Brendan Gardner-Young of Thrive Causemetics on AI, empathy, and automating half of support

“We answered every ticket ourselves”: the mission behind Thrive’s CX

Brendan (0:02) I am the COO and CTO at Thrive Causemetics. I’ve been with the brand since day one, so I’ve seen all the different evolutions of the brand and have been involved in all the technology decisions that we’ve made for a decade plus.

Lisa (0:13) How does that mission change what customer experience means to you guys?

Brendan (0:18) Customer experience to us has always been about how we have to approach it with empathy. We have to approach it with an understanding of what that customer is trying to speak to us about. Part of the reason is that early on in the business, Karissa and I both answered every single customer service question. And with all of that, we wanted to make sure that customers felt seen when writing in to us, and that we could recommend a product that’s actually going to help them.

So we started off with this ethos of: take care of them, even if we can’t take care of them. Now, today, we have a bigger team that does CX. We have Siena that we use. And we’ve tried to empower our team and Siena to still show up for customers in those ways. To know that someone could be going through something. Someone might be writing in to say that their family member is going through breast cancer, and they got products from a charity donation that we did. We feel like it’s our obligation to take care of that person in whatever way we can.

That’s still instilled today in the business, not only on the customer service side, but also inside the company, where people see those interactions with our customers and really feel like they’re part of that give-back mission.

Partners, not vendors: how Thrive picks its technology

Lisa (1:16) How is that impacting the way you evaluate your technology stack?

Brendan (1:23) They’re kind of opposing roles sometimes, but they work well together. One thing that I look at, as a COO operating the company, is that I want as little complexity as possible. You can see that in what we do in technology. We write almost no custom code internally, because we want our partners to be the ones that help us move into the future. I think that simplicity — the business has to operate, it has to be profitable, it has to make sense — flows into technology.

So we’re not the company that has a huge development team building out custom solutions to fit the business. As a DTC brand, we’re not a technology company, nor do we want to be one. We want to be enabled by the best technology, and my ethos has been that we’re not going to write that best technology internally. So we look to partners to do that.

Secondly, operations-wise, we use a lot of technology within the company. We have to, with the Shopify stack and all the different tools on top of it. We have always taken the approach — and this is similar to operations — that we have a bunch of partners, and those partners are all people that are really part of our team. We often tell a company when we onboard or sign a contract: you’re part of the Thrive Causemetics team now, and we’re going to give you feedback, and you’re going to work with us. We really want it to be a two-way street, to make the product better, and to make us better, and make it better for our customers. So we very much have an open line of communication with all of the partners we use.

Lisa (2:45) Yes, we partnered together on a lot of new developments of Siena, and we had the privilege to have you as a design partner. We always appreciated the direct feedback, and you guys knowing exactly what you want. That’s ideal for us. We’re always super excited when we get to work with brands that are open to that type of partnership.

Brendan (3:06) It’s a two-way street too, right? Sometimes we ask for things and you’re like, well, we thought about this, but this is why it might be better this way, and this is why we built it. So I think it helps us refine our thinking to be a better partner on that side too. We’re a paying customer because we believe in the product and we see it impacting our customers every day. But also, when there’s that two-way street — not just yes, yes, yes all the time, but some level of let’s work together to solve this the right way. The knowledge work, and some of the Slack messages back and forth of, hey, we’re seeing this, what should we be doing here — your team was really helpful in working through that, and then finding gaps so that we could build that together in the product and make it better. And we’ve seen that over time.

Lisa (3:44) It’s probably the same process that you follow with your customers, because you have such a big community and you listen very carefully. What do they want, what do they need, what else should we be introducing? And sometimes, you know, the customer doesn’t really know what they need, so you need to step in and drive that. It’s a balance.

Brendan (4:07) Yeah, and I think this is what our product development team does really well, our marketing teams. They have an idea of what they want to go build, but they also know that their answer is not necessarily the right answer. They’re willing to go ask questions. They’re willing to reach out. We have a community of early beta testers of our products, and they’re willing to reach out to those people and say, “Give me honest feedback about this. Am I on the right track for what we need?” That’s helped us in our product development efforts for sure over the years.

And also, we’re not going to make the product that every single person around the world is going to want, necessarily. We know that products work for different people, and that’s okay sometimes. As long as we can support them through that, then that’s great.

Lisa (4:40) Has AI accelerated your product development?

Brendan (4:44) It has accelerated some of our decision-making, and it’s accelerated some of our research. It has gotten us a little bit more organized, I would say. We do internal “build with AI” sessions, where it’s like pair programming, almost. And some of the team members that have really taken to AI — they have so much knowledge and expertise in their areas, and then they get to add on this other tool to help turbocharge their work a little bit, just to get to better decisions or better research, or what’s the latest thing here.

So we’ve definitely seen the benefits of pushing forward using AI and using research to make it so that we can be on the right track faster — without going overboard, like “we’re going to build everything internally with AI,” which we’re not. I’ve seen people take it a little bit too far.

The problem before Siena: the same answer, every time, at volume

Lisa (5:25) I’m curious, before you started using Siena, what was the problem that you were trying to solve, if you remember?

Brendan (5:33) The problem we were trying to solve was, one, consistency of our answers. We were on Zendesk for most of our support tickets. We were also on a different one for some of the social ones. And customers almost got a variation of answers based on macros and based on what we had in the knowledge base, but it was something that was ever-changing. We wanted it to be consistent. If you reached out one day, and then you reached out a month later, are you going to get the same answer?

And we found that Siena, working with our internal customer service team, actually enabled us to find better answers faster, support those customers, and also hand things off to our team members when they need to be handed off. We get everything from the typical “where’s your order?” to “hey, I have blue eyes and this skin type, let me send you a picture — can you help me with some makeup looks that might look good for me?” And everything in between. We really need a mixture of Siena plus our team members. It’s almost like Siena is a team member alongside those CX team members, in how we’re servicing those customers.

So the problem we were trying to solve was that we have a ton of volume, especially during peak times. It was inconsistent, in the answers that we were giving and the timing we were giving, to service customers. And we needed to find a way to break through that, to make sure that we’re reliable for those customers, because they trust us to get back to them in the times that are expected.

Two criteria and the moment it clicked

Lisa (6:45) And what were some of your evaluation criteria?

Brendan (6:48) One, it couldn’t be generic. We knew that it had to be brand-specific. We needed to be able to customize it as much as we possibly could, in order to make it follow our brand guidelines and what we wanted it to be. We really love and care about our Thrive Causemetics brand, and we really care about how it makes people feel and how it makes people think — not just receiving the products, but how anyone from the brand interacts with those customers too. So one was brand control. We really needed it to be in tune with who we were.

And I guess the second one is learning. It needed to be able to pick up on things that customers were saying, and then be able to adapt — to answer more questions, or answer them in a better way, or suggest whatever things needed to change. We obviously had a human loop that we closed, to say, “Hey, an agent saw this ticket, we need to go back into the knowledge base and update this over here.” But having Siena plus the agents work together on that knowledge base has been a game changer for us, because we can adapt it so much faster and support our customers.

Lisa (7:46) When was the moment that you realized, “I think I made the right decision”?

Brendan (7:48) I can’t remember the exact month after we onboarded. I think one of the moments was when we were reviewing — it wasn’t even deflection rates. It was something around customers that had been serviced in one touch, and then the timing it took for total ticket closure on the ones that got past Siena and into our queues. We saw immediate improvements in both of those. The timing of us being able to follow through on SLAs, and beat SLAs, was the first time we sat back. We had a pretty high rate that would get answered in the first answer. And so that’s when it was like, okay, this is something.

Also, the deep linking into the services that we use was something that was really important to us. We’re on Shopify — we were on the $19 a month Shopify plan 13 years ago, literally just the starter plan. Because of that, Shopify is the only thing that we know. So we always try to choose the best software solutions in the Shopify ecosystem that fit us. And we saw that you all were integrated into almost all the applications that we needed. We used Yotpo, we used Recharge, we used a bunch of other tools that you already had in there. Going back to the criteria, that was a big thing for us: if you can go into those individual applications and make changes too, then suddenly it’s really, really helpful for us to be able to surface that information.

Using AI vs. being changed by AI

Lisa (9:04) What do you think is the difference between a brand that’s using AI and a brand that’s genuinely being changed by AI?

Brendan (9:10) Using AI, to me, is typing questions into the chatbots, getting some answers, putting together some reports. Being changed by AI — if we use our brand as an example, it’s the team members that have had the aha moments. We go from not knowing how to use Claude or ChatGPT to its full extent, to two weeks later delivering some connected thing that’s pulling in data from one source and another source, running a report, and then giving it to another team.

I’ve seen so many people that are on one team but have built AI things that actually help other teams, just because they could, and they knew it was a problem. I think those are the ones that are getting rebuilt by AI, in some ways. You can be in service of not only your team — you start to help out other teams because you see opportunities to help them.

We’ve seen that across the board. Not that there aren’t teams anymore, but I have someone in finance that is heavily involved in product development sometimes, because it’s like, hey, I can help out with how we bring these new products to life and what the P&L of it is. I have someone in e-commerce that consistently wants to work with the performance marketing creative teams to get into ads and prove out some things. People have crossed lines in big ways, and I think that’s really helpful from a business perspective. It’s changing. It’s kind of like a built-with-AI company, I guess, because the lines of what your job roles are blur.

Lisa (10:28) What are the new skills that you would look for when you hire?

Brendan (10:32) It’s a lot of the skills that you look for anyway in hiring, right? We use “humble, hungry, smart” by Patrick Lencioni a lot in the company. We look for those traits in people. If they can be all three of those things, then it’s someone that’s going to be adaptable, be able to pick up change, be curious about the role, step into new roles potentially. We can definitely see it when people want to do that, and they have that hunger to learn and change and grow.

What are those new skills? I mean, change can be hard, but change can also be really exhilarating. It’s like, wait, that thing used to take seven hours, now it’s an hour. Okay, what else can I do now? So I really try to talk about AI as: what are the opportunities that we can go capture? Not how do we automate things away. To us, it’s more about how we reach more people, and what are our ways to not just grow, but to expand our giving mission, which we really want to be bigger than it is today. How do we do that faster, potentially, or in a more thorough way? That’s something we think about a lot. Expansion, rather than contraction, due to AI.

Siena Intelligence as an early-warning system

Lisa (11:30) Besides what you already shared about the impact on some of those metrics, like response rate and consistency, what do you feel has been the next big benefit after adopting AI in CX?

Brendan (11:43) For example, we’re using Siena Intelligence right now. It’s reading through all of the tickets that come in, and it goes into Slack. We have an automated report from Siena Intelligence that gives us early detection of the velocity of tickets — do we need to look at a potential product problem very early on? That sends directly to our product development team, so we can catch things way earlier than we potentially would have before, and understand: was there something wrong with the website, was there a quality issue, was there something that could impact customers? We now know it faster, basically.

With that, when we do sales data querying and whatnot, we’ll also ask Siena what’s going on, just to get a qualitative insight versus the quantitative. To be able to say, this is what happened, this is what we’re seeing, here’s our hypothesis. Can we ask Siena, and can we ask our data warehouse, what might have happened, and put together a report for it? So we’re pulling together early signals to make better decisions, and we’ll use Siena data to do that too.

Lisa (12:35) Who are the main people, or what are their roles, inside of Thrive that are leveraging Siena Intelligence?

Brendan (12:42) A couple of team members work on Siena directly — tuning it, looking at it, seeing the conversations. Siena Intelligence specifically: our e-commerce team uses it consistently to identify what website-related things are going on, what we can do about that, what people are struggling with. The product development team uses it consistently. Sometimes marketing will go in there and ask questions like, we have five shades of this one product and we’re thinking about more shades — is there any demand there? Are people talking about a purple shade of this product? So we’ll use that for early product-development-related things.

There are a number of teams that will go in to ask those questions. For us, with so many tickets, it’s almost impossible to keep up with all of them. What’s really the pulse of what’s going on, day after day, week after week, month after month? It’s constantly evolving, and monitoring that much data is also very difficult to do. So that’s where it helps us out, just volume-wise. We need to know the pulse of the words and the sentiment of what people are talking about.

Lisa (13:34) Before that, were you doing that manually, or what was the process?

Brendan (13:39) There would be some reporting out of Zendesk, out of the tool, or out of the other social tool that we used. We would pull sentiment from NPS-related stuff — if we do NPS questions, the answers to those things. We would do CSAT, which we still do. Those are all still super helpful. We just have way more data now to query with those things, to be more accurate, faster.

AI-native partners and the real cost of a fragmented stack

Lisa (15:21) How are you thinking about AI-nativeness? A lot is changing. I feel like there are new vendors almost every week. How are you thinking about it for Thrive?

Brendan (14:09) We’ve used a lot of different vendors over the years. A decade plus inside of the Shopify ecosystem — we’ve tried a lot of them. I gravitate towards the software solutions that want to build with us, at the end of the day. I think there are going to be fewer overall that we will work with and have direct relationships with, because you can build so quickly now, especially if you’re AI-native. But you still need the partnership to determine where that roadmap is, and whether it’s right for us or not.

What we’ve seen is that we have gravitated towards the ones that have the Slack channels with us, have the conversations with us, proactive reach-outs. “Here’s our roadmap, let me share it, let’s work through these things together.” The software vendor that says, “I know you said this two years ago and we didn’t build it yet, but we just released the alpha of this — can you take a look at it?” and circles back with our team who had been asking for something for a long time.

To me, the software teams that are really organized, that understand they’re building for customers, that are asking questions, that are initiating sharing sessions to get true feedback — they’re going to be the ones that we really continue to double down on and work with. We see the capabilities of all software companies getting larger and larger, and so we want the partners that want to build with us, rather than just build at us, following a roadmap without any input.

Lisa (15:21) How are you thinking about AI solving for both support and sales?

Brendan (15:24) I could see how it would be support and sales. We treat customers like they’re going to be around for a long time, no matter what — even if they’re just one-time buyers — because we want our cosmetics to be around for a hundred years. A long time. And because of that, we almost go over the top to be as supportive as we possibly can.

Sure, sales are important. The profits of the company go towards being able to have the big giving mission, to support so many products donated, to grow, to be in multiple countries. But supporting customers, I think, is still our top priority. We’ll answer customer calls that may not even be related to our products. We do all that stuff because we’re going to be around for a long time, and we hope there’s an impression of this brand that makes customers, or non-customers, want to try the brand at some point in time, or be loyal to it. As a result, with AI, we tune it to be empathetic and understanding.

One thing that we do that is really helpful is making sure that whoever is chatting with a customer — whether it’s Siena or one of our agents — has as much information as they can about that customer, if they’ve given us that information. Because you can show up in such a better way. Personalization can be a great thing for the interaction that you’re having overall, especially if you are doing sales. So the more integrations, the more data that we have readily available for that interaction, the better off we’re going to be for the sales process, and the better supported that customer is going to feel too. It’s super important to us.

Lisa (16:42) What do you feel is the real cost of a fragmented tech stack, especially in the AI era?

Brendan (16:49) I think there’s definitely some real cost. We’ve seen having to build multiple “AI brains,” let’s call it. The “what is the brand” has had to be replicated in so many different areas. In my mind, in a perfect world, that lives in one central spot and is then replicated to all the different software vendors. So there’s definitely a tech cost to that. Is it still maintained if it’s in a system that we haven’t been in for three months, because it’s been answering things fine? I’m not saying this is Siena, necessarily, because we’re in there every day. But there’s definitely a cost to that, for the team and for the maintenance.

The interoperability of the data now makes it so that you can connect so many services if you need data between two different systems. So it feels like the barrier is going down. But if you’re going to be AI-first, or curious around AI, you also need a couple of core central systems that you’re building on no matter what, and then the others can tie into them, almost. To me, you have to have those core systems that you probably aren’t going away from anytime soon, and really invest in those as we move forward — or else you just can’t switch your major systems. So there’s definitely a cost. There was a cost before, with so much SaaS that we had and all the different services. It was high for the team to maintain all those things.

Should CTOs own CX?

Lisa (17:56) Do you think CTOs should own CX?

Brendan (17:59) I think there’s a world where it could live under marketing. I think historically, when it was more like call center operations, they had a lot of CX. But if you’re really looking at it as a relationship and a partnership with the customer, then no, not necessarily. I think the CTO is the enabler of that from a technology perspective. It just still happens that I wear a lot of hats. But no, I think there are a lot of teams that are very customer-centric that could own, and should own, CX in companies.

Lisa (18:26) I have been seeing a lot of brands making this change internally. That’s why I’m asking, especially now that AI is in the picture. Typically, engineering teams, more technical teams, would own that part. But now I think it should be an expectation across the board: regardless of where CX sits, you need to be AI-curious, AI-native, around all that. But also know who you’re talking to. The AI is talking to humans at the end of the day. So that person leading it needs to be so in tune with what those people are saying about your product — good, bad, feedback, whatever — so that doesn’t get lost.

Brendan (19:01) Whoever leads it is the person for it. But we can’t just build, build, build without ever talking to the customers and seeing what’s going on.

Learning AI is like learning to paint

Lisa (19:06) I think Thrive, and what you guys are doing, is very advanced in terms of AI versus what I’ve been seeing since working in e-com, and since starting to build Siena and working with hundreds of brands. What’s your take? What do you think most brands get wrong about AI? Why are they still very much on the fence about it? What do you think is stopping them from restructuring their companies around it?

Brendan (19:30) I think leaders are one thing. If your leaders, your executives, are not really dialed in — not saying they need to be AI experts, but if they’re not encouraging it in a way that says this is something that can benefit the company, and it’s not coming from a mandate sense, but more from a curiosity sense.

What I tell the team internally is that I think AI is almost like learning how to paint for the first time. You don’t just get a blank canvas and start painting this beautiful picture. It’s going to take a lot of effort and a lot of practice and a lot of iterations. And I think that’s one of the missing links in a lot of organizations. It’s not just “go to a two-day training and become an AI expert.” What we do a lot is weekly sessions, where it’s me and you with a laptop, and I’m writing a prompt, and you’re asking, “Why did you write that prompt that way?” And you start to understand a little bit better. We build together.

Going back to the painting analogy: I have different levels of painters. If I’m a master painter — which I’m not saying I’m a master at AI — but if I have someone that’s at level zero, they need to go from zero to one. And how do they go from zero to one? Well, let’s get all their connectors set up in Claude for them, and let’s show them the power of a couple of key questions. Let’s show them the power of Siena Intelligence and what they can ask to help out in their job.

So I think when it comes to getting people in companies excited about it, it starts with the leadership giving them the why, and then showing them the path, and then encouraging along the way. I think there are some missing steps when it’s just “we need to be using this” — in what ways? And these are assumptions. I talk to friends, but we’re very focused just internally on us. Even with us, sometimes I feel like we’re behind on AI.

Lisa (20:58) Everyone feels that way.

Brendan (20:59) Yeah. And I talk to friends and they’re like, “Oh no, you’ve built these cool things.” My friends will have working sessions sometimes to ask, “What are you doing in your business?” The team has a really positive attitude towards it, which I’m really grateful for. I know that’s not the case for a bunch of different teams. But we’ve really tried to foster: this is here to help make decisions faster. Abundance. Abundance is what we talk about a lot.

Lisa (21:20) And I love that your team is always open to joining any workshop that we might host, even if it’s outside of Siena, asking so many questions and being curious.

Brendan (21:30) It’s also making space for people to be able to do that in their jobs, right? To encourage them: go learn that thing over there. There’s some way it’s going to translate back into the role somewhere. We try to encourage learning. We start every “build with AI” session with a sharing session. It’s like a show and tell. And someone’s like, “Look at this amazing thing that we built.” We didn’t even think that was possible, but that’s amazing.

Lisa (21:54) The biggest challenge that I’m seeing is that they think it’s an overnight success, and they are all about short-term ROI. But with AI, it’s a long-term investment. It does take a while to actually see the results, and you need to put in the work. What’s your take, your experience?

Brendan (22:11) I mean, budgeting is new at the end of the day, right? When I enter into contracts with software, and AI specifically, usually we’ll have metrics before we say yes and sign on the dotted line. It’s almost a hypothesis: after X amount of days, we expect to see this. We focus it on: what are we going to see in 90 days? What do we see at 180? What should we see this year? It doesn’t need to be perfectly correct, but it should be directionally correct overall.

And it doesn’t always need to be “we’re going to replace this many tickets, and this is our cost savings” only. We try to think critically about time saved, customers that would have potentially lapsed that are actually going to buy from us now. From a CX perspective, we just looked at all the factors to say: would this make sense, or would this not make sense? And sometimes with solutions, we’re just like, let’s give it a go. Let’s do a pilot. Let’s try it. Let’s do one small step and see what’s going to happen, with some hypothesis to prove it out — yes, no, or maybe.

In our space, moving fast is one of the names of the game. You have to do that. But it’s weird, because you implement and you move fast, and then you have to almost slow down a little bit, settle, and say, okay, how does this actually fit into our work going forward?

And then I think the other thing for budgeting is we work with our partners a lot to say: we want this to be a win-win partnership, and how do we make this into a win-win partnership? Ask it that way. It doesn’t just feel like a sales pitch on one side and negotiating down on the other side. It’s like, look, we think you have a great solution. We really want to work with you. We expected this over here, potentially, but walk me through, with some transparency, what we’re doing in order to work together and build something greater.

Often, on the software side, you’ll have a vision for what it looks like a year from now, because you’ve seen hundreds of brands go through it. And we’re just guessing on our side. We might have asked a couple of friends or seen some forums, that type of thing. So inviting in the partnership, getting away from the sales, and more into: what would it look like over a period of time for us to succeed here? What should we expect? And have those conversations longer term.

The metrics behind 50% automation

Lisa (23:58) So you are automating around 50% of your volume. I’m curious how you are tracking it. What are the most important metrics for you, that you would look at almost every week, to evaluate the success of your AI agents?

Brendan (24:12) Resolution — is the automation serving the customer’s purpose? We would do that in QA, though. The person that’s tuning Siena would be really looking at: is this the way that we would answer it, and is this what the customer expected?

And then second would be: does it even out our peaks a little bit? If we have a huge sales volume weekend, do we see the huge blip a couple of days afterwards, or is it leveled out, so we have less volatility overall in tickets? That’s the big help, because before, it would be: do we need to staff up? Do we need to staff more hours? Do we need to do overtime? That type of thing. And now it’s much more stable, more expected.

And then some of the other metrics we might look at: how fast we’re closing tickets overall. We look at CSAT scores a lot. So, are the scores of Siena consistently at or above — depending on the flow, obviously, the different flows that we have live have different ratings — the scores that we expect them to be at? We’ll look at that consistently, because quality is much more important to us than automating away 50%, 70%. Quality is the most important thing for us.

Lisa (25:12) Biggest benefits — what about, if you want to share: cost savings, team impact, projects that you were able to take on with the freedom or space that it gave you?

Brendan (25:22) Yeah, I mean, we saw a lot of cost savings overall with it. When we looked at Siena answering questions versus the team, there were some that were just more well suited for Siena. So we definitely saw, with the reduction overall, an ability to focus our team on the higher-impact questions that people might write in for. If we have makeup artists and estheticians on staff, well, suddenly they are more often answering makeup artist and esthetician questions, and they can be better stewards of the questions they’re answering there. The questions that Siena took on were often ones better suited for Siena to answer, with all the systems access and everything there.

And then secondly, we were able to look more critically at the systems that we had that were supporting customers, and ask the questions: do we really need all these things? Do we really need all these modules? Also the strategic questions that are longer term. We were able to finally settle and say, okay, what’s our strategic direction going forward? Is it this ticketing system? What do we do with social? We started getting asked some of those strategic questions. And that, to me, is space inside the team to be able to dream a little bit about what the best state to be in is, rather than just a steady state of constantly answering questions. Anytime a team gets into that dreaming kind of thing, it’s always a good time for us.

AI-native or add-on: advice for brands on the fence

Lisa (26:32) Thank you for sharing that. What would you advise other brands that are evaluating their AI stack, or their tech stack, at the moment, and are on the fence between going with an AI-native solution like Siena versus maybe an add-on that their ticketing system might offer?

Brendan (26:50) I think we’re seeing it consistently: the AI solution oftentimes is maturing quickly enough to be a suitable solution, even at high volumes, for us. So we will look at the ones that are more AI-native as: we can build with them into what we need, if it’s not there. But oftentimes, they’re really full-featured already.

We’ve honestly struggled with a lot of the more enterprise-y type solutions that are moving slower, that are missing things, or trying to build add-ons. It’s been a struggle for us from an AI perspective, and also just from the pace of change. The pace of development has sped up so much, and the expectations around that, that we see the large difference between the providers that are still working at the old pace, and the ones that are rebuilding themselves in a way that feels like they’re keeping up. Honestly, we want to be with the partners that are going to keep up in the future. We can’t lose some of those competitive advantages that we potentially have, by being with the right software partners and choosing the right tech.

Lisa (27:43) Velocity is everything. Thank you so much for sharing your experience, Brendan. Really appreciate it.

Brendan (27:48) Thanks for having me.