145. How to Transform Customer Experience at Scale ft. Jeffrey Bussgang & Teresa Anania

What I love about AI is it allows us to do QA at scale.

[Jeff Bussgang]
In a way, we’re going to do business, we’re going to focus on efficiencies, and we’re going to focus on running these experiments to become AI native.

[Teresa Anania]
When I first joined, I will say there was a lot of resistance in support.

[Josh Schachter, Host]
Okay, there’s this fear around labor displacement and replacement, right?

[Jeff Bussgang]
Showing the art of the possible and allowing people to imagine what the future state could look like.

[Josh Schachter, Host]
Where do you start that conversation, helping these leaders break that down?

***

[Josh Schachter, Host]
Hey, everybody, and welcome to this episode of UnChurned. I’m your host, Josh Schachter.

I’m very excited to be here today because I have two, two for the price of one, wonderful guests on the show today. And we’re going to have a little happy tension between theory and practice. So first off, I want to introduce Jeff Bussgang.

Jeff is the co-founder at Flybridge Capital Partners, one of the preeminent VCs. He’s based in the Northeast with a bunch of his firm there as well, focusing largely on seed investments and really taking a strong focus on AI in his investment thesis and his portcodes these days. Jeff is also a lecturer at Harvard Business School and is an author.

And his most recent book is The Experimentation Machine, How to Learn to Leverage AI to Become a 10X Founder and 10X Executive, in the case of some of our audience in this program. So, Jeff, thank you very much for joining us. Thanks, Josh.

Great to be here. Yeah. And Teresa, second time guest on the program.

So good that we thought we’d have her back again. Teresa Anania is the chief customer officer at Sophos, one of the preeminent cybersecurity companies. They are doing some cool things with AI in their product.

And she’s also at the Frontier Frontlines, figuring it out like all the rest of us of how to use AI as an enablement platform across her post-sales organization. So, Teresa, thank you so much for being here.

[Teresa Anania]
I’m excited to have the conversation.

[Josh Schachter, Host]
So, like I said, we’re going to start with a little healthy tension here, really talk about the theory and then kind of bring it into some of the practice practicality. Jeff, I want to start with you. You wrote, first of all, just tell us a little bit about the experimentation machine, and then we can go and make that relevant to the rest of the conversation.

[Jeff Bussgang]
Yeah, well, as you said, Josh, I work in two realms. I’m a seed investor of AI software companies. In that work, I’ve seen a ton of my founders leveraging AI in the product market fit journey.

And then in my work at Harvard Business School in the entrepreneurship faculty, I’ve seen all these students using the AI tools to go through the ideation and MVP process. So I’ve really seen really a sea change in how startups are being built. And I saw that cascading not only through startups, but also growth stage companies, a little bit larger companies like Sophos, and then even very large tech and Fortune 100 companies as AI is rippling through the entire ecosystem.

So that’s really the essence of the book is leveraging the modern AI tools in combination with timeless techniques for finding product market fit, focused on founders, but relevant really for all executives.

[Josh Schachter, Host]
Sam Ullman recently had a quote, it was something along the lines of where, I think, do you know where I’m going with this one? He was saying that we’re entering the fast fashion era of AI and all these tools and apps. I suppose that what you wrote and your principles and whatnot here are aimed at defending against that, right?

Setting founders up for the right fundamentals, so that it’s not just a flash in the pan type of experience they have.

[Jeff Bussgang]
Yeah, one of the things that I say in the book is, AI may help us get to a destination faster, but you want to make sure you get to the right destination. You still need to build an enduring, valuable, profitable business. You have to have business model fundamentals.

You have to have a good idea. And in many ways, and I call this, you know, kind of where the humans are inserted into the process now. In many ways, it’s up to us as humans to do that strategic thinking, discernment, judgment, taste, really decide what experiments we want to run, what kinds of companies we want to build, what customers we want to serve.

That’s really going to be driven by the discernment of humans. So it’s a really awesome moment for entrepreneurship to flourish, but it still is grounded in human ingenuity.

[Josh Schachter, Host]
Yeah, yeah. So let’s talk about more on the human side of things and building culture. So at an organization a little larger than your typical, you know, seed series, a startup, but let’s say a Sophos type organization, we’ll talk to Teresa about her group in a moment.

What are the conversations that you’re having with those leaders? What are some of the themes that you’re seeing around their wants and anxieties and desires towards AI?

[Jeff Bussgang]
Yeah, and look, I spend time with our more mature portfolio companies, we’re early investors in public companies like MongoDB, very large private companies like FalconX. I also get pulled in through my HBS activities to talk to CEOs of very large companies. They’re all trying to do the same thing, which is create an AI native culture and allow for a ton of AI experimentation.

I was talking, for example, to top 10 fortune, you know, 10 company, very large insurance company, who is a very close strategic partner of ours. They’re running hundreds of AI experiments right now throughout the organization, and it’s permeating everything they do. It’s not just software development, it’s not just in the IT organization, but it’s in customer success and sales and marketing and finance, operations, investor relations.

So really being thoughtful about how do you create that AI native culture? There have been some very public posts by a few CEOs, like the Shopify CEO and the Duolingo CEO, talking about AI is the way we’re going to do business. We’re going to focus on efficiencies, and we’re going to focus on running these experiments to become AI native.

So how do you actually do that? That’s really, I think, your point at the beginning of theory versus practice. It sounds good, but in practice, if you’re running a multi-thousand person organization with some history behind it, how do you turn it into a truly AI native organization?

[Josh Schachter, Host]
So when you’re talking to these folks, that’s the first thing that really comes to mind. It’s like, OK, Jeff, how are we doing this? We know there’s appetite.

We know the promise of what AI can deliver for enablement and these efficiencies, and top-line gains, but how? Where do you start that conversation, helping these leaders break that down?

[Jeff Bussgang]
Yeah, there are a couple areas where you have to begin. One is with data, because you as a company have a set of proprietary data, and there’s the classic garbage in, garbage out problem. You can use these publicly trained models, public data set trained models to some extent, but you really want to point it to your private data.

So you’ve got to have your data in the right shape to feed into these models. Secondly, you need to have the right sequencing and prioritization of what applications you’re going to focus on, and I’ll come back to that, and I’m sure Teresa has a ton to say about that in her own experience at Sophos. And then third, it’s about culture and people.

A lot of us are running companies that have existing processes and existing people who have been trained to do things a certain way. There’s some cultural resistance, perhaps, to doing things differently, and cultural resistance to experimenting and failing with new tools, cultural resistance to setting up pilots and ingesting new tools, and then finding a way to get those things into production as rapidly as you need to. And so that’s really what we spend, what I spend a lot of time talking to the larger CEOs, C-level suite companies about.

[Josh Schachter, Host]
Yeah, so getting your data, so three things, right? You said private data is really the key here. So how do you get your data in the right place and position to help you facilitate these experiments, sequencing and prioritizing those experiments, because you can’t do everything at once, and then the culture and the people.

And it sounds like the third of those, the latter, is where you’re focusing your time, helping to guide them. So yeah, so how do you build a culture, a digital or native AI culture?

[Jeff Bussgang]
Yeah, you’ve got to celebrate wins, you’ve got to celebrate when experiments are run, be okay to celebrate failures and create a culture where failure is allowed, but also have the hackathons, the AI entrepreneur, AI innovator of the month awards, the spotlights on what are the new application areas that you’re doing. You have to make it a part of your scorecard for people in your evaluations, how AI native on a scale of one to 10 is the staff member. You have to make it a part of your interviews.

One of our portfolio companies, Brighthire, which is an AI native intelligent interview tool that’s now sweeping through many, many companies. I don’t know if Sophos uses it or not, but Brighthire tracks matching job descriptions to the interview content and sees a gap between people who say they want AI native new staff members, but they’re not testing for AI native skills, asking about apps that they built or agents that they’ve constructed or initiatives that they’ve run around AI native activities. So really making it a part of the culture means rewards, incentives, dialogue, celebration, and having it really be featured in all of your systems, in your human talent systems and capital management systems.

[Teresa Anania]
Yeah, but I would just add to that, Josh, if I could, that those elements are so critical, but I also find that it’s really important to honor and empathize with the fear factor, right? With the impact that many of our employees think AI can have on their own jobs. I mean, we hear this all over the news, but it is real.

And, you know, the curiosity around what will this mean for the end state, or at least how we leverage AI to even deliver a better customer experience, frankly, has been at least top of mind in my organization.

[Josh Schachter, Host]
Yeah. Well, okay. So how has that manifest in your org?

How has it come to you or the leaders that sit beneath you? What types of anxieties and concerns have you heard?

[Teresa Anania]
I think the first impression is, especially when you talk about that external use case of digital agents, you know, where you are literally able to eliminate, and I don’t mean only for support. I mean, even for the renewals function, the sellers function, you’re able to eliminate that, what I call tier zero, tier one, repetitive, kind of menial work. The initial impression is, you know, what’s going to happen then to our hiring plans, to our growth plans.

And I just recommend that bringing those issues to the surface, having regular, either small group sessions, or I just came off an AMA and ask me anything literally on the topic of AI, so that our teams understand that while we’re removing some of those menial tasks and frankly, making it better for that external inbound customer, for us, it’s a combination of the external customer, but it’s also our reseller channel, who is also a customer in our processes. It’s not only improving their experience, we elevate a lot of the then skills within our own teams to actually address more advanced demand that is going to require the humans. And then Josh, we also focus a lot on the internal use case of the agent productivity.

Again, agent equal to persona, renewals rep, support engineer, success manager, sharing with them that we’re not only going to elevate a lot of their skills to be really able to focus on higher value engagement, but also we’re going to be helping equip them with more at their fingertips to make those engagements higher value and give them kind of more enjoyment at work.

[Josh Schachter, Host]
I mean, that’s the theme is you talk to leadership, myself included, in our organization and the customers that we speak to, and it’s, okay, there’s this fear around labor displacement and replacement, right? Rightfully so. I remember years ago reading Andrew Yang’s book, The War on Normal People, that sort of thing.

And I think those fears are all real concerns. But I do believe that you’re going to be able to push aside some of the more menial work that’s quite frankly, not as interesting to people, right? And have them focus on the more strategic, like you said, higher value activities.

And it makes sense. It makes sense. Now, that being said, that’s coming from you.

That’s coming from me.

[Teresa Anania]
Right.

[Josh Schachter, Host]
It’s coming from Jess, from leadership. How are you, like, what’s the vibe you’re getting from your team around that messaging?

[Teresa Anania]
Well, almost back to what Jeff was saying about the need to celebrate wins is really bringing some of the actual agents to the table to share how much more enjoyable their job is becoming on this journey to basically debunk and demystify the impact that AI can have on the organization end to end. So bringing them into those success stories, not just having me say it. You know, frankly, I have a driver in doing this that really, it exceeds some of the day-to-day needs of our people.

It’s important that we address those. But I am really trying to optimize to scale because we have a growing customer base. We’re at 600,000 customers.

And we’re, you know, growing in double digits each year, each quarter. And to be honest, at the leadership level, you know, I need to bend that cost curve. We can’t just keep throwing more humans at, you know, this growing base of customers, especially the long tail.

So while that may be my driver and when I’m in front of the board, I’m sharing our progress against that, you know, in the trenches and with the actual people doing the work, it’s more about showcasing their typical day job. It’s like the from and the to. It’s like the before and the after.

And when you could show real pragmatic examples where our own support engineers were doing swivel chairing, they were opening up five different systems to even imagine how they might respond to a very complicated, you know, answer that is required in response to our, you know, customers. When you can share with them that they can handle, you know, 10 customers but go deeper than they were ever able to go per day, that resonates. Because, I mean, they took a job because they want to have more customer engagement, less administrative research in the background.

So those are some of the use cases where you showcase the before and after that I think really resonates with the team.

[Jeff Bussgang]
I want to react to something that Teresa said that I think is really cool, which is showing the art of the possible and allowing people to imagine what the future state could look like. Most companies wait until they’ve got everything figured out and then do the long build cycle and then complete the build and the test and then show the art of the possible. What AI tooling allows us to do now is Teresa herself as an executive who probably doesn’t have a deep software development background, I think from your background, can do the vibe coding with the modern AI tools and demonstrate the art of the possible just by speaking to one of these AI systems, whether it’s a replit or a lovable or a V0, and create the art of the possible and that vision and show it to not only her IT team and product team, like this is what I want you to build for our use case, but also to her own team and get their input and get their buy-in and get their involvement more and more into what the tooling is that they want to see.

So I think it’s a really cool moment where we’re at a moment where we can articulate vision of the future, but then also build and show that vision very efficiently and rapidly without having to go through these long build cycles.

[Josh Schachter, Host]
Teresa, I think Jeff is proposing that you vibe code. How do you feel about that?

[Teresa Anania]
It’s a new one for me. I love it.

[Josh Schachter, Host]
Yeah, yeah. Vibe code, some case studies.

[Jeff Bussgang]
All you have to do is create an MVP, a pilot, a prototype, a few screen flows, just showing you the demo.

[Josh Schachter, Host]
Get on lovable. Yeah, yeah. So, but Jeff, as Teresa’s talking about the, you know, demonstrating the from, to, before and after case study, side swiveling and stuff like that.

Uh, what are some of your recommendations or, you know, probe on her a little bit here of like, you know, how can her organization create that best view of what they’re doing to motivate their people? What are some of the tactics?

[Jeff Bussgang]
I think the notion of buy-in and being sensitive to bringing people along, that’s really such a core leadership role. It’s not only, you know, painting the vision, but it’s also creating the process to bring people along and give them a sense of ownership and buy-in, whether it’s your middle managers or your end users. If the vision is we’re growing and we need to be more efficient and more effective in serving our customers as we grow.

If the vision is we have a bunch of clunky tools because we grew so rapidly and organically, and now we have a chance to step back and reimagine what the tooling should look like. If the vision is, hey, we’ve got an end state in the future where you’re going to have more resources. Those resources just happen to be, instead of people, AI agents, but it will allow you to be more effective.

And as I said, and you noted in the language of my book, a 10X professional, a 10X customer service or customer success rep or executive. That’s exciting. We all want to be better at what we do.

We all want to be more efficient. We all want to have our time freed up to do more meaningful work. So I think grounding on those principles is really the way to get the cultural buy-in, as Teresa was saying.

[Josh Schachter, Host]
Teresa, in your AMA that you just had, what you can share, what were the key themes of what came up from your folks? What questions?

[Teresa Anania]
Well, I mean, it was job security. What does this mean about hiring? Because there is a culture in many companies where the humans are, you know, viewing this AI as a competitive force rather than recognizing a partnership.

You know, so what I really focus on in that theme is the agent, I call it the agent assist, the use case of internal, how you equip those frontline goals with all the information at their fingertips to do a better job by that customer, that reseller, or that distributor. So I think that’s a big theme, is this fear that it’s a replacement, it’s either or, when really it’s both. And even where we’re using a digital agent to interact directly with our customers or our partners and channel, there is a human review of a lot of what we’re using in our chat bot, in our content email, you know, recommendations, in the workflow recommendations, which is continuously making this large language model better, but it’s also making our humans feel less afraid of AI as an either or, but rather a compliment to the humans. So that’s a big theme.

I think the other big theme is, you know, wanting to understand the expectation and how we’re meeting it on enabling them and giving them the training and the tools and the time to absorb basically a new job when it comes down to it, because they’re learning fast and we are rolling out use cases every single week. So what is the enablement? What are we going to do to invest, to make sure our people understand how to leverage and make this part of their story?

I think that’s the other key theme, Josh, that I’ve been really punctuating is that you can, like, when I first joined, I will say there was a lot of resistance in support because there was such a high touch engagement with each and every one of our inbound customer requests, which by the way is 30, 40,000 tickets per month, and we staffed to support that. So their fear was that we would do this to the detriment of the customer experience.

[Josh Schachter, Host]
Because they have to go, they’re gathering so much context to use it, right?

[Teresa Anania]
Like, I have to go to this system and that system. Right, and they’re feeling that our customers expect that almost white glove treatment each and every one of them, no matter how small, how large, and their fear was that we would basically to the detriment of the customer, we would be driving this AI innovation track. The reality is the use cases we’ve rolled out and tested, our CSAT, our TNPS, our average handling time, our number of handoffs have all been higher under this world.

So our customers don’t want to have to always talk to a human. And when you can have the perfect balance and nuance of where AI can self-serve and provide them a frictionless experience, that’s when your level of effort completely matches the customer expectation. So I think that’s a big theme that has come out in our AMAs.

Like, how are we doing this and not completely eroding the customer experience? And the answer is, it’s quite the opposite. It’s actually improving.

[Josh Schachter, Host]
It’s improving and it’s augmenting. Go ahead, Jeff.

[Jeff Bussgang]
Yeah, I want to throw something in here. Teresa’s got an insight that I want to emphasize, which we’re seeing across our portfolio as well, which is that the idea that we have to make trade-offs between quality and personalization versus mass, you know, sort of service handling for larger, more enterprise-y, valuable customers versus the smaller customers. Like, those trade-offs are now changing dramatically where we are able to see tooling and platforms and processes that allow us to give personalized care tailored to small businesses and small, high-transaction throughput customers who previously had to maybe deal with less personalized care.

So I think that’s really an interesting dynamic. And that’s true in customer service. That’s true in sales.

That’s true in all sorts of, you know, human engagement where we’re seeing the personalized touch now being able to be scalable for even small customers.

[Josh Schachter, Host]
Yep. And I mean, that speaks to what Teresa said is one of our primary goals, long tail, which we’ll talk about in a moment. I mean, so I’m helping to run Atlas at Gainsight.

And our first slide of our pitch deck is you’ve got high-touch, mid-touch, and low-touch, you know, in customer success. And we want to basically take your low-touch and turn them into the equivalent of what would previously be mid-touch or higher-touch, right, through the scale of AI.

[Teresa Anania]
So I think, yeah. Right. And Jeff, let me add one other thing.

I’m sure Jeff would agree with this, is that everything we do in certainly the customer experience world needs measured and monitored and adapted to the needs of the customer without waiting for the lagging indicator of like what I shared or the metrics we’re seeing, like CSAT, TNPS. What I love about AI is it allows us to do QA at scale. When you think about the whole world of quality assurance and making sure, you know, in the moment, the sentiment of the customer, in the moment, their experience is a good one and ever-improving, it was very difficult to do that except for your traditional QA tools that were like one out of every 100 experiences were dissected.

Now with AI, you don’t have to wait for those ultimate measures. You’re constantly reviewing the customer sentiment, even the tool that Gainsight recently bought, being able to, you know, sift through thousands and thousands of emails and back and forth with the customer to get the true sentiment surfaced so that to Jeff’s point, we can be more personalized, whether it be right on the call because there was pre-identification of a pretty frustrated customer or whether it’s post-call to circle back and make sure that we landed in the right place. This is the beauty of using AI over the entire customer life cycle.

[Josh Schachter, Host]
Yeah, yeah, you’re referring to staircase, I think. And actually that’s one of my pushes is to, because we’re building the plane as we’re flying here as well, right, with Atlas. And one of my pushes, especially coming from update AI, you know, also focusing on voice to the customers, how can we integrate what we’re doing as soon as possible into staircase, into that customer intelligence so that I can see at scale then what the sentiment is if I’m placing these calls through AI and AI is having a conversation with my customer coming up towards renewal or to help them, like, what’s the vibe? Not by coding, but what’s the vibe?

How do I analyze that at scale? Or, you know, even though it’s a long tail customer, if there is something that’s acute enough, how do I escalate that in the moment? It’s a very good point.

So it sounds like, okay, so Teresa, it sounds like you guys are actually doing very well so far, right, everything’s in experimentation these days.

[Teresa Anania]
I mean, there are pitfalls along the way, Josh.

[Josh Schachter, Host]
Well, I want to hear those. We want to hear those for sure. But you piqued my interest when you were talking about how, you know, your CSAT or your NPS and all these things that have been, there’ve been upticks in that.

I’m assuming that you’re focusing right now at least one of the prominent use cases is case resolution, case management, deflection, those sorts of things. What’s in your stack? What are some of the use cases, the experimentations you’re doing, the technologies, the companies you’re using, what you can share?

Because now you’ve got listeners, you’ve got other leaders, including our CCO, that are listening to this podcast saying, hey, I want to do what Teresa’s doing. So what are you guys up to?

[Teresa Anania]
I mean, I will say, I think we’ve learned a lot and I would highly recommend a very crawl, walk, run approach. We definitely did, and I believe was the right move. We started with the internal use cases of agent assist and we looked broadly so we knew what our requirements were and the tech solution that met our requirements so that we were not limited to a point solution that was kind of custom built for let’s say the support moment of truth.

That’s what I love about our initiative. We have a governance steer code that is far broader than just like what the support leader wants to drive in the very obvious use cases of AI for support. And so even our tool selection and our use case prioritization was all around, we do not have perfect data, let’s just call it out.

I don’t know any, I fought like jumping around from like an Autodesk SaaS company that had 30 years under their belt and then to a born in the cloud SaaS company of Zendesk and now to Sophos, which is kind of a blend of the two that maybe the data challenge would, yeah, it changes but they all have them. And I believe every one of us can recognize data is not perfect. What I love about the use cases that we’ve chosen and the ones that have the biggest impact, we are just basically getting the value out of the content we have but ever improving the content in our knowledge base that was disparate across all these systems better and better.

And we never before had a real motion to do that. So for the first time ever, like our data is getting better while we are increasing productivity. So it’s all about like workflow recommendations.

It’s about email, draft responses, obviously leveraging Gen AI with all the great little, add this tone, make this a little more casual, make this a little more formal. Then we’re deploying the use cases that even before digital agent. So we didn’t go there yet.

We are deploying the capabilities around like intelligent routing. And again, I do not mean just for support. I mean, think of your business as having an intelligent front door.

Gartner speaks of this all the time and I truly embrace it. And what it means is that you think about all the inbound customer demand. You don’t think about just those that end up in tickets to support.

You think about where are they on the community searching for content and it was potentially undiscovered. Where are they in our forums? Where are they going in our support portal, on our website?

That to me is all the inbound demand. To me, a ticket creation is already a failure because they didn’t find what they need. So we looked at the use cases across that customer life cycle.

We have a lot of inbound demand on our renewals team and that process, again, thinking customer and channel, not just the direct customer. And so we tackled our prioritization to focus on giving them the workflows and the content and frankly, things like auto-quoting on demand where the rep is literally able to type in, give me a quote. So they’re interacting with the digital agent, which is the quote builder agent, and they’re telling it what they need and it’s building a quote on demand.

No more searching for entitlement in six different systems. So those use cases have resulted in better employee engagement, at least 25% improvement in productivity. And frankly, for us, which may be unique, it has allowed us to take some of the savings of attrition that we aren’t backfilling in let’s say the more reactive roles and moving it to the more proactive roles, which were in need of more resourcing so that it’s like a flywheel effect.

Because the more you do in that self-serve proactive experience with AI and humans, frankly, we use both, the more you’re saving on the customer having to create that ticket or that inbound demand in the first place. And that’s kind of the beauty of this flywheel effect that we’ve just started. I mean, we have about a year in, so we have a ways to go.

[Josh Schachter, Host]
Jeff, what do you think? I think it feels like she’s kind of got it covered. You know, this is impressive.

She’s doing it.

[Teresa Anania]
Well, yeah, I will tell you, it’s been a Herculean effort to get the entire company looking at this more holistically instead of these siloed tests that were being done. And now that we’ve unlocked, like not only tooling, but process that really thinks about the company front door, I feel like we’re making way faster progress. So that was a big learning, you know, taking it out of the silos, not boiling the ocean, but thinking about the agent equals persona X and not always thinking about just support as your starting point.

And therefore our customers have said, we don’t have to navigate your functional silos. We don’t have to think about where to go to get a quote or to get a, frankly, a help on a adoption goal or a fix in the product. They basically are able to say, here’s what I need.

And it intelligently routes to where it belongs. That’s what I think about when I think of that front door.

[Josh Schachter, Host]
So it’s augmenting, it’s not automating effectively, right? It’s helping them. And, you know, listen, we know things are moving quickly at some point, maybe there will be some of that autonomy, right?

The virtual agency there.

[Teresa Anania]
And we are, we are getting to that next Josh. Well, Josh, we’re going to do the internal enablement first and all the assist for the persona equal X. And really it’s just, it’s renewals and CSMs, it’s TAMs and service and it’s support.

That’s our goal and sellers. That’s our goals. And then we’re going to move to the digital agent use case.

And that one we feel has to be like spot on because we have done some proof of concept work and in parallel. And, you know, I will tell you, not every market firm that has a tool to share is being clear about the reality versus their marketing hype. I think there’s a lot more of like, oh, this is push button.

It’s been purpose-built. LLM is going to, you know, be very tailored to your use case. Like, no, we’ve had to do a lot of work and there is a lot of work to deploy that well.

[Jeff Bussgang]
One of the things that Teresa kind of touching on, which is a world I live in with my startups, putting my venture capital hat on, is that most large companies are accustomed to working with other large companies. There’s a reason Microsoft and Oracle and Google are so beloved within IT organizations. At Zendesk, they’re safe, trusted choices, and they’re used to working with large enterprises.

And historically, large enterprises are very precise about where they let a startup come in and get access to the data and run these experiments. Well, in this moment, all the innovation is coming from the startups. And so companies need to have this new mentality of taking a bit more risk with startups that are less mature and maybe less robust and resilient, but are at the cutting edge because if they wait forever for Salesforce and Oracle and Workday and all the big, you know, software companies to bring the AI, you know, A game to the table, they’re going to be waiting forever and they’re going to fall behind their competitors. So it’s a different dynamic that exists right now for a lot of scale companies, which is they have to get used to and be willing to run experiments with startups.

[Teresa Anania]
Yeah, so true. That’s the thing that keeps me up at night. That’s exactly, I’m so used to, like we’re a full Salesforce shop and I’m just so used to that enterprise C360, single source of truth and all of the different.

And you’re exactly right, Jeff. We are currently looking at several startup-ish companies and it’s kind of terrifying.

[Jeff Bussgang]
And I would say on the startup side, one of the things that we see is that a lot of companies are piloting, but then we get lost in, you know, the pilot ghetto and they don’t have that path to production because people are running these siloed experiments as Teresa described, but they’re not really ready to invest or willing to invest in young startups, scaling enterprise-wide or they haven’t got the buy-in from the top and the budget and the organizational prioritization lined up to do the scale-wide, you know, production rollout.

[Josh Schachter, Host]
I will say as a vendor in this space now, and by the way, it’s interesting because Gainsight in many ways is reinventing part of the company. You know, this Atlas group that we’re in now is saying we are a startup within the larger organization because we recognize that we need to have that cultural shift to be able to move at pace with all of the other startups that are out there. So I think it’s absolutely necessary for larger companies to kind of inherit that.

Yeah, embrace that experimentation machine, as Jeff would put it. One of the things that we also experience, ABC, Jeff, ABC. One of the things that we also experience is like you do get a lot of folks that are coming in for QBRs or whatever the case might be and are just kind of curious about AI.

Some folks are more intentional about it than others. Some folks want to or even need to check the box because there is executive mandate, but they don’t quite necessarily know yet how they want to use it. And or they’re kind of taking a bunch of folks and kind of piling it into lots of experiments.

And then there’s a little bit of a risk there, right? That like, okay, great. So you’re saying we are one of a dozen experiments that you’re all running in parallel.

How’s that going to work out? You know, what’s the end game for us on that one? So it’s interesting from that.

I don’t know if you see that on the commercial side at SoFlo’s, Teresa or not, but it’s interesting to see how kind of the AI sales process is playing out.

[Jeff Bussgang]
I have one question before we wrap, Josh. Yeah, please. I see the Wimbledon hat.

We got the U.S. Open coming up. It’s the big question on everybody’s mind. Emma Raducanu, Carlos Alcaraz, are they an item or not?

[Josh Schachter, Host]
Oh, I wasn’t even aware. That’s the gossip these days.

[Jeff Bussgang]
They were mixed doubles partners in the U.S. Open mixed doubles qualifier.

[Josh Schachter, Host]
Ah, okay. They’re both great players. I’m a big Alcaraz fan of the guys.

Just such a tremendous athlete.

[Jeff Bussgang]
Me too.

[Josh Schachter, Host]
Yeah, good. Jeff, give us, actually, Teresa, I want to start with you. What’s one question that you have on your mind that, you know, others in the audience, listeners might also have, something that you’re trying to figure out in your turning year?

[Teresa Anania]
Yeah, I mean, I do think that the question that we constantly navigate is this balance of readiness in the organization, the maturity assessment of our readiness and our desire to deploy very quickly. So, I really believe there’s some really good frameworks out there on maturity assessing, your readiness. I do think everyone should start somewhere, for sure.

Don’t wait. But I also think that it’s really critical that deployment have all the pieces, including the change management we talked about. Otherwise, you know, we’re going to get ahead of ourselves.

And it’s going to be some of those bad experiences that kind of taint the greatness that can be achieved with AI.

[Josh Schachter, Host]
Yeah. Jeff, end us here. Tell us, what’s the biggest takeaway that you’d love for readers of your book to get out of what you wrote?

[Jeff Bussgang]
I think the thing that I would really encourage your listeners and readers to do is just keep thinking about what the future is going to bring as these AI systems keep getting better and better and better. The AI tooling, the AI platforms we all use today are the worst we’ll ever see. The rate of improvement is stunning.

And we’re getting to a point where not only are we seeing things that these AI platforms can do that we never could have imagined and now seems quite normalized, but we’re getting to a point where we’re going to see self-improvement, where the AI systems are going to improve themselves and take actions on their own. So this agentic future and the superintelligence future is really just around the corner. And so just rethinking, what would it mean if in your organization, if in your life, you had AI systems that were not only helping you do your work better, but also being able to self-improve and coach you on how to self-improve.

I think that’s going to be a really exciting moment in our future.

[Josh Schachter, Host]
Jeff, Busgang, Teresa, Ananya, thank you guys so much for being on the episode. It was a cool mix, the yin and yang of bringing together, you know, again, an author and VC and professor with an executive leader in the space. So appreciate your time today and wish you both the best of luck.

[Jeff Bussgang]
Thank you. Thank you.