In this episode of Mastering CS: Candid Leader Insights, Irina Cismas sits down with Eric Mistry, Customer Success Manager and AI and Automation Transformation Lead at Zapier, the automation platform that connects apps and workflows. Eric’s role sits at the intersection of customer success and automation strategy, and he spends his days helping customers shift from using AI casually to building repeatable, agentic workflows.
He shares what the Automation Strategist title actually means, how he thinks about breaking down impossible problems into small solvable ones, what he automated first and why it changed how he shows up for customers, how Zapier’s MCP integration and Claude Code have become his personal cheat code, and why empathy is the most important technical skill he has ever developed.
What You’ll Learn
- What the Automation Strategist title means and how it differs from a traditional CSM role
- Why Eric is his own first customer and how internal builds become customer-facing workflows
- What the first things he automated were and what the results actually looked like
- How he thinks about the mindset shift required to move people from zero to one with AI
- What infrastructure you actually need to run a scalable automation stack
Key Insights & Takeaways
The best CSMs are their own first customer. Half the automations Eric shares with customers he built for himself first. That’s not a shortcut, it’s a proof of concept.
Impossible is just a lot of small problems bundled together. Breaking down a complex goal into individual, easy-to-answer questions is how you make automation feel accessible to anyone.
Automate the research so you can be more human in the meeting. An automated morning briefing with company news, support history, and LinkedIn activity means walking into 10 calls fully prepared instead of doing hours of manual research.
Share things that are actually useful. The reason people adopt what Eric shares is because he only shares things that have proven value. The credibility of the share matters.
Podcast Transcript
Intro
Irina (0:06 – 0:27)
Welcome to Mastering CS Candidate Leader Insights, the podcast where we dive into the world of customer success with industry leaders. I’m your host, Irina Cismas, and today I’m joined by Eric Mistry, Customer Success Manager and AI and Automation Transformation Lead at Zapier. Eric, I’m really happy to have you here!
Thanks for joining!
Eric (0:28 – 0:29)
My pleasure! Thanks for having me on.
Irina (0:30 – 0:43)
Eric, you are a CSM at Zapier, and you also lead AI and automation and transformation, a lot of big words. What does that actually look like day to day?
Eric (0:44 – 2:09)
Actually, the funny thing is I have to update some of my things. Our title has changed from CSM, which most people think of as the traditional CSM, to Automation Strategist, which I think reflects a lot more of the, hey, we’re thinking about the strategy of how people are going to choose to use this. So when we’re talking with customers, it’s less about, oh, we’re going to just review your numbers and do the standard customer success thing. It’s more about how do we actually help you with those business goals and how do we bring the strategy side? That’s what you’re actually paying for with a CSM: how are you going to help me use this tool to achieve the outcomes we actually bought the tool for?
So we try to have that name really reflect that focus, but yeah, the rest of the roles too. Everything combines. I wear a lot of different hats. It’s hectic, but it’s a lot of fun. They all really do mix together.
The AI and Automation Transformation Lead work is about designing workshops, education, and sessions that directly move people either from zero to one. I’ve never used AI in my life, how do I use it in a way that’s actually effective? And those folks who are, hey, I’ve used AI in a basic way, I’ve had that chat, I’ve had that conversation, how do I turn that into repeatable workflows and think less about AI and more about agentic? So helping move people from zero to one, and also from one to a million.
Who Eric Works With Internally and Why
Irina (2:11 – 2:28)
Who do you interact with the most in this part of the role? Not from the customer side because that goes without saying, but I’m curious, who is the team around you and what are the roles that contribute to you delivering value to the end customer?
Eric (2:29 – 3:46)
I think that’s the most exciting part of the whole thing. One of the things that customer success in general gets to do is the role is by its nature very cross-functional, and the work I get to do on a daily basis is even more that way. I’m talking to people in product directly. The product manager is building the tool: how and why did you build this tool? What should I know? What should I bring in? The designers: why did you design the flow a certain way? Why did you make these decisions? The support team.
We’ve got hands down one of the best support teams in the entire world. Understanding what are customers missing? What are those gaps? Where can we close those gaps? Product marketing: how are we trying to position ourselves? What are we saying? What are we not saying? What’s landing with our customers? What are they actually interested in? Our sales team: why are people actually buying the tool? Because that in this world is shifting every single day with everything that’s changing in AI and in SaaS. People’s motivations and reasons for doing things are changing constantly. We’ve got to keep up.
Getting to be cross-functional with everyone is the super exciting part. And yeah, the answer to the question of who am I working with is literally everyone. And that’s the best part of the whole job.
Being Your Own First Customer: Internal Builds That Become Customer Workflows
Irina (3:47 – 4:14)
Based on what you are just saying me, I was thinking of one thing, asking so many why’s. It’s like you are a CSM also for the internal team, for your internal responders. So basically it’s a dual role as for external customers, but also for internal customers.
You have to, I would say harmonize along the user journey.
Eric (4:15 – 4:55)
We are our own first customer. We have a value called build the thing, which is: you see something you can build to fix it, so build it. And I’m not just helping people use Zapier. Half the flows I share with the customers I work with, I’ve built for myself: an external meeting briefing, a thing that looks at my week and says, hey, here’s where you can improve these things, AI analysis, all these things I built for myself. I just turn around and share them with my customers.
And that’s a lot of what we do internally. Our RevOps team, our EA teams, everyone who builds on Zapier is really good at sharing that. And then we just help transmit those internal builds to the rest of our customers.
Irina (4:57 – 5:00)
What’s the one thing you are trying to move the most now?
Eric (5:01 – 5:51)
I think the biggest thing that I’m thinking about constantly is how do we unlock the mindset shift for people, but how do we get people to see, oh, you can do these particular things. You can actually connect these tools. And I always talk about what’s on your list of things that are impossible.
And for a lot of folks, what they would have written on that list six months a year ago as impossible is just, oh, that’s an afternoon of work, or that’s just combining these two tools that wouldn’t talk to each other otherwise. There’s a lot of things that are entirely possible, and you just have to get the mindset shift for people to say, let me break the problem down. Let me solve those small problems.
And suddenly impossible is easy.
What Eric Automated First and What the Results Looked Like
Irina (5:52 – 6:13)
What were the first things you automated when you joined the team? And what did you try and end up dropping, if that was the case? Give me a few quick examples because that’s useful for our audience as well. They can learn from what you’ve automated.
Eric (6:14 – 7:42)
One of the most useful things I’ve automated, which used to take a good portion of my morning, or the night before if I knew I had a whole day of meetings ahead, is meeting research. I want to have some context on exactly who I’m meeting with and what’s going on with their company. So every single morning at 7am, I get an automated Slack message that says, here’s all your meetings with external people, here’s the company, and here’s the latest news on that company. Here’s all the things that are happening with that company: upcoming conferences they’re hosting, a release they just did, something they were in the news for, a podcast their CEO was on.
The same sort of search happens for the person. So it can say, hey, saw you were just on LinkedIn, or it also looks at our internal support systems and says, hey, I saw you were talking to support last week about this issue. Is that resolved?
I don’t have to do any of that research now. If I have 10 calls in a day, that’s hours of work I’ve saved. I come into those meetings better. Customers have a better experience. That’s why I built it. I want to come in with the best possible preparation and have them feel heard and seen. That’s the kind of work that robots should do. I don’t need to be doing that research. I might dig a little deeper if I see something interesting in the summary, but having the basic research done for me is a great use of automation.
What Happens to the Time You Win Back
Irina (7:43 – 8:04)
And all this time that was gained because it was automated and you didn’t have to invest into the research part, you filled it up with what, with more customer-facing interactions or maybe with more thinking about what else can you build?
Eric (8:05 – 9:10)
Really it’s a mix. The joke is always that meetings will fill any vacuum of time, but I think when you remove the busy work, there is higher work. Anytime you see work where a robot should be doing it, eliminate that and turn that time into human work. A robot right now can’t do some of this inference drawing between things, doesn’t understand the nuance of a customer call. If I watch a call recording, if I watch a customer going through a flow live, I can pick up on some emotions that a machine right now cannot. So I can spend my time doing those more human things.
And sometimes that’s also just more time to meet with customers and making those meetings better. It’s less, oh, we need to figure out all the research during the call. It’s, no, we walk in with all the research already done. Now let’s sit down and work on your problems together. Let’s talk through your strategy.
How to Get CSMs to Actually Run Their Week on Automation
Irina (9:11 – 9:20)
Building is one thing, but getting people to use it is another. How do you get CSMs to actually run their week on it?
Eric (9:21 – 10:27)
Partially we’ve just got a team that is naturally curious and wants to do these things. And also it’s leading by example. I show, hey, here’s the past week I’ve been running this thing, here’s the results I’ve gotten out of it. Do you also want it? And everyone on the team has a little bit of that builder mindset too. They’ll take it, they’ll tweak it in their own little way. They may add a weather report to it. If they commute to a co-working space, they can tweak it for their own purposes.
That’s one side of it. And then second, I’m not going to share something if it’s not useful. There are a lot of things I build that are just fun for me. It might be a little meme maker, it might be something that does a very specific personal task. Unless it’s more applicable to folks, I’m not going to share it. So they know that if I share something it’s useful. Or I’ll preface it with, hey, this is the seed of something potentially useful. I figured out a way to grab RSS feeds and get a good analysis out of them. Is there any purpose for that for people? And they might jump on that piece.
Irina (10:29 – 10:54)
What is all of this running on? What’s the infrastructure that you have to have available so that you can run all of those automation? Of course, besides Zapier, because I’m assuming that you are basically using your own tool for most of them.
But where does Zapier plug in?
Eric (10:54 – 12:51)
95% of what I do just runs on Zapier. I’m also using Cursor, Claude Code, and Codex, all the harnesses, using those on the daily. The big unlock for me is MCP, which is the hot buzzword right now. I’m using the Zapier MCP constantly. I have it set up in Cursor, so if I need to draft a hundred customer emails that need some variation built in, some conditional data, I just tell Cursor to use the Zapier MCP to use my Gmail to draft those emails. It’s kind of the cheat code for everything for me.
So what is it built on? It’s basically all Zapier. But the important things I tell people if they’re thinking about this more generally: you need to have your automation platform, your tool that’s going to move everything together. You need to have a set of data, whether that’s Snowflake, Databricks, any sort of internal data system, an S3 bucket, product usage data, any data you can get your hands on. You need your email tool, internal chat tool, calendar. If you connect those things together, that handles so much. And if you’ve got a CRM, if you’ve got a customer success platform, make sure those things talk to each other. Understand your data and where things live. I know my product usage data lives right here. I know my customer education data lives right here. What lines do I need to draw between those two? Do I need to, hey, whenever someone completes a course, write it to their record in the customer success platform? That’s an automation I have built because it’s super useful to know this person completed the course. All these signals you need: if you connect your tools together and keep it relatively simple, you can do so much.
Irina (12:52 – 13:20)
Awesome. When you were talking, I wanted to ask you, you’ve been building this, but what did you have to learn in order to be up to date with this? Is this more towards a technical CSM role where you need to have a technical background or in this moment, AI so commoditized that everyone can learn it and put it in practice?
Eric (13:21 – 15:00)
The technical barrier, while it exists, I have a slight technical background but I’m never going to be an engineer. Even for those folks who are relatively non-technical, if you can learn the basics of systems thinking, which I think is really applicable to everyone, you can do this. Can I think in systems? Can I think in larger things? That’s what strategic CSMs do already. They may not do it on the technical side of things, but if you can talk to an executive about outcomes and connect your tool to those outcomes, you can think about automation systems. It’s all just that same framework.
And I think it’s just a natural curiosity. The biggest thing where people get stuck is they look at it and they see this giant problem. But when you look at those giant problems, it’s actually just a bunch of really small problems bundled together. If you want to create a giant health score that’s done automatically, that sounds really daunting to any CSM. But if I said, hey, can you just tell me how many times they’ve logged into the platform? That’s easy. Can you tell me how many times they’ve done the key action on your platform? That’s pretty easy. How many courses have they done? All those little problems are pretty easy. You solve the little problems, you package them together, and you’ve got a big problem solved.
Getting folks into that mindset is the biggest challenge. It’s not necessarily a technical skill. It’s just understand where your data is, think in systems, and break down the problems.
How to Deal with Messy Data Across Systems
Irina (15:01 – 15:15)
Data is one of the biggest barriers. How do you guys at Zapier make sure that you have it consistent between all the systems that share the data points?
Eric (15:17 – 16:14)
We’re really lucky to have a fantastic data team, but I’ve worked in higher ed, I’ve worked in a lot of different places, and I’ve not always had the cleanest data to work with. The two things that I look for are: what can I find that’s a viable source of truth? Sometimes you step into a CRM and there are eight different company name fields or a bunch of different customer ID fields. Choose one that’s your source of truth and lock it down. This is the source of truth for that. Just nail the basics first.
And once you have the basics nailed, you can take that to leadership and say, I need better, cleaner data, because here’s what I’m doing with these three basic data fields that I can trust. If I can get this data field and this data field to be consistently updated and trustable, I can scale these processes that lead to these business outcomes. That’s the data puzzle.
But again, it’s like with everything: if you start simple and start small, a lot of the problems are much more solvable than they seem.
The Habit That Has Carried Eric Through Every Role
Irina (16:15 – 16:28)
Your whole career has been about getting people and technology to work together. Is there a habit, a way of thinking or something you learned the hard way early on that still shows up in how you run accounts today?
Eric (16:29 – 18:25)
Yeah, I think the kind of early act to my career was working in higher ed. And my job was to essentially be that person who lives between the IT department and the faculty. So much, oh, let’s fix your computer.
But how do you use virtual technology to teach medical simulations? How do you use the LMS in a creative way? How do you use instructional video?
How do you build programs that do all these interesting things? The IT department was not by nature really good at talking to faculty and faculty because they didn’t have that technical language, weren’t really good at talking to the IT department, like articulating what they needed. So I think the biggest thing that I’ve learned from all that, it’s just like an empathy for everyone’s an expert at what they do and everyone’s like, no one is.
Like deliberately trying to not do things. Everyone is trying to do their best. So like every, every single problem is like we’re solving this problem together.
And how can I understand the way you think and the way you bring things in with faculty? It’s like, okay, I translate what we’re talking about technically to something they handle. If it’s in their field, I’ll talk about it in their field language.
If it’s just like pedagogy, I’ll talk about it in language pedagogy. I’m never going to use technical terms unless they’re absolutely necessary. Even then I’m going to say, oh, but it’s like this.
So like really strong use of metaphors, those kinds of things. But I think it’s really just having empathy for every single person that’s trying to solve the problem together. So let’s bring that forward to everyone.
And I think that just helps across the field, working with folks who may not ever think about automation technology. So it’s like, okay, how can I guide them? They’re not like, it’s not something they’ve worked with before.
So if they’re missing the language, if they’re missing the metaphor, like how can I put myself in their shoes and explain the way it’s got work for them? I think that’s a really useful thing just in general for CS. And it’s certainly really well.
Irina (18:27 – 18:34)
I think this was a really interesting conversation. Thanks again for joining and taking the time to speak with me today.
Eric (18:35 – 18:37)
Yeah, absolutely. Thanks for having me on. This was great.
Irina (18:37 – 18:44)
To everyone listening, thanks for tuning in until next time. Stay curious, keep learning and mastering customer success.
