Meet Ashley, Celigo’s manager of IT engineering
00:05 Jeremy Erwin: Hi, I’m Jeremy Erwin, host of the Builder’s Hub podcast. I’d like to introduce Ashley Esposto, manager of IT engineering at Celigo. Ashley, welcome to the podcast. Before we get started, we like to ask everyone who comes on to tell us a little something that we can’t find on your LinkedIn profile.
00:22 Ashley Esposto: Yeah, I’d love to share that. I would say something you probably wouldn’t find on my LinkedIn profile is that I’m a big crafter of every type. So, of course, I’ve got my Cricut, got my sewing machine, paints and everything in between. I love a good craft. And if you could see the rest of my house right now, you would see that it’s fully ready to go for Halloween. We’re making decorations and stuff.
00:49 Jeremy Erwin: That’s awesome. So when you’re not working on IT support tickets and such, you’re crafting at home, which is always a nice way to decompress.
00:58 Ashley Esposto: It is. It is. And I definitely love to decompress with crafts. Clothes, all that good stuff.
01:06 Jeremy Erwin: So when you’re not decompressing, tell us a little bit about what you do at Celigo and what your day to day looks like.
01:11 Ashley Esposto: Absolutely. When I’m compressing… I’m kidding, I’m totally joking. At Celigo, I manage the IT department. We are a team that interfaces with our entire global company, which is really cool. We support the internal IT functions for Celigo, so we get to support all of our awesome colleagues. Our customers are our colleagues, which is really cool. And we touch a little bit of everything. We touch access. We touch infrastructure. And with that, we’re constantly making improvements. We’re constantly doing things just, frankly, to make our Celigo colleagues’ lives better in their day to day.
01:56 Jeremy Erwin: That’s awesome. And how technical would you say you are, Ashley? I have a follow-up question based on your response there.
02:04 Ashley Esposto: You know, I would say I’m actually pretty technical. As I moved into a manager position here at Celigo, it’s been a little bit more managing, but there’s definitely no lack of hands-on technical collaboration, we’ll say. I get to build every day, and I super enjoy building. Celigo makes it really easy. We have a lot of really great tools that allow us to do that. We’re definitely a hands-on team, and so the technical needs remain.
Why Ashley’s team left asset spreadsheets behind
02:38 Jeremy Erwin: Got it. And what was it that made you build your first flow on the Celigo integration platform? Was there a point in time where you logged off and said, “I can’t manage this much anymore,” or was it, “I have too many other priorities that I need to focus on”? What was the tipping point for you?
02:57 Ashley Esposto: You know, I think a lot of our greatest ideas come out of necessity. When there’s a need, you kind of build a solution for it, and there are a lot of things that need attention. I’ve built quite a few things here at Celigo, but I would say one of my biggest builds was what we call our ITAM, or IT asset management. We were living in spreadsheets. And I know, I know. 2007’s calling, give them back their spreadsheets, right? Just kidding. But we were living in these spreadsheets. And with spreadsheets, human error can happen, outdated data. All of these things can really happen.
So we built an autonomous system that allows asset tracking and asset management to occur through a few API connections. Different systems own different pieces of data, and it can track the lifecycle of a device: moving from a depot, moving into a user’s hands, being repaired, replaced, retired, whatever the case is. It tracks it all automatically based on changes to where it is, who has it, whether it’s being signed into, all of these different triggers. And like I said, those different systems have that respective information, and they govern the narrative of the lifecycle of the asset in the system. So now we actually don’t have to update a single spreadsheet for any of our assets.
04:27 Jeremy Erwin: Impressive.
04:28 Ashley Esposto: It’s relieving.
Inside the ITAM flows
04:30 Jeremy Erwin: Yeah, I can imagine. And if you can, Ashley, could you walk us through that flow, or some of the flows you’ve built on the platform, and explain what’s happening there?
04:40 Ashley Esposto: Yeah, definitely. I’m going to share my screen and show you a little bit about those flows. Let me know if you can see my screen.
04:55 Jeremy Erwin: Yeah. And for all of our listeners, we’ll try to narrate this as best we can.
05:00 Ashley Esposto: Yes. So right here, we’re already in the tile. I don’t want to give away all my state secrets, so I’m going to stay inside the tile. But we’re inside our ITAM tile, which is really cool. You’ll see here we have Jamf Protect. This is a secondary flow that we built. But coming into our unassigned section, we have quite a few flows pulling data from different sources.
For a little bit of context, our IT asset management resides in Jira Assets right now, or JSM Assets. So you’ll see references to JSM Assets. Just know that’s where everything’s filtering into, and that’s where it’s compiling all of the data for each asset.
05:53 Ashley Esposto: So we have a couple things in here. I’m going to start from the bottom up. Sycomp is our equipment partner. They’re our depot, so they hold all of our equipment. They ship it. They receive it. They collect it when a Celigo employee departs or when they join us. They take in repairs, they send it out. So basically, this is going to have a shipping, an assignment and a disposition. It’s going to tell us: is it an active machine? Can it be redeployed? Is it retired? Does it need repair? And all of that translates into JSM Assets, which is really cool.
You’ll see some other ones here. These are all the older spreadsheets we had. I had it reference all of those sheets to pull in all of the data, because there was a lot of really good data we could build a historical narrative on for all of these assets. So we were able to build an asset management system that pulls all of Celigo’s known asset tracking history, from day one of the spreadsheet being created to today, which is really cool.
Then I decided to also include our networking equipment in this asset management system. The idea is that we’ll expand. We’ll be able to add printers. It won’t just be about Celigo laptops. It’ll be about Celigo hardware, Celigo equipment being used across our offices, across our workforce, all of these different places.
07:33 Jeremy Erwin: Yeah.
07:34 Ashley Esposto: It’s really that single pane of glass that we’re going for for assets. We also use Jamf for our Mac devices. Jamf is our Mac MDM. It pulls daily, actually twice a day. It pulls details from Jamf to Assets, and it’ll say: is it checking in? What’s the OS? Is it up to date? Is this person still assigned to it?
Let’s say we have a device being sent back for repairs. That device is no longer going to be signed into, so it’s not going to be checking in. So if there’s a discrepancy, like, okay, this is being returned to Sycomp, and Sycomp hasn’t received it yet, we’ll still know who it’s assigned to. And then when it gets to Sycomp, there will be information pulled from the depot saying it’s arrived, it’s in the depot, it’s under repair. So it really keeps the story of the asset and where it is in its lifecycle, which is really cool.
08:42 Jeremy Erwin: Awesome.
08:42 Ashley Esposto: We have the same with Intune. I will tell you we are moving off of Intune and going on to JumpCloud, but we still want to keep track of those devices during the migration. I’m going to circle back to Ingrid in a moment. That was a later iteration, and it was actually a really cool way to leverage some AI capabilities.
But to wrap up some of the other things we have here: we have daily check-ins, and we have a couple groups in our workspace. We use Google Workspace, where we have all Mac users and all Windows users. Based on the information in here, those group memberships update automatically based on the machine each person has in this asset system, which is really cool. And these are two variations of that check-in, just to make sure it’s up to date.
Ingrid, the ITAM bot
09:37 Ashley Esposto: Circling back to Ingrid: we had this really comprehensive database. If you know anything about Jira or JSM, it’s a robust system. If you’re not super familiar with it, it can be a little bit hard to search using JQL and all these different search queries. So I wanted to make it easier for anyone on my team, at any skill level, whether they’re coming in as an IT help desk support person, an engineer, someone above that or anywhere in between.
I wanted a system that makes it easier to communicate with the data that’s here, so I built a Slack channel with a bot, and the bot’s name is Ingrid, the ITAM bot. We use Claude as the LLM in this case. Using plain language, you can go ask Ingrid, “Hey, what’s Jeremy’s serial number?” It’ll come back pulling data from the ITAM system and say, “Jeremy’s serial number is such and such.” Then you could ask, “What are the specs on it?” And it’ll say it’s this much RAM, this much hard drive capacity, this year, this model. It just makes it really easy to find this information in a very plain-language way.
How automation changed Ashley’s day
10:56 Jeremy Erwin: Awesome. So you spoke a little bit there, Ashley, about how your team’s day to day has changed by automating some of these processes. As someone in a leadership role, managing a team, how has building these flows changed the way that you work?
11:18 Ashley Esposto: Well, these have definitely changed my day. For a long time, I was core support for US-facing hours. Most of my team resides in India, so they do a lot of APAC-facing work, and we do cross-functional work. But I would have to update a spreadsheet for every single person coming into Celigo, every single person who ever departed Celigo, every single person who ever needed a replacement, an upgrade, a repair, whatever it was, asset related. You have to update that spreadsheet, spend time, make sure you’re updating the correct line, make sure you’re getting the correct data. If there’s a new device coming in, you have to bring it in and update it. It was a lot of hours, accumulatively. Of course, this was over years. It’s a lot of hours and a lot of time spent.
And then there’s auditing it. Once a quarter, more often to be honest with you, you go in, you audit it, you make sure all of this data is correct, and you make sure it reflects the actual realities of the asset disposition. None of that happens anymore. None of that happens. No time is spent on this, and we really don’t have to worry about inconsistency in the data, because it’s based off API triggers in different systems, the systems that make the most sense to own that data.
Sycomp is going to know about repairs and replacements and retirement way better than Jamf is going to know about it. But Jamf is 100% going to know better who is on that device than Sycomp. We know who we shipped it to, but there’s a lot of data validation in the system, which is really, really cool, and it brings a lot of reassurance in the data quality.
13:09 Jeremy Erwin: And managing that manually, as you said you were, just shows why you needed to decompress with the arts and crafts you mentioned earlier.
13:19 Ashley Esposto: Yes. Completely.
What Ashley is building next
13:22 Jeremy Erwin: Ashley, final question for you. What are you building now? It could be technical. It could be not technical at all. But what are you building, and why are you building it?
13:32 Ashley Esposto: That is a great question. I feel like I’m building six things at once, and I like it that way, so don’t worry. We always have these great ideas, and we jump into them, which is really cool. That’s one thing I really like about Celigo: we are empowered to build the things we have ideas for, to make our day better, make our colleagues’ day better, make our company’s day better.
But right now, I’m really focused on our LCM project. We’re working on automating onboarding and offboarding on a company scale. I’ll admit it, it’s overdue. But it’s in the works, and it’s really exciting to see. I’d say that’s my biggest project right now, and I’m really proud to be working on it. It’s a little bit of my baby. Hopefully I get to debut it on a bigger scale shortly. There’s been a lot of work on it. The HR team’s been great. Jamil, who I think is also on this podcast, has been helping out with that, which is great. It’s really exciting.
14:44 Jeremy Erwin: Awesome. We’ll definitely look to slot you in for a future episode to talk about that project once you get it live and learn from it.
14:56 Ashley Esposto: Absolutely. I would love to.
15:00 Jeremy Erwin: Thanks for joining, Ashley.
15:02 Ashley Esposto: Of course. More than happy. Thanks for having me on Builder’s Corner. It’s been a real pleasure to showcase a little bit about what we’re working on.
15:10 Jeremy Erwin: We loved having you. Thank you.
Spreadsheets as a source, not an enemy
15:18 Jeremy Erwin: Wow, Nate. There’s a lot to unpack from everything Ashley said in this episode. One of the first things we’ll start with is something Ashley said: our greatest ideas come out of necessity. She never framed it as an IT asset management integration project. It was a spreadsheet pain project. So, a question for you, Nate. When you talk to Celigo prospects and customers, how many of their best use cases are hiding in spreadsheets that nobody even thinks about as an integration?
15:52 Nate Briant: I think almost everybody we talk to, it is out of necessity: either the necessity to change their broken processes, or the necessity that they want to grow. They need infrastructure in place that can scale with them. And so much of it is a modernization project, where you’re going from something very basic plus spreadsheets to maybe a modern ERP.
Look, the reality is we use spreadsheets every day. Right, Jeremy? How many spreadsheets are we in together? People have been trying to kill spreadsheets forever, but it’s such an easy place for people to spin up a new spreadsheet and work. I think what really happens is, if you’re trying to take a business action in a spreadsheet, that’s when it’s like, okay, we need to shift this. If I’m having to have a button that runs a macro that does some action in a system, that’s not where a spreadsheet belongs. It’s a great place to store data and compare data and run some basic formulas. But when you’re trying to create advanced models or, like I said, take business action, that necessity is where people can see a platform like Celigo come in and take it from just a data source into real, controlled business action.
A system of record by domain
17:03 Jeremy Erwin: That’s right. And I always think about the auditability of things. The version control within our Google Sheets probably wouldn’t suffice for most CIOs out there.
Another thing Ashley brought up in this episode, she didn’t come out and say it, but what she got to was that there’s a system of record by domain, not one master. The design principle is that the system that makes the most sense owns that data. So Nate, is this the right mental model for prospects obsessed with having a single source of truth? What are your thoughts on that?
17:46 Nate Briant: I think it definitely depends, Jeremy. I really liked how Ashley described it, especially for her use case with asset management and tickets and stuff like that. I think it’s obvious that those things maybe don’t need to live together. But when you talk about financial records, you definitely want a single source of truth. So I think it’s important that you have a model that’s correct for the business vertical you’re in, and not try to force it.
If you’re trying to force everything into one system, you end up buying a bunch of extra modules or something like that. Versus, you buy applications for the need they’re going to fix, and then if you can sync that data together, you don’t have to cram something that doesn’t exist into another application. So I thought it was really interesting, and it made total sense for her use case.
Backfilling legacy spreadsheet data
18:36 Jeremy Erwin: Yeah, I think so. And I do think it’s case by case. I think that’s the overarching theme here: what makes the most sense for the use case at hand. I don’t think it necessarily makes sense to say, “Let’s load our production database with spreadsheets,” when there are migration tools and data ingestion offerings, like we have here at Celigo, that allow for that historical sync.
And that leads into the next conversation: backfilling some of these applications or databases with legacy spreadsheets. Is that something, Nate, that you feel we undersell here at Celigo? Most people hear “automation” and think, this is my ideal state in a month, six months, a year. Do we undersell the fact that those legacy data points need to be uploaded into these applications?
19:38 Nate Briant: At the end of the day, people are just trying to get their job done. They don’t necessarily care if something’s “legacy” or “modern” at the ground level, when you’re just trying to accomplish your tasks. But it’s important for people in leadership roles, strategic leaders, to understand that while people are trying to get their job done, this can ultimately be a limitation on future growth or the flexibility of the company, if people are stuck working with a legacy system or stuck on a spreadsheet.
And change management is a big part of it. It’s like, “I don’t want to change how I do work, because I’m just trying to get it done.” But if suddenly you have a huge influx of data, can your old systems handle it? So the old systems need to be integrated, absolutely. In some cases they can’t be thrown away immediately, but ultimately you’ve got to be looking at the long-term business processes to figure that out.
20:38 Jeremy Erwin: I’d agree with you 100%. One of the things I talk about most often is master data: ensuring the master data is correct. I feel like a lot of organizations out there have Google Sheets or Excel spreadsheets full of their master data, and that’s the source of truth. So I think you’re spot on there, Nate. It’s crucial to still have and support it. Maybe not for the next year or so, but at least during that initial transition into a fully automated organization.
21:10 Nate Briant: And I would love it if someone listening has a great experience with master data to reach out to us and explain what master data really solved for their business, because the concept makes total sense. You need to be able to keep the same record schemas and definitions all the same. But to go back to the previous point about sources of truth and golden records, what is segmenting out master data going to do for the long term? It seems like it’s more to manage, and maybe that’s just a naive take. But I’ve heard lots of people complain about master data more than they’ve praised it. So I’d love for someone to come and explain it to us like we’re five years old.
21:52 Jeremy Erwin: That’s an open invite to any of our listeners to please reach out, and we’ll have you on one of the next podcast episodes.
AI as the last mile, not the first step
22:00 Jeremy Erwin: Nate, let’s talk about Ingrid a bit. I love the name Ashley gave her bot. This is the Slack bot over JSM data. To me, it seems the AI layer wasn’t necessarily the project. It was a later step, a later iteration, to solve specific skills gaps. I know JQL is hard. Help desk staff shouldn’t really need to rely on that too much. Having AI act as an accessibility layer on top of that is, I think, beneficial. But in your mind, was Ingrid an agentic automation or an AI automation to start? Or how could that agentic functionality have been incorporated from the start?
22:55 Nate Briant: I think it’s a super interesting question, and I think the framing really matters here. Again, it goes back to people just trying to get their job done as easily as possible. Just because you throw AI at a problem doesn’t mean it’s going to make it better. On top of that, accessibility to an agentic output is something a lot of people are solving. If you move it into a productivity app like Slack or Teams that you’re already using, I think that has huge benefits, because it’s where people are working already.
When it comes to something like querying data, I think it’s a great use case. AI is really great at generating structured queries. We even see that ourselves when we want to query Snowflake data. It’s a great use case to convert your natural language request into a query you don’t have to write. So I think it’s really about least agency, where you can say, “I want to use AI to assist in what I’m doing,” but it’s not the whole game. I’m still trying to accomplish my task. It’s assisting and making that faster, and it’s the accessibility layer. It’s obviously really valuable, and it’s not a totally autonomous use case, but it’s a great place to put agentic into a business process.
24:20 Jeremy Erwin: Something you said there was having users work out of the apps they’re already working in. Meeting users where they’re at. It reminds me of Field of Dreams: if you build it, they’ll come. This is a little bit different. You have to meet users where they’re at, because as soon as you incorporate a new process or a new application, their entire current process is thrown awry. Then they have to learn the new process and the new tools. When you meet them where they are, whether it’s agentic automation or a new application, you’re going to have a higher adoption rate, in my opinion.
25:07 Nate Briant: Totally. It’s change management. When you’re asking people to change their processes, there’s going to be resistance. But you can really see here how the agentic piece is slotting into something that probably already happens naturally, where someone on the team asks a question of someone else on the team, maybe someone who writes the queries or has the knowledge, and then they go out and do the manual work. This is very much where someone who maybe doesn’t have the insight to write the query themselves, instead of asking another human, is asking an agent, and they’re just doing it through Slack. So there’s very little change management there. And I think Ashley saw high adoption because it just made sense.
Measuring value beyond the hours saved
25:52 Jeremy Erwin: And if we talk a little bit about Ashley’s role: she supports US hours. Core support. There’s a team in India that also supports our India office. She was personally updating spreadsheets every time an employee was hired, every time an employee was offboarded, if there was a repair. And her quote was, “None of that happens anymore.” She didn’t quantify the hours saved. In a way, I wish she had, but I think “I don’t audit this anymore” is even more convincing. Because you have a near 100% accurate automation that’s working. Not needing to audit it means there’s a task taken off your plate, so you can focus on more important or higher-revenue tasks.
26:54 Nate Briant: Totally. You and I love to dig in and find the exact hours and tie it to what people are getting paid, to get the perfect return on investment everybody’s trying to chase. But when Ashley’s just trying to make everybody’s lives easier, including her own, it’s one thing to automate it. It’s another thing to have the confidence that you don’t have to sit there and monitor it.
There’s a lot of talk about hallucinations and inaccuracies with AI, which absolutely is a problem, and I think we talked about it in another episode about when things go wrong. But humans make errors as well. What’s nice is you can put guardrails around agents that you really can’t put around humans. That final step is what gives Ashley the confidence to not sit there and watch the spreadsheet, make sure it’s updated, or do the task. The task is being completed with the confidence and guardrails that she doesn’t even have to think about it. Like you said, now she can focus on something better.
28:00 Jeremy Erwin: 100%. You and I, being solutions and former solutions people, get wrapped up a lot in the dollar amount, the return on investment. This may sound corny, but I feel the return there is having happier employees who aren’t jumping through hoops to get their tasks done. In past roles of mine, the most frustrating thing was knowing I had to get something done but not having access to A and C to actually complete the task.
28:34 Nate Briant: Yep, 100%. It’s beyond just the dollar amount. You can get into other things like employee retention, SLA improvements, general efficiencies. You could probably calculate a dollar amount if you sat down with a spreadsheet, maybe, and put all the numbers in. But you can clearly see it, especially in how many flows Ashley has built. She immediately saw the value and just kept converting business processes into automations. You wouldn’t do that if the value wasn’t there, despite not needing to quantify it necessarily.
Empowering builders with guardrails
29:12 Jeremy Erwin: I’d agree 100%. Something else Ashley mentioned that you and I are very well aware of, being employees of Celigo, but for our listeners, a sneak peek behind the curtain: Ashley said we’re empowered to build the things we have ideas for. I think that rings true here at Celigo. We’re empowered to build anything we possibly can, as long as it makes us more productive and more efficient. And having that freedom to brainstorm and think outside the box is exactly what set Ashley up for success. She knew she had a problem and chose to do something about it, because she had the autonomy to do that.
You and I know what that looks like internally, Nate. But for our listeners, what does that empowerment look like, to allow your users to build things they have ideas for? Is it having the correct toolset? The correct permissions? Knowing the right people within the organization?
30:17 Nate Briant: I’m sure that helps. It’s funny you brought up the right permissions. I immediately thought of that as something that’s maybe underappreciated. Permissions can feel like friction at the time. But having the right tool to build an automation doesn’t necessarily mean you’re going to build the best automation, and you can automate the wrong thing.
What I really like here is our IT team does a great job. I have the tools with Celigo to do this, and now, especially with AI, you can almost build anything with Claude or Cursor, almost to the detriment of IT’s sleep schedule. But they do a great job of saying, “You have the idea. Let’s talk about the right permissions.” Because if you get the permissions wrong… too much access is the obvious one. I don’t want to get an email from the CFO that I adjusted all of the invoices in our ERP when I only wanted to automate understanding where our customers are geographically.
So that’s really critical to think about: once you have a toolset, you’ve got to manage both the expectations and the governance of it. Otherwise, you can really get yourself in trouble. It’s feeling empowered to do the automations, but also feeling confident you’re not going to break something. That’s the empowerment I feel. I can work within these guardrails, know my work’s getting done and know I’m not breaking anything for anybody else.
31:54 Jeremy Erwin: I feel the same way. I know I don’t have access to production databases and instances where I can really mess things up. So I do have that confidence. That’s a great point, Nate.
32:10 Nate Briant: And I think the lifecycle is a big part of that, where you can prove an automation has been built. On our team, Adam built a great automation for building web pages, and it saved so much time. We proved out the point in a sandbox, and once we showed it to IT, they said, “Okay, let’s go through this and promote it into production.” Now you’re getting all these efficiencies in prod, with the confidence that IT has the reins on it. So it’s about the lifecycle, not just the initial build.
Episode recap
32:41 Jeremy Erwin: You’re exactly right there. Nate, I loved what I heard from Ashley. Some of the things to call out as a recap: we should treat spreadsheets as a source, not an enemy. They have their place in the world. We have to let the system that knows own the data. A single source of truth holds in some use cases, but in others you may not have one. And lastly, AI is the last mile, not the first step. You may have automations and integrations running perfectly today, but adding AI can gain you even more efficiencies and make other users more aware of those data points and datasets. Any additional points you’d like to make about anything Ashley brought up?
33:45 Nate Briant: No, I think you summarized it great, Jeremy. We’re having a lot of fun interviewing builders who are using the platform and building automations in general, and we have more episodes coming out real soon. We also love to hear from people listening. If you want to reach out with use cases, you can find us at [email protected], or message me and Jeremy on LinkedIn. We’d love to talk to builders and see what everybody’s doing, because solving your day-to-day tasks and building really fun automations is something we love talking about.
34:16 Jeremy Erwin: Absolutely. Thanks, everyone, for listening in, and we’ll catch you in two weeks. And one more thing: if you’re enjoying the Builder’s Hub Podcast, subscribe on YouTube or any of your favorite podcasting apps. We’re everywhere audio lives. And if you know someone who’d get value out of this conversation, send it their way. Thanks for listening.
