, if you’re doing a promotion, you might wanna say something like, “Every time you run a promotion, you gotta have the words ‘limited time offer’ or ‘while supplies last’ at the bottom.” And so we can now enforce that through an agentic check to make sure that you’re actually adhering into that sort of thing there. And so it’s been really exciting because agents can just be much more granular and, and thorough than I think, [laughs] some- sometimes it’s just really easy to miss, checking the checklist of 50 things, but an agent can be really thorough and go through all of that stuff.
[00:07:07] Greg Kihlström: Well, yeah, ’cause it’s, I mean, it’s one thing to catch, I, I wouldn’t even say the obvious, but the easily quantifiable errors. But to your point, when something’s a little off-brand, a marketer knows that, right?
[00:07:37] Jon Hyman: Yeah, so really, like, what we see marketers do is that they have this, like, giant QA spreadsheet or document that they go through, like, before they send out a campaign. ‘Cause sending out a campaign’s pretty daunting. Like, the, the anxiety between, creating something and then hitting the send button and launching it to, like, 5 million or 10 million or possibly even more consumers, there’s, there’s a lot. People wanna double, triple, quadruple-check it’s working there. And, so what we, we’ve built this product, Agentic Standards, that, lets them put in any kind of freeform rules here. That’s what’s so great about LLMs is it can, it can… It’s kind of like a smart coworker, right?
[00:08:14] Greg Kihlström: Yeah.
[00:08:15] Jon Hyman: It is a smart coworker. And so it can take a generic rule that you might say of things like, if you’re going to send to users who are unsubscribed, that must be or can only be a transaction email. If you’re going to send, an SMS message, it has to be to the double opted-in SMS group. You can, you can define rules like that, and then we provide, do, we do kind of the, I’ll say, the context massage and context stuffing to giving to the large language models in order to m- for it to know whether or not your campaign a- adheres by that. And you really can do anything and everything. We’ve worked very hard throughout our beta period. We’ve had a beta with almost 200 customers, I believe, in order to really just kind of look at what are the types of things people wanna do. And we’ve got everything from, like, cultural sensitivities to make sure that, certain campaigns don’t run on days
[00:09:00] Jon Hyman: that, might be culturally significant, to, to folks, all the way into, just legal and compliance issues there. And so there’s, there’s kind of a lot that we support. And lastly, we, we know that a lot of marketers create campaign briefs with their creative teams or their, or their agencies, and so with this product, you can drag in your creative brief and attach it to a campaign. So when we go to check it, we actually can check it against the brief, which is great.
So your agency delivers you something, says, “This should be the subject line. These should be the assets that you’re using. This should be the way the email should look.” We can just check all that stuff now, so it’s, it’s really cool.
[00:09:37] Greg Kihlström: Love it. Nice, nice. So I wanna talk about, another thing, another announcement, so Operator Connect. And so I saw the, the demo on stage, as far as, being able to bring Braze into AI environments like ChatGPT, Microsoft Copilot. historically, the platform itself was kind of where all the approvals and, and permissions lived. So what, if you could talk a little bit about, how does this, how does this change the, the dynamic of, how marketers are working as well as how do- does that, does it change anything about governance as well?
[00:10:16] Jon Hyman: So one of the things that I really liked about what we did today on stage, and then we were deliberate in making this decision, is that all the demos that we showed, we had Operator power, the creation, the editing, the management, and the fixing of any of the things that were on stage, meaning that the human who was running the demos didn’t actually do any dragging and dropping or any clicking themselves. They simply used Operator to drive forward the demo. And we’ve been working hard on making sure that Operator can, can agentically operate our entire dashboard. We now have 100% coverage, meaning that anything you, a marketer, can do in our dashboard, Operator can do in the dashboard. We do have some boundaries, so for example, we don’t let Operator launch a live campaign into production. We want human approval on that.
[00:10:57] Greg Kihlström: That’s probably good. [laughs]
[00:10:57] Jon Hyman: But, exactly. We picked a few things deliberately, but, but, when it comes to creating, editing, fixing, tweaking, setting up drafts, composing stuff, Operator can do all those things there. And we know that a lot of companies, though, are doing a- agentic tooling outside of Braze, and they want to say, “Hey, why can I only build my campaigns or my, my journeys inside Braze?” And so we released a product called Operator Connect. I think it’s a very special product because when you think about, like, like, headless SaaS offerings, what I see in market a lot of businesses have done is they essentially took their REST APIs and then they put a semantic layer in front of it, which lets the agents understand how to use them, and then they said, “Great, now we’ve got MCP support for headless support.” But really what that is
[00:11:43] Jon Hyman: is you can think of it about being like Claude or ChatGPT, whatever it is. Claude is reading your developer documentation, figuring out what API calls to make, and it’s making the API calls. That’s basically what it’s doing. But customer engagement requires a lot of context. It requires strategy. And so when we go back to what I was saying earlier, that we have all this structured understanding of your business, your brand guidelines, stock guidelines, what works, what doesn’t work. We know your data. We know what data you have on your consumers. We know what catalog data or other tabular data you might have inside Braze, and we know what you use. We need all that stuff in order to figure out what’s the best way to accomplish something. We also need to know how to use Braze the best. Some use cases are gonna be really good as decision studio use cases.
[00:12:28] Jon Hyman: Some things are gonna be really good using some of our other AI products there. And why trust that to just how good cl- what, what Claude’s doing? A great Claude reading some developer documentation, taking your goal, and then trying to figure out and map it to what API calls, that’s just not gonna be nearly as successful as Operator doing it, because Operator knows Braze, it knows your environment, and it knows how to do everything. Furthermore, Operator can do validation testing. It can preview emails. It can test them for spam checks. It can preview them on different inbox types and different screen resolutions. It can do accessibility testing, all of those kinds of great things that we wanna do to check and, and make sure things work. We’re not gonna expose that stuff via APIs. That, that seems kind of, kind of crazy, right? And so what we did is we said, let’s build a managed cloud agent platform where what we’ll put behind MCP isn’t, isn’t really an API to, like, say, create a journey. It’s an API to launch an agent inside Braze. And so what Operator Connect is, is it is a full-blown agentic creator, the same thing that’s in our dashboard that our customers are using, that you can now invoke remotely. And this is so cool because whatever you can do now in the dashboard with Operator, you can do remotely with Operator. And I just think that this really changes the dynamic of how businesses can, can build, because inside their agentic tools, like Claude or ChatGPT, they’re connecting their Google Drive, their Gmail, their Slack, their Confluence- their Figma, their Notion, all these other systems here. And so you just have this rich ecosystem to be able to say things like, “Hey, grab our assets from Figma and read this Notion page on what we’re doing and take this, this, this, PDF that’s in Google Drive, and this is the strategy, and then make it happen,” and actually pass that stuff along over to Operator. So, I’m really excited about what, what’s possible, and I’m also really excited about the, the near, kind of next step of, of, of, like, kind of agentic interaction of becoming more multiplayer. What I mean by that is, if you think about Operator on our dashboard, it’s a single marketer talking to Operator w- doing their own work. It’s the same as, like, me using Claude Code in order to code a feature. It’s just me doing my own work. But
[00:14:37] Jon Hyman: I’ve got teammates. I wanna understand what they’re doing. We wanna all have opinions on each other’s stuff. And how do we get to be, quote, unquote, “multiplayer” where we are all giving input into the agent and together we’re creating this content and this strategy? And because Operator Connect now brings the power of Operator to other tools, we get to do that. So one of the things that, that we… I’m actually showing this off tomorrow, and we’re gonna do a webinar on, is, is bringing Operator into Slack. And so that way, marketing teams that are in Slack talking about their strategies can ask Operator to pull analytics and stats and previous stra- and previous campaigns and ideate new ones, and then have Operator go off and build it and s- keep the whole Slack channel updated with what’s going on. So I’m excited about what that means. y- to, to your point on governance, really what, what we say is people are adopting AI, I think, way faster that they would be. I mean, it’s really been staggering how I’m just looking back, like, nine months ago, some people were, like, writing software code by hand, and now it’s, like, 100% code. I think that nine months from now, most marketers aren’t going to be building themselves. And so you think about the fact that more builds are happening, you, you have just less, almost kind of, like, close understanding of the exact thing that is being built because y- you didn’t build it, so your, your brain just cognitively c- doesn’t have the connection to it. Just like writing versus typing, right? [laughs] writing gives you more recall. And so you need to make sure that you’ve got QA checks, and the stuff like the agentic standards becomes more important. The stuff like Operator being able to test for accessibility and do previews, all that stuff becomes more important. And so
[00:16:14] Jon Hyman: we’re, we’re continuing to, to drive forward in that direction, ensure that customers feel that they can launch confidently and quickly with all the agility and ambition that they’ve got.
[00:16:25] Greg Kihlström: Back to kind of the, the multiplayer, and I, I know some of that’s, forthcoming, but, I als- I also think, I mean, this is the way that marketers are working is just fundamentally changing in not only the collaborating and the way that they do that, but also collaborating with what and whom, right? ‘Cause I mean, there’s going, there’s already this concept of virtual teammates and agents and, and things like that, and so it seems like in, being able to collaborate, we’re not just talking about collaborating with humans anymore, right?
[00:16:58] Jon Hyman: Like, what’s exciting, as an engineer, as, and the builder of, of Braze with that Operator Connect is it now gives us the power of Operator as kind of a background agent, something that can work, that we can, we can send a task to and let it chug on its own without you being in the dashboard. And so then that kind of next future vision that we’re gonna be moving toward is one where AI is able to be much more proactive. It’s not you typing into a prompt and then getting the creation. You’re still gonna want to do that, but the monitoring of the strategy can all now be done agentically, where we can look and say things like, “Your conversion rate is flatlined. We can try a couple different things. We can have more testing. We could do more variants. We can do more channel selection, all those things. What do you want?” You know. “Should we go off and experiment with that?” And, and AI agent can do that. Or maybe
[00:17:44] Jon Hyman: your subject lines are underperforming, and we say, “We have another subject line idea that we think is gonna do, perform better. Are you okay with us swapping that in?” There’s just so many things there where we can get to be insightful and, and proactive and, kind of s- like, just make, again, make judgments and opinions around, “Hey, is this performing well as we thought?” Like, we start to forecast how well we think something’s gonna do and then see does it hit that expectation relative to what you’ve sent in the past, and be like, “Oh, you know what? This got a lower click-through rate than the other things here. We think that it could be X, Y, and Z. You want us to change that?” Operator can, w- will be able to do all those things because now we have the ability to invoke it when you’re outside of the dashboard.
[00:18:23] Greg Kihlström: And so, a l- a lot of what we’re talking about is just kind of- As I was saying, like w- how work is evolving i- even in marketing. I mean, so you’ve been doing this with Braze since 2011, I believe.
[00:18:37] Greg Kihlström: So, how do you, how do you keep, ahead of things, or how do you keep thinking, and keep your teams thinking in the right way to… you keep- You’re, can’t see the future specifically, but how do you try to anticipate and, and kind of see where, see around the corners?
[00:18:56] Jon Hyman: There’s, there’s a number of different ways here, and, one of it I’ll just say is very customer-centric. So a lot of things that we’re building are just a such direct connection to the things that our customers are trying to accomplish or what they wanna do, and I love that we work with such a plethora of different businesses across a different number of sizes and digital prof- proficiencies. We work with really digital upstart companies that are just on the leading edge of… it’s basically a, if you have your whole team engineers and they love to build. We work with, like, really large Fortune 500 companies where, sometimes they may just mu- mu- sometimes they move a little slower, they got some great ambitions, and they, but they, they need to go at a different pace. And because we have both sets of companies as well as
[00:19:41] Jon Hyman: smaller startups who really are just trying to be as nimble and move as fast as they can, we get to see what I essentially say, like, what, what these digital upstarts are trying to do today, everyone, the Fortune 500 companies, these big behemoth multinational businesses are gonna wanna do in the next year or two. So we gotta make sure that we’re gonna be able to support this, ’cause the demand signal is, is, is just starting to be heard, but it’s gonna be so loud in two years from now. So we kind of have a, have the luxury of having and servicing a wide round of customers who can give us really great feedback. But in addition to that, we have a very strong product vision around what the future of engagement is going to look like, and this comes from the fact that we employ so many marketers, we live and breathe the customer engagement all the time. We are, are very focused on it as our, as our key, like,
[00:20:27] Jon Hyman: business challenge. And also, we just are very observant around what’s happening in the technology world, and then what are the parallels and learnings that are gonna come to marketers. I was talking here just a moment ago about the cloud code, people not writing code by hand. It’s like marketers are gonna have that same thing. It’s just a natural extension of what we see in one area, as it, as the technology begins to migrate to other parts of the, the business structure.
[00:20:50] Greg Kihlström: Yeah, yeah. And, and so what do you see still belonging to the human marketer? I know we talked about a lot of things that will be done by agents and things like that. What, what still stays with the humans?
[00:21:02] Jon Hyman: The things that really you can’t substitute are creativity and strategy. Bill, my co-founder and our CEO, just had a really fantastic, point that he made a few, like about a month ago, where at the World Cup a lot of brands thought that they would… A lot of shoe companies thought that they would stand out, and they independently all made pink shoes for some of the athletes, and it was kind of this kerfuffle over the fact that, like, everyone thought that they were standing out, but they ended up looking all the same.
[00:21:28] Jon Hyman: He showed something during our keynote yesterday of, of 30 different cars that all looked pretty similar, and he said, “Well, when you put cars in the same air tunnel and you’re just going through and just testing for wind resistance and trying to get the most fuel efficiency you can, you’re gonna optimize out anything that’s gonna look, unique and, and special.” And he said, “You gotta be mindful to not have, to not shave away in the wind tunnel of AI, your creativity and what’s gonna make you stand out, and you’re gonna melt into the sea of sameness.” And I think that it’s really only humans that are going to be able to push back against that. Like, and, and, and an LLM can help you with some brainstorming, but there’s just no substitution for human ingenuity, creativity, and strategy. And so we really see that the marketers are going to
[00:22:13] Jon Hyman: be the directors of all of these things, and it just lives and starts with the marketers. And agents can do, go and bring things to life, but if, if you’re not having a human touch, I just don’t know why your brand is going to do well, because, like, people wanna connect with people.
[00:22:27] Greg Kihlström: Yeah, yeah. Love it. Well, Jon, thanks so much for joining. got two last questions as we wrap up here. first, as I mentioned, we’re here at, at Braze FORGE in Las Vegas. I know we have one more day to go, but what’s been a highlight so far?
[00:22:40] Jon Hyman: Oh, my goodness. Just seeing the energy in the crowds and talking to people. There are so many great companies here that are doing different things and, and at different levels of sophistication here, and I just love kind of the bustling energy of the groups and, and, people just asking each other what they’re doing with Braze and what are their best tips and tricks and best practices.
[00:22:59] Greg Kihlström: Yeah. Love it. And last question for you, what do you do to stay agile in your role, and how do you find a way to do it consistently?
[00:23:07] Jon Hyman: I, I love to adopt technology, and one of the things that I am a firm believer in as a CTO is that I need to stay very involved in actually building of the product. So I have not gone post technical. I submit code, I work with, I’m on, I get escalated for on call and help our teams during, during troubleshooting. I work directly on set of support escalations, just really to make sure that I’m staying really in the thick of it, being on the ground, so that way I can just really very closely represent our technology back to our customers and have a very, very solid, firm grasp of everything that we’re doing. And in that, I learn all the things that are kind of on the cutting edge of how are all of our engineers building. I build a lot of code, I write a lot of code myself, and just, just kind of really get to stay forward in this, like, AI future that we’re just rushing into.
