Episode 120: HubSpot Sensitive Data for 2026

Product Management and innovation enthusiast, Hank Lander, currently serves as a Group Product Manager at HubSpot, focusing on Customer Data Protections. Hank's expertise lies in launching new products and building product pipelines. Prior to his current role, he was the Director of Product at ZoomInfo and worked at multiple early-stage startups. Hank holds an MBA in Entrepreneurship from MIT Sloan School of Management and a BS in Computer Science from Georgia Institute of Technology.

What you’ll learn about in this episode:

How AI is reshaping sensitive data governance inside HubSpot.
Which portions of HubSpot support HIPAA compliance via sensitive data
Why AI features that touch sensitive data must be opt-in, scoped, and auditable
How HubSpot is embracing the concept of “toggle off” data access so that customers can decide if they want AI to not touch specific data.
How HubSpot is building in ‘undo’ capabilities with AI integrations like Claude, to provide more control to users.
How AI can improve quality and streamline workflows before the work reaches your team.
Why AI-powered audit logs will become a proactive tool for security and governance.
What the future of CRM looks like as humans and AI agents work together.

Additional resources:

Transcript

Josh Dougherty:
Welcome to a Brave New Podcast. This is a show about branding and marketing in the healthcare space. But more than that, it's an exploration of what it takes to create brands that will be remembered and how marketing can be a catalyst for those brands' success. I'm Josh Dougherty, your host. Let's dive in.

Well, welcome back to A Brave New Podcast this week. Today we are going to dive into a topic that we covered last year, which is HubSpot's sensitive data functionality. And to do that, I've got Hank Lander from HubSpot joining me again. He was kind enough to join a year ago and has come back to tell me about all the updates because there have been a lot of changes that have happened with AI and with sensitive data and I think the general evolution of the HubSpot platform in the course of a year. So we thought it would be a good time to do a refresher. Hey, Hank, welcome back to the show. Thanks for coming back on.

Hank Lander:
Really excited to be here and boy, have things changed in the last year.

Josh Dougherty:
Exactly. Well, we have a lot to talk about, but before we do that, I'd love to give you a chance to introduce yourself and provide a little bit of background for those who weren't able to listen to our first episode.

Hank Lander:
Yeah. So to provide a little bit of background, I've always been interested in entrepreneurship and startups, tried going my own route with a startup, didn't go so great. But that mission has always been a part of me, which really led me to HubSpot. And so I've been at HubSpot for around nine years and through line of my career as I reflect backward is really being drawn to the unsexy stuff, but that really matters for enterprises and adoption of platforms. And that started early with looking at how sales territories are configured to now really leading up a lot of our governance efforts here at HubSpot. Part of what really drew me to this role was really empowering and helping admins to be able to do more without having to hire a team and without having to get a PhD in HubSpot to administer HubSpot. And so that's what really brings me back every day and super excited to be here and talking with you.

Josh Dougherty:
Yeah, awesome. I mean, I think that is what drew me to HubSpot as a partner originally too, because we were looking for a tool to equip our clients that they could, I think, have a superpower and a platform that didn't require them to have a standalone admin. So it's cool to see another kindred spirit who has that same passion. I think the product has gotten more complicated, but it is still easy to use, if that makes sense, 10 years later.

Hank Lander:
I should put that on my LinkedIn, giving admins superpowers. I love that. Or hopefully that. Or I should say I should aspire to that.

Josh Dougherty:
Yeah, exactly. So when we had you on the show last, I think it was April last year, there's a lot that's changed since then. And you alluded to that, but I'd love to hear what you think are maybe the biggest shifts for you and your product team in that time. What are the things that maybe have changed the most that have been top of mind for you?

Hank Lander:
I think the elephant in the room is the first one, is AI changed everything. I think when we last spoke a year ago, people were maybe dipping their toes in AI, but it wasn't really where it is today. And now we're having to think AI first as we're starting to think about governance, especially about things like sensitive data where before it was, let's limit access to AI from sensitive data. And now it's, well, how do we give better control and do this in the right ways or do this safely? And so the whole conversation I feel has changed. And with that, we've hopefully built out a lot of features to give that control back to users so they can tweak the dials to the level that they feel comfortable in when using AI and sensitive data.

We've also invested a lot in our platform since then, as it relates to sensitive data. I think when we spoke last April, it was very much focused on appropriately protecting properties, whether that was a property you created or a property like engagements and really being able to do, let's say, the basics in terms of controls. Now it's much more about improving the quality of how we're scanning for where sensitive data may be in your portal, making it easier to access, looking for additional categories of sensitive data that can be stored and monitored within HubSpot. And then this is a little bit of a teaser, but we're also now, maybe by the time this podcast goes live, also we'll be talking about scanning attachments for sensitive data. So both say, the breadth and the depth of sensitive data has expanded greatly in just one year.

Josh Dougherty:
Yeah. And I should have said this up front, but I will say it now, that we're going to delve into talking about sensitive data, PHI, HIPAA compliance, and I just want to be 100% clear, neither Hank nor I are attorneys or HIPAA compliance experts. So we are sharing from our knowledge about how these compliance frameworks work, but you definitely need to check in with an attorney or with experts to check on what we're saying here. So do not take it as legal advice or compliance advice.

Hank Lander:
Yes, 100%. I'm not going to be providing any legal advice. I will purely be stating best practice, but you should consult your risk or compliance folks first.

Josh Dougherty:
Absolutely. So I think we talk with a lot of clients every day about sensitive data functionality. People are excited about it, but they're also confused about to what extent ... People will ask, "To what extent is HubSpot HIPAA compliant?" Which I think is a big question that I'm like, "Well, it's a little bit more complicated than just putting this label to, I think what I consider a complex process that isn't just like a stamp of approval." So can you talk about maybe the overview of what functionality can support HIPAA compliance and what doesn't within HubSpot?

Hank Lander:
Yeah, and let's talk about that framing. So HubSpot allows customers to be HIPAA compliant. Just being on HIPAA or just clicking that flag doesn't make you HIPAA compliant out of the box. And I just want to reiterate that from last time. So what I mean by that is HubSpot gives you the tools to be HIPAA compliant. So you can sign a BAA, enable sensitive data protections, restrict access, maintain those audit logs, and it gives you the compliance tools to configure and use the platform as you so choose. Technically, what we're doing behind the scenes is we are—I should say when you activate sensitive data with HIPAA compliance in mind—is we are storing that data in an encrypted fashion and then using a unique rotational encryption key system that is on a per tenant basis or said differently, on a per portal basis. So there are multiple keys being involved. It's not just one big key per portal. And we're encrypting that data. And with that, we're putting application level controls. So you can see who's accessing that data, who is encrypting or decrypting that data and creating that traceability.

In terms of, let's say all the technical details, I would suggest going to trust.hubspot.com. That is our source of truth as it relates to this, but to wrap it up with those protections in place plus signing the BAA, that should allow you, let's say, the foundation to be HIPAA compliant if you are doing all the checks and balances internally to meet that regulation in the US.

Josh Dougherty:
Yeah. And I think most people should keep in mind that HubSpot isn't the only system that if you're a healthcare organization that you're using. So it's really there's a HIPAA compliance posture across the whole organization and this really should be seen as something that plugs into the rest of that approach, not just a one size fits all solution.

Hank Lander:
Exactly. It's one tool in the toolkit of the broader ecosystem. Exactly.

Josh Dougherty:
Awesome. So I think many people will know about the general or, hopefully, they know already about the general being able to store engagement, storing different properties securely, but you've been rolling out a beta for sensitive data and AI, which I think most people will be excited about because this is a thing we're all grappling with as marketers, as data ops people. How do you balance the power of AI with the need for data security? I don't think anyone knows that answer perfectly yet, or maybe we will not know that answer perfectly, but how are you and your team thinking about that? How are you balancing those two things?

Hank Lander:
Yeah. And this is where reflecting back on our conversation from last year, it's night and day different. I think last year we were all talking about, nope, it's like a wall. This data will never go over, sensitive data will never go over the wall to AI. And it was more about limiting access to AI. Now what we're finding is that some of the sensitive data is actually some of the most relevant information for AI in that it gives the context to that unstructured data and allows you to do interesting things and that's for each, that's for a customer to figure out. But with that, our stance has changed. And so now it's more about how do we do this safely with the right guardrails and not making this one size fits all. And so, one of our core design principles is that all AI features that touch sensitive data must be opt-in, scoped, and auditable. And what I mean by that is we want to give the power to the super admins to make risk-based choices for the reportal with the information they have.

And so let me give an example of if you have an AI environment, whether it says ChatGPT, Claude, Gemini, whatever, and you have a BAA agreement with one of those vendors and you want to send information to those vendors, fantastic. Who are we to tell you, "No, you can't send that data to that vendor to be processed in some interesting way." Instead, we want to make sure we give you the controls of when you want to send that data, that you can choose to select it. And so it isn't, "Oh, I'm going to send all engagement information" or "I'm going to send all ... " I shouldn't say ... Let me back up. It should be, "I don't want to send all sensitive data. I want to just send notes information," or "I just want to send engagement information," or "I want to send this property's specific information over to the AI frontier model to do the operations you're going to do there."

That's where we're really at and providing that structure and that control and then making it auditable because as we know, everyone's risk tolerance is a little bit different and AI is a little bit unique in that it's not just reading the data, it can synthesize, it can output it, but this also puts more pressure on that trust layer. And so we want to make sure people understand that and can make the right choices. And so for us, that toggling ability is essential for our enterprise customers so they can say, "AI can't touch this" or "AI, yes, can touch this." And then it's enforced at the data layer. And if I had to summarize this down to three words is choice, is power and we want to give that power to the people who know the context the best.

Josh Dougherty:
Yeah, that's what I was reflecting on as you were talking was this idea of power but also transparency because I think all of us have struggled at times over the last year as we're trying to integrate AI into our operations and you get a button that's like, "Connect these platforms." And you're like, "I think theoretically I want to do that, but what am I connecting when I'm connecting these platforms together?" I think that's the big scare.

Hank Lander:
Exactly. Did you read from my Gmail?

Josh Dougherty:
Exactly. So it sounds like that clarity of choice, is it a toggle off for specific pieces of data or how deep is that on/off toggle for access going? And certainly this is probably something that's evolving over time, so you don't have to give me a finished answer.

Hank Lander:
Yeah, I would say this is going to be an evolution in that right now we are doing it at the property level in most places and then at the engagement level. I see this level of granularity only expanding in that there's going to be more choice versus less choice. We don't want to revert back to a world where it's like one toggle, it's either all off or all on. We really believe that the more we can break it down, the more choices we can give customers, the more likely they are to do what's right for their organization. And at the end of the day, that's what we're trying to do is empower those admins.

Josh Dougherty:
Yeah. How are customers responding to that toggle on off features so far in the beta? What have you heard from people?

Hank Lander:
Yeah, it's funny that there are three camps developing in that,, "I don't care, just give me everything, I'll manage the risk." Then let's say the diametrically opposed of, "Never let AI anywhere near my sensitive data, never turn that toggle on." And then the third camp, which is really what I've been talking about is, "I'm so thankful I have this control, let me set it up how I want it." And so far it's been an even split, which I would not have expected where I wouldn't say any one camp is 80% of the volume or anything like that. It's a lot more evenly distributed than I would've thought when going into this work.

Josh Dougherty:
Fascinating.

Hank Lander:
And I'm assuming as adoption becomes more widespread, we'll see more and more acceptance of the risk and going up that maturity curve, let's say.

Josh Dougherty:
Totally. This is a little bit off-script question, but it led me to think back to our conversation about wanting to make sure people can administrate HubSpot without specific admins maybe doing it. I mean, I think enterprise customers often invested in an admin, but how are you thinking about controlling complexity here when you're ... Also, we're talking about a new setting on almost all the sensitive data fields that exist.

Hank Lander:
Yeah, it's a tricky balance, I'll be honest. And the way we've been trying to approach it is—visibility and transparency is when you set up that integration, make it very clear, what are you granting access for? What are you not granting access for? Once you've set up AI, making it very clear, and this is something we're adding on the sensitive data page—where are you sending sensitive data in the list of all the integrations that may be touching sensitive data? And we believe that by providing that visibility and that transparency and making it easy to action, that will empower the admins to make the right choices for their organizations quickly without having to invest a large amount of time. I don't want someone to feel like they need a PhD in HubSpot To administer HubSpot. And so let's make it really easy for them to figure out what's happening.

Josh Dougherty:
Yeah. So I guess this is, again, something that's like necessarily we can't make this less complex, but we can make it easy for someone to make a decision around it.

Hank Lander:
Exactly. And I know we're talking about AI, but this is true for any integration-

Josh Dougherty:
Totally.

Hank Lander:
Anything that is taking data in and out of HubSpot. We want to make it super clear, super easy to understand and be able to answer the questions of what happened to whom and when in our audit logs so that there is that traceability as well-

Josh Dougherty:
Totally.

Hank Lander:
... if something bad were to happen, let's say.

Josh Dougherty:
Totally. Well, let's talk a little bit more about a specific integration. And I think you mentioned Claude a little bit easy earlier about sending data over, but this is something that many people are excited about because Claude has, I think, grown in popularity a lot over the last six months. How are you approaching connecting Claude to sensitive data, ensuring there's a safety net? I think you've mentioned it in our pre-call, there's ability to undo some changes as well as you're looking at this. So can you share a little bit about how that's being approached?

Hank Lander:
100%. And let's take a step back first and say this isn't limited just to our Claude integration. This would be true for any integration. And then more broadly speaking, it also would be true for, let's say our OpenAI integration and whatever other AI integrations we have, but we can keep going with Claude. But let's start with, let's say the final part of the question. What is this undo thing? So what we've found with AI is that it really empowers users to do things that were, I'd say, tricky to do in our UI. Unfortunately, sometimes that leads to mistakes, whether that is the AI hallucinating, the person granting access when they probably shouldn't have. Things happen. It's just life. And so we wanted to make sure it was really, really easy to go back and correct those because HubSpot's data model is highly relational where one action can have a lot of ripple effects that infect both people and other agents with any mistake that's introduced.

And so we wanted to start at the foundational element and say, "Okay, I'm looking at a property history" or "I'm looking at an audit log. I see something that's off. I want to be able to correct that." Historically, the way we enabled that is, we we're like, "Oh, just create a workflow and run the changes backwards." But that put the burden again on the admin. And I feel like I'm beating a dead horse here, but our goal is to make life easier for our admins. And so we really reflected on that and said, "Okay, how could we make this better?" And with that, we came up with the idea of being able to restore, from property history or from audits directly, any of these types of changes.

And we originally rolled that out. It's released now for all users, but with AI, we now are working on exposing that through our APIs as well. And so in the same prompt where you ask Claude to do something and let's say, you went to an event, you collected users at the booth or what have you, and you accidentally uploaded the wrong list to Claude and you're like, "Oh no, this wasn't the list I wanted to put into HubSpot." You can now ask, or I should say, in private beta right now, you can ask Claude, "This is the wrong list. Please undo these changes." And it will be able to call that restore capability itself. And so you can get back to that earlier state. And we really believe that this safety net is critical to really defining that trust layer so that people feel more comfortable using AI to do more things within a HubSpot and really empowering those end users to, not to sound too cliche, revolutionize the way they're working on HubSpot.

Josh Dougherty:
I think that's huge because I think it's one thing to be able to roll it back in platform, but you've got to get out of your workflow and then you have to go and do something separate, move over into the platform. But if you're able to use that agentic interface in Claude and OpenAI or what have you to actually do the fix right there, it's not only a trust building thing, it's a massive efficiency builder for someone as they're working with the system.

Hank Lander:
This is probably a whole separate podcast, but I know the way I've started working is very different. I have, in my case, Cowork open on one screen and then I have the SaaS product I'm working with on the other one and I'm working on both and I wouldn't say I'm going, let's say through the SaaS product and clicking through like I used to. Now I'm asking Cowork to do a lot of that, let's say the more manual tasks on its own. And so I believe that as HubSpot opens the door and really allows a lot more access via APIs to do that dual experience, I actually think that's where the industry is heading and the companies that can do that well and provide tools like the one we just discussed, they're going to be ahead of the curve as that way of operating becomes more of a norm.

Josh Dougherty:
Absolutely. I think that's 100% true because I think everyone's working that way these days. They want to be able to give autonomy to their chosen model to be able to do some work and to be able to trust that it's going to get it done effectively. So another thing I think that is important here, and this is we're talking a lot about safety, which I think is an appropriate conversation, but you are working on an evolution of roles in the portal. I know this is a future looking thing, but you were willing to share and talk about a little bit, but this was specifically making agents users in the portal with their own responsibilities. I would love to hear about how that shifts the operational dynamic and how you handle the concept of agentic reviews and where an agent can maybe see something before a manager sees it. This is really two questions. So first, maybe you talk about this role shift where you're trying to treat agents more as users and then maybe we can talk about this idea or concept of agentic review.

Hank Lander:
100%. Yeah, so let's split these two up and I'll tackle that.

Josh Dougherty:
I was like, let's push it all into one question, but it's really two. Let's talk about agent-

Hank Lander:
... really difficult here.

Josh Dougherty:
Yeah, I know. Let's talk about just agent roles first. Yeah.

Hank Lander:
All right. So our philosophy, our approach to AI has always been about enabling hybrid teams and we're really starting to make strides on that vision. And the first part of that, as you rightfully called, is that safety and trust aspect of it. In HubSpot's philosophy, the way we can best do that is by treating our agents as users, meaning the same permission sets, the same permission scopes that you give to a user, you could then apply to an agent. And so you have much more of that control of saying, "Yep, this agent can just view context" or "No, this agent can view or edit a contact."

And I, again, sound like a broken record, but we really believe providing that control is what allows admins to then configure their portals to the level of risk they find acceptable. And by making agents first-class participants within HubSpot, then we can have those defined identities, we can have that audibility, we can have those permission roles defined just like you would a human user and be able to, lets say, get a lot acceptance of the risks that you may have been concerned about if we wouldn't have given that extra control.

Josh Dougherty:
Yeah. So obviously this is something you're building towards. It isn't a current reality. What do you think the scope ... Which agents... Are these the agents that are available, customer agents, others that you're thinking of having those scopes or are there others as well?

Hank Lander:
Yeah, we're thinking this will apply to all our foundational agents and we will really be part of our philosophy moving forward. So if a new agent were to be created, it would follow that. And then we're also, because it's baked in at the foundational level, if you, as an end user, were to use Breeze Agents Studio and created your own agent, it would have those same controls. So this is going to be uniform, whether it is a first party agent, a second party agent, or even a third party agent that you're leveraging. So it's much more about giving you additional visibility and control over how agents leverage and use and manipulate the data within your portal.

Josh Dougherty:
Yeah, I love that. And I love it being a foundational piece. So even if it is third party agents that you're bringing in, it gives you the same framework to work within, which I think creates for ease of use across the board.

Hank Lander:
100%. And then the second part of your question around agentic reviews—and this is a little bit closer to going out to a private beta than the all agents are users—is exactly what you called out earlier in that when we talked to a lot of enterprise customers or more sophisticated organizations, I should say, they mentioned having the challenge of pairing quality with velocity where they want to move really quickly. They want to be able to create that blog or that marketing campaign really quickly, yet everyone is terrified of, let's say, having that subject line that has, let's say, first name in brackets or something like that and making a mistake like that or need to fill out later, it gets actually sent to our users or it just has incorrect information because it never went through X, Y, Z check.

And so this is where we think agents can actually take that first pass, read whatever the piece of content may be, check it for tone, check it for grammar, check it for spelling, check it to make sure there are no broken links, check the subject line, make sure it is something that would likely generate clicks because AI is actually really good at things that would drive engagement and do a set amount of things that you check that content against. And again, I'm going to say this so many times, give the user the control to figure out what are those steps that should be done in that agentic review so that when it does go to the human reviewer, you're working at, let's say, a base level of quality and you aren't going to have to do that, "Hey, there's a grammar mistake in paragraph one. Please rewrite that."

The bar of quality has been raised so that when the content editor is now looking at it, they can move with more velocity and ideally, get that out the door quicker, which is better for the customer and ideally also better for the person who's being engaged with as they're now getting something that is much more of higher value for them and their day-to-day, let's say. No one likes getting that email in their inbox that has no relevant information or no context. Now we'll be able to provide that at a much higher level of fidelity.

Josh Dougherty:
Yeah. I think this is an amazing, really, use of AI because even for people who are maybe concerned about using it to generate content, we all know that AI is an amazing reviewer and can collate content together and give good recommendations on an unstructured data set, which at the end of the day, looking across your other communications, comparing performance and then saying, "Okay, edit this email," for example, "in this way to make it more effective," it's going to be better almost than a human can do with their naked eye to do that type of work.

Hank Lander:
And AI is so good at context switching. So if you have multiple brands and you have different voices for your different brands, AI can keep that all straight in their head and know, "Oh, this email is for this brand. It needs to use this voice versus this email for this other brand." While we, as humans, sometimes struggle when we're context switching of like, "Oh wait, who is this for? Oh no, I put this in the wrong voice or I described the problem not the way that this segment describes the problem." And so AI is really good at doing things like that. I don't know how your evolution with prompting has gotten, but I know when I write a doc, I always ask AI, how can I make this better? What gaps have I-

Josh Dougherty:
Absolutely.

Hank Lander:
... left out? And it's really good at providing that feedback. And this is, again, where we think we can really leverage AI to not just write the content, but also just make it better so it's more actionable and more valuable to the end user.

Josh Dougherty:
Well, this type of functionality, and you can tell me if you're far enough along here, but how will it pull context? Is it going to use the same kind of AI context and knowledge vaults that are built into the system now, say for example, for Breeze, or is that how it'll pull context for your specific brand?

Hank Lander:
I would just say HubSpot, more uniformly, we're all going to start building on top of the same Breeze tool engine, we'll call it. And so it doesn't matter if it's, let's say this review agent or some other thing that's going to come out of HubSpot, we're all now building on a shared framework and tooling. And so you should see a lot of consolidation and where HubSpot is going as we release a lot of these capabilities. And with that, we believe that because we're crafted together as a platform, we'll be able to take better advantage than being cobbled together tools that you see in our product. And we actually think that's one of our advantages as an organization going through this inflection point in the market. And so we are really leaning into that as a differentiator. And so great question and hopefully, that'll come through not just in this feature, but all the features that we're releasing.

Josh Dougherty:
Okay, cool. Well, couple other questions. I think one more about, I know we got off talking about things that aren't necessarily around sensitive data but are very cool things that are coming up, thinking about the review cycle, et cetera. I'd love to zoom back in to audit logs and talk about how AI is impacting how teams interact with those logs and what improvements you're seeing in documentation. How is that changing with the full integration of AI into the sensitive data tools?

Hank Lander:
Yeah, I think this is a really fascinating question. Reflecting again back on where we were a year ago, we weren't talking about audit logs.

Josh Dougherty:
No.

Hank Lander:
Audit logs are like that thing that everyone needed to have for compliance reasons, but wasn't front and center of how you think about your product or your tool. Now audits are that paper trail for trust and it's the first resource people go to when something went wrong and needs to answer that question of who did this? What changed? And so audit logs are going through an evolution here as part of not an afterthought, but something like a secondary consideration to front and center of how we're thinking about that trust layer in empowering AI. And so with that, we are revamping our AI, investing a ton in it to make it a more discoverable experience, baking in search from the onset so that you can filter through those massive tables and figure out and really answer the question of what's changed and where are these data discrepancies coming from and do it in a more natural way than having to, like, what is this category? What is this subcategory? Let me filter to this. And then I think it was this user.

So really just reinventing that experience and making it more from a reactive experience to proactive of something unusual happened, let's flag that early so that it doesn't become a problem and can be course corrected prior to, let's say this is going on for multiple cycles and the change is getting worse. And I'll be honest, this is hard. Anomaly detection has never been easy. It's becoming even harder because there's more data, but this, again, is one of those unlocks that AI allows that just wasn't there even 12 months, much less 18 months ago. And so, we're really leaning into that to make it a better experience for everyone, whether that is an end user using a portal that has a higher level of quality and trust in the data to the admin that then administering it, the data is much simpler and it's more about, "Cool, I'm getting flagged when things are happening," or "If I find something, I can quickly investigate what is going on and troubleshoot that seamlessly."

Josh Dougherty:
Yeah. I think it's becoming more useful and I can see as you, in that future when you're able to roll out agent specific permissions to this audit log, become even more useful then too, because you have more clarity around what's happening and when you're using the AI's ability to look at unstructured data and understand it and comprehend it really quickly. And your final point about making sure you're alerting people when something happens, but only when something happens that is important, is essential to make sure that there's not too much noise in their day at the same time.

Hank Lander:
100%, and-

Josh Dougherty:
I can make it sound so easy and this is probably something that you're thinking about deeply with teams and engineers.

Hank Lander:
You just put six months of roadmap in two sentences.

Josh Dougherty:
Exactly.

Hank Lander:
I don't know what that says, but in all honesty, the philosophy of what we have of AI is a tool that people do use to make themselves more efficient or to be more productive. Well, with all tools, you have to have guardrail or checks.

Josh Dougherty:
Totally.

Hank Lander:
And us, we believe this trust layer that we're investing in and growing is the way to safely use AI and we truly believe in it and that's where we see ourselves going.

Josh Dougherty:
Love it. Well, as we close this conversation out, I think it's important for us to acknowledge this is a point in time. We're talking in June of 2026. This will be published in a week or so, but a lot's going to change pretty immediately. The rate of change is faster than it ever has been. And so I'd love for you, as we close out, to look in the future a little bit and say, what are you most excited about as the platform continues to evolve and grow? Obviously, you probably won't talk about specific functionality or specific things because a lot of what we're talking about today we wouldn't have imagined 12 months ago. And so maybe talk conceptually what you're excited and thinking about as you go forward.

Hank Lander:
Yeah. I don't even want to put any predictions on tape because I will be proven horribly wrong. But in terms of things I'm most excited about is AI is changing the way we work with our SaaS products and in my mind, for the better. But with that, it is making things like access management matter more than they ever have in the past. It also makes things like fixing mistakes and traceability more important than it ever has as we are trusting these agents with more and more tasks. And so, the whole governance layer has got to develop and has got to develop quickly to support those foundational things.

And in my mind, the way you delegate more to AI is by getting confident in the identity, the permission, the infrastructure, the audibility, the traceability of those actions that AI is taking. And so, when you ask me what excites me more, it's thinking about how does this evolve? What does a team look like a year from now? How do you bake in that agentic user with that human user? How do you make that mix happen? How do you create that orchestration level as we were talking about what the agentic reviews, so that we are enabling people to be more efficient and effective within their sphere of ownership and yet, have that accountability to know this is what the agent should do, this is what I am doing and lead that groundwork.

And so to me, it's really moving towards that hybrid vision where it's a collaborative experience versus one versus the other, or me creating tools for the agent or the agent creating tools for me. It's we're going to partner on this to create the best outcome possible. And I'm super motivated by that. We'll see how much of that transpires in the next 12 months, but that is really where we are investing and where I think the market's going and that's what I'm most excited about.

Josh Dougherty:
Yeah. Awesome. Well, thank you for sharing about the future. And I think also I want to really say I appreciate you coming on and sharing about complex and challenging things to describe and breaking them down in ways that are easy for people to think about and to digest. So appreciate your expertise and your willingness to share.

Hank Lander:
Of course. Happy to come back on and be a return guest. I feel very privileged. And thank you so much for having me.

Josh Dougherty:
How can people connect with you or follow along with the work that you're doing throughout the future here?

Hank Lander:
Yeah, I think the best way to stay in touch would be LinkedIn. My name is Hank Lander. I don't know if you could provide that in the show notes, but-

Josh Dougherty:
We'll stick it in there, for sure.

Hank Lander:
Awesome. Yeah, so I check LinkedIn regularly and always happy to have conversations with folks via LinkedIn and go from there.

Josh Dougherty:
Awesome. Well, thanks for the time and thanks for coming on, Hank. Good to see you again.

Hank Lander:
All right. Thank you.

Josh Dougherty:
Thanks for listening to this episode of A Brave New Podcast. Go to abravenew.com for more resources and advice on all things brand. If you enjoyed this episode, show us some love by subscribing, rating, and reviewing A Brave New Podcast wherever you listen to your podcasts. A Brave New Podcast is created by A Brave New, a branding agency in Seattle, Washington, that crafts bold and memorable healthcare brands. Our producer is Rob Gregerson.