AI in customer success: how to strike the right balance

Ingmar: Hello everyone, good morning, good afternoon, good evening. Thanks for joining our October webinar panel from wherever you may be.

I'm Ingmar Zahorsky, the VP of Customer Success at ChartMogul, and I will be your host today. We're joined by two CS leaders, Diana and Steven, who have worked with many top SaaS companies both as operators and advisors. With capital efficiency being a top priority and revenue retention and expansion more challenging currently, we’d like to discuss how AI can add value to your customer success operations. We have prepared several questions that will guide our discussions, and I will also be able to answer any of your questions, so please leave them in the chat.Without further ado, Diana, could you please introduce yourself?

Diana: Sure, thanks Ingmar! I’m Diana De Jesus, and I’ve been in the SaaS space for about 10 years now, doing a variety of customer-facing roles from customer experience to customer success and customer marketing. During that time, I’ve worked at companies like Hotjar, Spoke (which was acquired by Okta), as well as a customer success platform called Catalyst. I’ve seen a lot of customer success—the good, the bad, and the ugly—and I’m happy to share any stories to give you more context. I’ve also worked directly with customer success teams in health tech, edtech, and large enterprise companies, helping them structure their teams to be more efficient and enabling customer success managers to become more confident.

Now, I am the founder of a company called Strategic CS Labs, the evolution of my previous company, The Customer Success Project. Within the Strategic CS Labs community, we share a ton of resources to help you gain practical guidance on how to be more efficient as a customer success practitioner. We also discuss how to use things like AI to help you make sense of your job and be more efficient, so you can feel a bit more in control, because customer success can sometimes be overwhelming.

Ingmar: Thank you, Diana. Steven, could you round us off with your introduction, please?

Steven: Sure thing! Hi, nice to meet everybody. I’m dialing in from Berlin, Germany. Like Diana, I’ve spent about a decade now in B2B SaaS. I started my first customer success job as a customer success manager around 11 years ago and have spent my entire career in support and customer success roles. I started my customer success journey at Signavio, where I was the first customer success manager. I built up the team and led it until the SAP acquisition after 8 or 9 years. After that, I moved into the HubSpot customer success team. Same role for DH and grew the team there before becoming the first Chief Customer Officer at a company called SalesManago. During that time, I saw the Enterprise architecture, CRM, and the e-commerce space. I started the first European Customer Success Meetup in Berlin in 2014, before there was another one, and I’m still participating in the community events nowadays. I focus a lot on helping B2B SaaS companies understand the value they’re driving and shaping customer growth strategies. I provide advice in product and customer success, aligning with the business goals of the customers. I’ve been following Diana's work very closely, so I’m very honored today to be speaking alongside her. I’m really looking forward to this webinar.

Ingmar: Thank you both for joining. It’s going to be a lot of fun, as we have a diverse set of experiences here, working in B2B SaaS. I'd like to start by getting a baseline of where we are today. Diana, how do you see the role of AI having evolved in customer success over the past few years, and what do you see as the most promising use cases today?

Diana: So, I’ll be honest, when we started talking about AI, particularly tools like ChatGPT, it was like this promised land—it was supposed to revolutionize the way we did our jobs in customer success. Like, if I were a customer success manager, I’d just kick up my feet and let AI do my entire job. At least that’s the picture a lot of us were sold on at the beginning. Fast forward a bit, and now it’s more about, well, can AI help me do some tasks here and there? So those are the tools that are more like standalone AI tools, but the tools within our customer success ecosystem have really ramped up their AI game, helping CSMs become more efficient. Before we joined this session, we were talking a little bit about tools that have integrated AI into their systems. For example, Gong has a really neat feature that allows you to ask questions and have a conversation to gain insights about what you can do next. That’s already built into the tool. We have a lot of enhancements now to help us become more strategic, find information faster, and even brainstorm.

But I think the true power of AI now lies in helping us develop skills. AI tools like ChatGPT or Gemini can help us with skills development and practice. Let’s be honest—many customer success teams are not as well-funded as, say, sales teams, so we don’t always have the budget for training. We often have to train ourselves, and AI can fill that gap. You can use it for brainstorming, to shake off the cobwebs while thinking through strategies. Those are some of the things I’ve seen over the years as AI tools have evolved.

I even wrote a guide on tools with AI built into them—like Zoom, or the co-pilot feature in Salesforce, which many of us use. How can these tools work for us and help us feel less intimidated about leveraging AI in our day-to-day.

Ingmar: So what I’m hearing is that the high-value points right now are really increasing productivity, and also skill building. I’ve seen some of those products recently at SaaStr, where someone pitched a product to me, actually using AI to prepare and talk to me at SaaStr. It was quite clever, I must say. I watched the entire demonstration. Steven, I know you have a lot of experience with process automation in your previous roles. How do you feel about where we are today with AI in customer success?

Steven: Yeah, I think I see a few similarities in the way companies approach process management and also the AI topic. What we’ve seen over the years is that understanding processes—especially the customer success process, defining it, and automating it—is becoming more and more important. Measuring the impact of AI reminds me a bit of how some companies are now considering AI in their processes. What we've seen is that sometimes the effort required to implement such a process or system is underestimated. Because of that, some initiatives appear to have a big impact but aren’t feasible in a way that actually makes sense. Therefore, the way companies approach AI has shifted. Like Diana said, it’s moved from this idea that AI will automate everything we do, to now identifying specific topics where AI actually makes sense for our daily work. This includes things like identifying personalized recommendations or automating repetitive tasks where AI can help us do a better job and have a bigger impact.

I think the most promising use cases for AI are in areas like improving the way we measure customer health, making sense of large data sets, automating follow-ups, identifying at-risk accounts, and then combining AI insights with human interaction. This allows us to still have the human touch for sense-making, which I think is key. Another area we’ll probably discuss later is how we approach Tech Touch or new forms of customer success management, where you need a deep understanding of use cases, data, and interactions with customers in order to successfully automate these processes.

From my background in process management, I usually have a different perspective from others. I say that you first need to understand the most complex cases before you can automate them. This is key to ensuring that your automation is effective. I usually focus on working with the most complex customer cases first, understanding them thoroughly, and then simplifying those insights to scale. So, there are definitely similarities in how we approach AI in customer success, and I think it’s important not to over-engineer what we’re trying to do.

Ingmar: That’s a really good point, Steven, especially when you look at the stage and size of the company you’re in, as well as the resources available. To implement a more complex solution, you often need significant resources, and if not managed properly, it can actually lower the quality of the experience. It’s important to consider your baseline. For larger companies with tens or hundreds of thousands of customers, there’s more of an opportunity to automate parts of the process. However, for smaller companies, say those with hundreds of customers or even just a thousand, the percentage you can actually automate efficiently and still deliver a good customer experience might be so small that it might not be worth the effort.

That was my experience earlier this year when I was exploring some AI tools. As a product in an analytics platform, trying to onboard a customer using an AI chatbot wasn’t working well enough yet. I felt that with the small number of customers we had, we could provide a much better experience manually than what AI could offer at that stage. We didn’t have enough customers for it to be a high-value tool. We’re a team of about 70 people, and AI just wasn’t adding value for us at that time.

Diana, what have you seen with the companies you work with? Where do you see AI helping with processes, especially in areas like onboarding or in customer communications at scale?

Diana: I really like what you mentioned about the size of the company and how feasible it is to implement AI enhancements in processes. I was talking to a client yesterday who works at a large company. They don’t have many tools in place and are still working off spreadsheets. All the data points are dumped into these spreadsheets, and they were saying, “I just started taking this Excel course because I don’t know how to do all these things like VLOOKUP. It’s stuff I learned so long ago, and now I have to redo it all.” They just don’t have the time to analyze all the data. I said, “This is where a tool like ChatGPT can step in and help.” Even this big company with the potential to have the tools to analyze data properly, due to privacy and security concerns, is opting not to use these tools and instead keep everything more manual and slim.

This is a big challenge for enterprises, but AI can still play a role here. We can anonymize the data and have AI help us identify trends. That’s one way AI can make a big impact. Another thing I’ve seen—because I recently went through this myself—is that AI can be particularly helpful in streamlining repetitive tasks, like data analysis or customer follow-ups, so teams can focus more on the strategic and human aspects of customer success. A lot of the work in customer success comes down to finding simple solutions that help you get to the answer faster. For example, I’ve figured out a few basic VLOOKUP formulas, which help me solve a very specific problem without needing an all-encompassing tool. It’s not about having a system that analyzes everything for me, but rather finding a specific formula that allows me to do my job more effectively. These seemingly simple solutions, like VLOOKUP formulas, might appear trivial, but for a Customer Success Manager, especially at scale, they can make a big difference.

When you’re managing customer data that’s spread out all over the place, it’s just not feasible to manually go through hours of data looking for insights. This is where automation can really help—at least to identify trends or speed up certain tasks, so your team can work more efficiently. These kinds of shortcuts, although small, add up over time for the entire Customer Success (CS) team. So, one major use case I see for AI, like ChatGPT, is helping identify trends in customer behavior and interactions.

For instance, when working on an onboarding flow, I’ll use tools like ChatGPT to optimize my email sequence. I’ll input the details of my campaign, including the five emails I’m planning to send, the goals, and call-to-action. Then I ask the AI to evaluate what might be missing or what could be improved. The AI highlights things I might have overlooked, which makes me think more critically about the campaign and offers insights I might not have considered. These little hacks help refine our processes and make us more effective as CS managers.

Ingmar: At ChartMogul, we’ve found some specific use cases where AI and automation really add value. For example, we use AI to enrich lead data by gathering industry information, which helps our sales team understand the lead’s Ideal Customer Profile (ICP) more quickly. This is powered by GPT and built in collaboration with our Revenue Operations team. We also use similar tools to automate parts of our feature request process, reducing the manual effort involved in collecting and organizing customer feedback. This kind of automation increases productivity and speeds up our internal processes.

While we haven’t yet used AI for personalizing our customer interactions at scale, I see its potential to be a powerful tool in that area. Steven, what do you think about how customer success can be personalized, especially considering that every customer might have different needs and goals depending on their department or use case?

Steven: When considering tools for customer success, I always think about the "three Ps"—People, Process, and Product. These are three key areas to focus on when implementing anything, whether it’s a new process, a new tool, or a new way of working.

First, identify who you are solving a problem for. This could be a specific customer segment or department. What is the pain point you’re addressing? Then, gain a high-level understanding of the process you’re trying to improve or automate. What needs to be done and how? Finally, the third P is the product—the tool that will help you implement the solution.

This approach reminds me of when we at Signavio introduced a customer success platform to manage our 3,000 enterprise customers. You can’t underestimate the work that goes into maintaining such a platform—creating campaigns, health scores, models, and playbooks for managing a large customer base. Similarly, with AI tools, you need to understand the problem before you choose the product. AI tools, like any other tool, are meant to fix a problem. They take input and produce output. But before choosing a tool, it’s crucial to understand what specific pain point you’re addressing, how the process works, and how the product can help.

I still see too many companies picking tools too early, without fully understanding what they need. It’s important to do a thorough evaluation of your process and your needs before deciding on a tool. And while I come from the vendor space, I’m always in favor of trying to work without a tool first—understanding the process before you add automation. Then, you can find the product that truly solves your pain.

Diana: I completely agree, Steven. When I worked at a customer success platform, I realized how important it is to have a solid process in place before you start relying on tools. Sometimes, we think the tool itself will solve all our issues, but it’s not effective unless we already have a good understanding of our data and customer journey. For instance, if we consider using a customer success platform, but we don’t have the kind of analytics that a tool like ChartMogul provides, it becomes much harder to make use of the customer success platform effectively.

Starting with process is key, and then tools can help us once we’ve established that. I’ve had a similar experience with AI—if you don’t fully understand your customer journey, AI might end up magnifying your problems instead of solving them. For example, if you’re using AI for customer outreach but don’t have a proper cadence or understanding of customer needs, you might annoy customers instead of delighting them. So, you need to really dig in, talk to customers, and identify pain points in the onboarding process before applying any automation. That groundwork takes a lot of effort but is essential for success.

I really like how you framed the discussion about the shift from “easy mode” to “hard mode,” Steven. It’s like what we were talking about earlier with the value of understanding the product before trying to automate or introduce AI into the process. It’s essential to truly grasp how things work manually before diving into tools and automation. I find that diving into things by hand, even if it's painful or time-consuming at first, helps you uncover issues and gaps that might not be apparent in the initial hypothesis. It's a great learning experience.

Building on what you said about involving people in the process: this is something I've seen work really well. When you have the end users—whether that’s Customer Success Managers (CSMs), Sales Development Representatives (SDRs), or Account Executives (AEs)—participate in the improvement process, it builds more buy-in, and they can provide insights that make the implementation of new tools and processes smoother. Pilot testing with a smaller group allows you to work out the kinks and adjust before scaling. Change management is tough, and I see a lot of tools fail when they haven’t considered the people using them. So, gathering feedback and input from the frontline workers who actually face the pain is crucial.

Ingmar: That brings me to my next question: How do you think the role of Customer Success Managers (CSMs) will evolve as companies move toward more AI-driven processes? What skills should CSMs focus on to stay competitive in this changing landscape?

Steven: Great question. I think one of the biggest shifts we’ll see is the increasing importance of the three Ps: People, Process, and Product. CSMs need to understand their customer’s challenges and pain points. But they also need to master the process of addressing those needs, whether that’s through direct engagement, tools, or a mix of both. In this age of AI, the product knowledge of a CSM is more critical than ever. They’ll need to be product experts—understanding the ins and outs of the tool, how to use it to its full potential, and where it fits into the customer's broader goals.

On the other hand, they also need to become more consultative. AI will make things more efficient, but it won’t replace the strategic conversations that CSMs need to have. In my experience, there’s still a gap in many sales and customer success teams between knowing how to make the product work and how to articulate the value of that product to the customer. CSMs need to become more adept at discussing the ROI of the product and helping customers reach those outcomes. This consultative approach is key—AI might help gather data, but it’s still the human CSM who needs to interpret it and guide the customer.

Diana: I couldn’t agree more. We’ve already started seeing the shift from being just relationship managers to becoming revenue owners. I think understanding the customer’s “why”—why they bought the product in the first place and what business goals they want to achieve—is crucial for CSMs today. It’s not enough to be a friendly face anymore. CSMs need to have those business-savvy conversations, dive into ROI, and understand the financial implications of the product for the customer. That’s where the skills around value-based selling and tracking impact will become essential.

One thing I’ve noticed is that many CSMs still focus too much on the product features, but they often miss the bigger picture. It’s about understanding how the customer measures success internally and tying the product to their broader business strategy. That’s where AI tools can really help CSMs—by giving them data to support those consultative conversations and making the customer success journey more prescriptive. AI can help you identify trends and patterns that you can then use to guide the customer’s journey, focusing on long-term value rather than short-term fixes.

And in terms of AI, we’re seeing it help automate tasks that would normally take up a lot of time—like follow-up emails or analyzing data to identify customer trends. This leaves more time for CSMs to have those important, consultative conversations. But we also need to be careful not to lose the personal touch in customer success. AI can assist with efficiency, but we still need that human connection. I think the challenge will be balancing automation and personal interaction, using AI where it makes sense and leaning on human expertise where it doesn’t.

Steven: Exactly. And I also think the next phase for customer success will be about how we can optimize the use of data in our day-to-day work. As we see more AI-driven tools, CSMs need to understand how to interpret this data—not just in terms of churn predictions but in how to strategically guide customers. AI can surface insights, but CSMs still need to be the ones who deliver them in a way that resonates with the customer’s business objectives.

Diana: Absolutely. AI can surface patterns, but human intuition and relationship-building will remain essential. CSMs will be able to step into a more consultative role because they’ll have the time and data to make more informed decisions. They’ll move from firefighting to really helping customers grow and reach their business goals. AI will help by giving CSMs the right tools, but the skill-building part comes from understanding how to use those tools strategically and maintaining that human touch. It’s about being proactive and value-driven.

So, if I may jump in with a slightly off-topic but relevant comment, it’s funny how much we’re all talking to AI tools like ChatGPT these days. No one should feel weird about it; it’s becoming part of our everyday lives, just like in the movie Her. AI can help with everything from crafting emails to even understanding your customer better and preparing for those tough conversations.

But the bigger takeaway here is that AI isn’t just here to replace us—it’s here to make us more effective, so we can focus on the more strategic, consultative parts of our jobs. AI will streamline the operations side, but CSMs still need to own the relationship, the strategy, and the revenue impact for their customers. This is the next step for customer success in the age of AI.

I remember mentioning this ten years ago during a podcast—I probably annoyed everyone at the time—but I’ve always said that Customer Success isn’t just a function, it’s a strategy. We all need to think about Customer Success. I think we’re finally there now, but with that shift has come a lot of expectation.

If you’ve ever had a Customer Success role and sat in a boardroom, you’ll know what I mean. Investors, whether it’s private equity or venture capital, will ask you the tough questions: What value are you providing? How will you improve the net retention rate? It's great to be in that position, but you also need solid answers. Just saying that your health score went up won’t solve those questions. To be honest, I was in a room with 120 people recently, and when the question was asked, 'Who considers their health score to be predictive?', very few hands went up. So, always be realistic about what you can actually provide. Be careful about what you bring to the table if you want to be part of the conversation."

Ingmar: That’s so true. And when we think about data and AI, especially in terms of metrics, we need to measure the impact AI has on Customer Success. What are some of the key metrics to focus on? Steven, you mentioned earlier the importance of finding one task to focus on. If we apply that same approach to AI, then we should measure our AI tools the same way."

Steven: Exactly. If you focus on one task and measure that, it gives you a clear sense of whether you're on track."

Diana: In my newsletter, Strategic CS Labs, I often include a section about how to measure AI's impact. For example, if I’m using ChatGPT to help me write like a marketer but I'm a Customer Success Manager (CSM), I would measure metrics like open rates, response rates, and whether customers are taking action. It’s about breaking down those small tasks to create a bigger picture that will ultimately contribute to the overall outcome. If I’m using AI tools to analyze data, I’d measure how much time I saved in the process and what actionable insights I gained that lead to a desired outcome.

Steven: I love that approach, Diana. It’s so important to have very clear goals around why you're using a tool and to track specific metrics tied to those goals.

Diana: Exactly. And if you're using a tool like ChatGPT's co-pilot feature, for example, it’s essential to track how much time is spent operationalizing it, how many times it's being used, and whether the tool is being adopted. This will help assess if it’s actually improving productivity or not. For an Operations team, metrics like tool consumption might matter, whereas for CSMs, it’s more about productivity. For CS leaders, it's about aggregating all of those metrics to get a clear picture of the team's efficiency.

Steven: And that’s a great point. The more specific you can get about your reporting, the easier it is to understand the results and make decisions based on them. Clear metrics allow for real insights.

I don’t have anything else to add, Diana. That was a perfect answer. I’d just emphasize focusing on customer metrics. At the end of the day, it’s about the impact you’re making for your customers, not just internal efficiency.

Ingmar: I completely agree with you. It’s all about ensuring that the technology augments, not replaces, the human touch in Customer Success. AI can help streamline certain tasks, but it shouldn’t take over everything. It’s essential to retain the human element in customer relationships.

Diana: I couldn’t agree more. In fact, that brings me to something I’ve seen as a common pitfall—thinking that AI can replace jobs. Sometimes leaders, or even CEOs, might look at AI and think, 'Can we reduce headcount because of this?' But if you actually benchmark AI tools against the capability of a skilled person, you'll realize that human expertise, especially in B2B SaaS, still brings so much value. A well-run team, especially with great CSMs, is a competitive advantage that AI can't easily replicate.

Steven: Exactly, and that’s why, even as AI becomes more prevalent, we need to focus on augmenting human efforts, not replacing them. In B2C environments, there may be some valid use cases for AI, but in B2B, the personal relationship and deep understanding of customers are irreplaceable.

Ingmar: That’s why I love this quote from a report on AI in Customer Success. It’s from the SVP of Product Management at Gainsight: 'Despite its technological roots, AI is a path to more human-centric interactions. We’re working to ensure that AI enhances rather than replaces human connections, because at the end of the day, humans are the best at navigating nuanced conversations.' This really resonates with me because you need that human skill to truly deliver value to customers.

Ingmar: Absolutely. That’s the perfect way to wrap things up. AI is a tool that can support us, but it’s our human touch that truly makes a difference in Customer Success. Thank you both for such an insightful conversation.

Diana: Thank you, Ingmar and Steven, for the great discussion. It’s been a pleasure!

Steven: Thanks, everyone. It’s been fantastic. Have a great day!