Adam Coffey on this episode
Adam Coffey is Director of Analytics and Reporting at Ivanti, sitting at the rare intersection of go-to-market, finance, and technical software development. With five years at Adobe in GTM finance and previous experience running GTM Finance at another company, Adam brings both operational credibility and hands-on engineering skills—he's not just talking about building instead of buying, he's done it at scale globally.
The core insight: AI applications allow companies to replace expensive SaaS tooling by building custom software in days, not months, but only if your data foundation is clean and your organization is structured to support it. Ivanti faced quotes of several hundred thousand dollars each for tools like Gong and Clari, so Adam's team built their own Sales Command Center using GitHub Copilot, VS Code, and Microsoft Power Apps. It handles meeting prep, flags risks from call transcripts and emails, and automatically suggests MEDDPICC field updates—saving reps roughly 30 minutes per call that they can redirect to selling and pipeline generation.
The reframe here is critical: most companies believe AI will magically fix dirty data. It won't. Adam explicitly warns that "you can't slap AI on a dirty CRM"—and he hasn't seen any tool that can. The deeper the conversation goes, the clearer it becomes that data architecture, team structure, and organizational alignment matter more than the AI itself. Finance, RevOps, and GTM teams stay siloed because they don't share data, coordinate models, or plan together. That breaks the organization's ability to do accurate annual planning, sales capacity modeling, and forecasting. Building AI apps on top of broken foundations is expensive and ineffective.
The rest of the episode covers the build-vs-buy math (saved ~500K in subscriptions while spending 80-100K on hosting and development), why SaaS faces existential risk if it can't adapt, the security rigor required before deploying homegrown tools at scale, and a roadmap for what comes next: a Finance/Ops Command Center, live-call coaching agents, and an extensible data harness so other teams can build their own AI applications without reinventing the hard infrastructure.
What we cover in this episode
-
3:04
Finance and RevOps silos How broken collaboration between finance, go-to-market, and ops leads to duplicate models, Excel files, and inaccurate planning.
-
9:52
Sales Command Center overview AI app that handles meeting prep, surfaces deal risks from transcripts and emails, and auto-suggests MEDDPICC field updates for reps.
-
11:59
Build stack: Copilot, VS Code, Power Apps The specific technology choices Adam made to build and deploy the Sales Command Center at global scale.
-
15:43
Build vs. buy: real numbers Saved roughly 500K in annual SaaS costs; invested 80-100K in hosting and development; scale matters for licensing efficiency.
-
19:57
SaaS vulnerability and adaptation Why certain SaaS categories face extinction risk and how vendors must bundle defensible products or adapt their business model.
-
24:33
Security review for homegrown tools Extensive vetting by security and IT teams required; risks include web scraping toggles leaking private data and vibe-coded vulnerabilities.
-
35:20
You cannot slap AI on dirty data AI requires clean, contextualized, curated data with clear ontology; Adam has never seen AI work effectively on poorly maintained CRMs.
-
43:17
Centralized data model over fragmentation Single semantic model across all functions beats hundreds of isolated dashboards; reduces work, ensures consistency, enables AI to scale.
The lines worth sharing
You can't slap it on your dirty CRM and somehow it magically knows how to clean that all correctly. It is not today.
Adam Coffey · 35:20
If a SaaS company only sells SaaS and they don't also sell hardware or something, it's adapt or die time with this AI revolution.
Adam Coffey · 20:28
The goal is not to cut costs. That is a byproduct. The goal is around efficiency gains and giving AI superpowers to each function.
Adam Coffey · 11:06
Data goes to IT to die. A data team that sits in IT, I personally haven't seen it where stakeholders are fully satisfied.
Adam Coffey · 40:27
Common questions from this episode
Can AI replace Gong and Clari?
Yes, if your data is clean and your team has the engineering capability. Adam built a Sales Command Center using GitHub Copilot and Power Apps that handles meeting prep, flags risks from transcripts, and suggests field updates. It saved Ivanti roughly 500K in SaaS costs.
How much does it cost to build your own sales AI tool?
Ivanti spent approximately 80-100K on hosting and development while saving 500K in annual SaaS subscriptions. The exact cost depends on company size, licensing deals, and your team's existing infrastructure. Smaller companies may not see the same ROI.
What data setup do you need before using AI apps?
You need clean, contextualized, curated data with clear ontology showing how it all works together. Adam warns you cannot slap AI on a dirty CRM. Use Gartner's Data Maturity Scale to assess your current state before building.
Should data teams report to IT or another function?
Adam strongly advocates for data teams to sit in operations or report directly to the CFO, not IT. Data teams buried in IT become service-ticket takers rather than strategic partners. They should be part of planning and annual strategy decisions.
Why do dashboards still matter in the age of AI?
Dashboards provide deterministic, repeatable views of KPIs, while AI outputs are probabilistic and variable. A centralized data model powering both dashboards and AI ensures consistency, reduces duplication, and creates a reliable foundation for AI applications.
How do you get finance and RevOps to collaborate?
Break silos by moving shared data into a central warehouse with controlled access, not locked in isolated systems. Finance and RevOps need to build capacity models, contribution models, and forecasts together, or annual planning will be inaccurate guesswork.
SEO meta description
Adam Coffey, Director of Analytics at Ivanti, built AI apps to replace Gong and Clari, saving 500K in SaaS costs. Learn why data maturity matters more than AI hype.
Target keywords
Adam Coffey
Ivanti
replace Gong and Clari
sales command center
GitHub Copilot
Power Apps
build vs buy SaaS
AI sales tools
data maturity
MEDDPICC
RevOps data
sales AI automation
Read the complete conversation
Auto-generated from the episode audio and lightly cleaned. Expand to read or search inside.
Read the full transcript · 57 KB · Adam Coffey
EDDIE0:04Welcome to Go to Market. In this podcast, we share tangible, actionable playbooks from the trenches working as go to market strategy and rev ops consultants for our clients here at Union Square. Consulting and candid conversations with revenue leaders in the market that have been there. Now let's get into it.
RACHAEL0:23My guest today is Adam Coffey, director of analytics and reporting at Avanti. He's also one of the rare people who sits right at the intersection of go to market finance and the technical side of building his own apps. He built his own AI tools, has some cherry opinions on data and where it belongs, and a clear view on what revenue teams can start building instead of buying.
RACHAEL0:42Adam, welcome to the show.
ADAM0:44Yeah, thanks for having me.
RACHAEL0:46Awesome. So, yeah. Tell us a little bit about your background. First of all, I know you have so much experience in all these different areas that go to market side, the finance side, the technical side. How did you end up living at that intersection of all these specialties?
ADAM0:59Right. I mean, I think you said it well, it really is this intersection between finance go to market operations and data analytics. So it definitely wasn't the plan from the beginning. I a few companies back, I spent five years at Adobe doing go to market finance, lived and breathed in Salesforce and bookings policies and bookings forecasts. My the last company I was at, I was running go to Market Finance so very heavily involved in the go to market in rev ops world from a financial perspective.
ADAM1:35So we were able to build a good bridge and a good relationship between those worlds, which are often fractured at that time with some of the technical skill sets. I had acquired, the CFO came to me and asked, can we create this new role to where I can actually sit within a within a finance function, but doing business analytics to actually officially bridge the gap between these siloed worlds?
ADAM2:04From there brings me my current role. My team currently owns all global reporting and analytics for our entire company. We own power BI as our reporting stack and data visualization tool. And and now building out a new team for AI applications where we'll be targeting certain SaaS subscriptions that we have, especially the loving fruit that will build out our own custom software version using AI tools.
ADAM2:36And the hope is give AI superpowers to each function. The two not just trying to cut costs with subscriptions, but really
ADAM2:46maximize productivity and efficiency. Let people focus on what matters most. Sales reps focus on sailing, not updating Salesforce things of that nature.
RACHAEL2:55organization, what does a revenue or finance team consistently get wrong? Simply because nobody can speak on all three of those languages?
ADAM3:04Yeah. Good question. So what? It depends on the company. Of course. It depends on what the culture is like and how mature their their data is. And really where the data team sits matters a ton in my belief and how that data seems structured. But what I have seen at times is finance sits in their silo. Their perspective is the data is too sensitive for anyone to look at, so they will continue to do their own thing, but they won't have as much of the inputs from the other orgs that they're partnering with from go to market.
ADAM3:40And they'll do their own forecast, their own modeling and all of that. Then rev ops struggles if if finance doesn't work in a good partnership or also like HR and recruiting, you know, there's all these different pieces to the puzzle. If they don't work in a good system. Rev ops I just see them living in these glorified Excel files so much and it's painful.
ADAM4:07Someone, some unlucky soul has to maintain this and update it, and they'll have something like a quota workbook with all their territory planning, and they're trying to keep it updated with who gave notice when they're leaving. What's their end date? Oh, we hired someone. Now we have an open position of recruiting for. And it's it's often the headcount that is a huge value add.
ADAM4:29If they weren't trying to replicate it and maintain it in some Excel file somewhere. So there's things that I think a perfect example would be certain models that finance often owns, but in conjunction with revenue ops. So one would be sales capacity model or marketing contribution model or pipeline efficiency model, but go into sales capacity modeling. I personally believe you cannot do accurate annual planning for go to market without a good sales capacity model, which what that is, is basically, let's look at your sales reps who have a direct or primary quota.
ADAM5:10Let's have a attrition forecast for headcount. What's our backfill time if someone leaves, is it takes three months for them to get filled? What's your ramp schedule for them? Are you planning to have new hires? So you need to know what finance plan is, what's in budget, and all of this. And then you look at the sum of all their quota for ramped sales reps compared to your target breakdown for that territory.
ADAM5:35Do you even have enough quota at each point of the year to cover what the target is, so that it works in both ways? It tells you your targets and your quotas, right? Are you deploying enough quota? What's your attainment assumptions of that quota to hit the target. But it works also in reverse. It tells finance how many headcount they need to budget for sales reps, because from there you have your sales rep headcount.
ADAM5:59That should be a big piece of how many bars do you need in place to support the sales reps? How many sales engineers will you need? So that gives you your full expense profile for annual planning. And if you don't have a good model like that, that's working together between teams. I my perspective is you're just guessing.
RACHAEL6:20And how can go to market ops teams and revenue leaders become like better connected and closer friends with finance and work together more. Because I know a lot of the like, a lot of times finance is always butting heads with different revenue leaders wanting budget for things and finances like, why do we actually need this? And it's always kind of a push and pull.
ADAM6:40Yeah, yeah, that's that's the hard world that each, each company is to figure out for themselves. But there might be a stigma that, oh, finance wants to talk. It's about budget and I can't. They're stealing my piggy bank. Or maybe, maybe marketing always gets a tax every time they need to cut costs. We always come to marketing to to save a dollar.
ADAM7:06So there might be some of this that finance is just policing everything. Definitely not the case. Yes. A part of being a is to manage expenses. However, if finance has a strong enough team that they're not just the finance do ruse, but like they actually do understand go to market to where they have a seat at the table for planning and strategy to where they're a valued partner.
ADAM7:32And if finance is willing to and from a privacy security perspective, have their data certain things shared. So what I mean by that is they have their their tools, they have their different systems and forecasting tools. I believe it needs to be in your data warehouse. So if you have snowflake or something, allow a data team to have access to that.
ADAM8:00See them as a value partner to where it can be secured. Someone needs to have access to it, but you can have a centralized lockdown database that no one else has access to. You can control permissions, roll up security, whatever you want from there. You're no longer going to be just caveman reporting, repetitive Excel files and all of that.
ADAM8:17You can actually do things with that data. You can see even if it's if it's anonymized or names are taken away, or it's just the aggregates like the sum total of expected attrition or sum total of expected new hires. I think it's that breakdown of, well, no one can actually look at any of this. It's too sensitive. We're going to keep it in a bubble of six people.
ADAM8:41Well, that's how your plan is never going to be operationalized because we're running the business over here a different way. So I mean, I think that's step one is just be willing to collaborate. Sometimes the willingness is only there for like one person finance and one person go to market. And they talk occasionally, but it has to actually be there more tangibly than that with where the data lives.
ADAM9:04And do the automated pipelines, are they set up in the right way?
RACHAEL9:07I love that because then it takes away that that barrier and that bottleneck of needing to set up a meeting or needing to have like a cadence of connecting with somebody on finance to update them on these things. I feel like every time you make it like one step harder, it gets a little bit less probable that it's going to happen.
RACHAEL9:26So having that data accessible to the people in finance or select people in finance at least, who can look at it and answer their own questions for themselves to make the best decisions for their department. Make so much sense, right?
ADAM9:38Exactly.
RACHAEL9:39All right. So let's get into the fun part. Not that finance and stuff isn't super fun. I'm sure it is for like a lot of people, probably a lot of people listening to this. But I want to know, tell me about the coolest things that you've been building with AI.
ADAM9:52Yeah, this is the meat. This is the fun part. So AI applications, like I said, there's that new team I'm building out and it's just AI is probably way too much of a word that everyone's hearing, too much that your eyes gloss over when you hear it. However, still, everyone goes into their job each day and they're being told by the board or the company that owns them or whoever like implement AI, use it everywhere.
ADAM10:18And it's not as easy as it sounds for a lot of reasons. This is the AI revolution. It's like taking over the world. But where do you put it? Where do you dedicate your resources? And really, what's the driver of doing this? I think that there's a misconception that implement AI because we can cut cost if that's the driver.
ADAM10:42I think oftentimes people are going to be disappointed. You know, there's stories out there about people getting sticker shock when they see their their anthropic bill of how many tokens they've been using. And, and some are saying like, this is more expensive than just having the head count. You're doing this. So it can definitely be done in a profitable, efficient way.
ADAM11:06You definitely, absolutely could reduce costs in certain ways. My view is that's a byproduct, that that's not the goal for doing it. But that is a added benefit if you can do it the right way. I would say the goal, like I said previously, it's around the the efficiency gains input to the game. So the coolest thing I've built recently is something that I'm calling the sales command center.
ADAM11:33So basically there's a tech stack that I used. It's not the only way it can be done. It can be done a lot of different ways, but I guess I'll quickly go through it. For anyone who wants to try to replicate something similar, you can use GitHub Copilot for vibe coding essentially, and using it in an IDE called VS code, and you need a way to host it in stores and authentication.
ADAM11:59I used Microsoft's Power apps and you're just off to the races from there. So that's the short version to get you started. So what it is basically is AI superpowers for sales reps. The long term goal might be they don't need to be in Salesforce, updating deals and updating fields and and doing all their meeting prep and all that, that the AI within the app is doing most of that for them.
ADAM12:27But I think some people don't realize that, you know, teams, for example, if you record a call, it has a transcript, that data is available, you can use their API connection and you can access that transcript. You don't need Gong just to get access to a recording and a transcript. And for for clarity, being used for writing back to Salesforce or a visa or all these other expensive tools.
ADAM12:54Instead of that, we built our own. It was a pretty steep learning curve at first. We'll hopefully get better in the future, but it's been great. We've deployed it globally to all sales reps and sales managers. The main features that we've that we built in the first draft are things like here's your calendar. You have upcoming call. We're going to do the meeting prep for you, give you an agenda.
ADAM13:22Here's the questions that you need to ask in order to get information. You'll need to progress the deal through the pipeline. Here's the risk the reviewing, the call transcripts, the emails, and Salesforce fields and all of that. Here's the hurdles you realize your economic buyer hasn't engaged in the past three meetings, and he hasn't responded to an email.
ADAM13:43Other people are responding. So, you know, it'll highlight a few of those things after the call. Take the transcript. There's a lot of details of how it all works. So we have a convert to English process and all of this. And then from the transcripts it will give recommendations. So I med pick field updates. Next steps current situation in a right back to Salesforce instantly.
ADAM14:06But you have a chance to make some manual adjustments. But the agents will give you a first pass, a recommendation that does pretty well, and it goes on and on and on from there. Things from actions you should take follow up, generate an email response. Really the goal for the sales rep side, if I made an assumption that a sales rep spends 30 minutes per sales call doing meeting prep and Salesforce updates, so between the two, let's say it's 30 minutes.
ADAM14:39If I can cut out 90% of that and they give all that time back to selling, you could you don't need to be a data scientist to figure out that's a good thing, that that will increase the ceiling for the capacity of how many deals they can manage any given time, or they can spend more time on pipeline generation activities.
ADAM14:58So that's the premise. There's so much more we're wanting to do, so much more that I want to test and build out. We also built out manager forecasting tools in there as well. So it's like turning into the one stop shop to where eventually if we hit all the bases, why do you need to go to Salesforce? Why do you need to open it?
ADAM15:21You need to kind of becomes this expensive data storage repository that we're writing back to and reading from all the time. But, you know, that's where we're starting.
RACHAEL15:31So how much money do you think you saved the company, like without by not needing all these different SaaS programs and just doing it all yourselves?
ADAM15:43Yeah. So there some of these tools we at my current company, we did not actually go purchase and I had to go I'm trying to rip out and replace them. We got quotes and we're looking into several options and several different things. I can say that the combined or honestly, each quote was several hundred thousand dollars each, but going this route, it's not it's not free.
ADAM16:07To be clear, you need a really strong team who can solve this, and you don't want the agents just running off of raw data needs to be very clean and curated. And there's context, and it's called an ontology. So it understands how it all works together. But also you're doing storage and hosting. The reason I mentioned I went with Microsoft PowerApps direction was because when we looked into the direction of what if we just built in cloud code and then hosted ourselves their own storage and authentication, time to delivery was significantly longer and more painful and a lot more work.
ADAM16:45And also it was it was more expensive going our own route versus it depends on the company size. So if you're a large enough company that you're going to have, you can get bulk pricing and licensing. It's a lot cheaper actually to go with a one time or a subscription with Microsoft PowerApps to where I can then go publish 50 applications, and that same user can solve access to all of them versus the other route.
ADAM17:14You're paying a ton. So yes, this first project, you know, we might have saved half $1 million in subscriptions, but it cost us, you know, something like maybe 80 to 100 grand in hosting and storage. But now I can add more on top of that, I don't have an additional cost associated with with billing applications now.
RACHAEL17:36And so what was the biggest challenge of going the build versus by route? Was there anything that surprised you during the process of making it all?
ADAM17:44I mean, honestly, I was first surprised by the capabilities of how far GitHub Copilot has come. That was kind of mind boggling. I guess the challenges are really there's no one specific challenge of I would say you don't know what you don't know, and this is just something that isn't. A lot of people don't know this as possible yet.
ADAM18:10I'm sure they will. I honestly believe that this AI application space of building your own tools or replacing SaaS tools is just going to be a wave. That's that's people are starting to figure it out. So I would say like my superpower is the ability to just figure it out. So if you don't know how power apps works or what did surprise me is going from, you know, the power BI world where, like I said, my my team owns power BI, I've lived a power BI for a long time.
ADAM18:44I competed in the power BI World Championships. That's all very.
ADAM18:50Structured. There's guardrails and parameters around it. Now you're going over this route, building your own application. You have to figure out how to make it work performant, like you have to figure out how to fetch data because you're trying to pull all these records of opportunities and it's slow. You have 300 people using it at the same time globally and and all these different things.
ADAM19:10There was a lot more I had to figure out how to make it work. How do I convert Chinese transcripts to English? And, you know, so many little pieces. I've never built my own software, anything from scratch myself from like the first line of code to the millionth line. So yeah, it was a it was a hard problem to solve.
ADAM19:29It was very frustrating at times. And it's so new. Even Microsoft Solution architects didn't have the right answer because they're like, well, we're trying to figure out what this code app thing means. And so again, it was just like, hey, there's no instruction manual, no YouTube video. You have to get your hands dirty and figured out.
RACHAEL19:48Exactly. It's so new that, yeah, there's no way to to learn from other people yet. You can be one of the first people making those tutorials. Adam.
ADAM19:55Yeah. We'll see. I'm excited.
RACHAEL19:57Should. Yeah. And it's it's such a it's such a new and snowballing thing. And I keep having these conversations with other go to market leaders around this idea that like the SaaS is in trouble, you know, like as people become more and more knowledgeable about how to do these things and build them for themselves. And as these tools become more mainstream and accessible, like and significantly cheaper in a lot of aspects, I think there is a risk there with these, these companies.
RACHAEL20:28And I'm not saying that like Gong is released, are going to go out of business anytime soon. Of course not. But there there is a shift happening in the market and something is going to have to change. I think as this gets more and more popular and people figure out how to do it for themselves.
ADAM20:43Yeah, and I'm glad you brought that up, that that is such an important point. And like you said, yeah, they might not go out of business immediately, but it is the adapt or die time. It's already starting. Like, don't be the next Kodak. You need to actually adapt with this. You know there's there's the Adobe's out there. And you know I've heard some crazy things.
ADAM21:06You know, my my brother in law was at Adobe for a very long time running a huge creative team. And it was just unheard of like to ever not use Adobe. Of course, the creative world, they're all using Adobe products now. He's left Adobe recently and and he says there's not a single person on his new team who uses any Adobe product.
ADAM21:29Like it's just a crazy shift. So I would say that it's worth looking into that. If a SaaS company, if they're only SaaS and they have no other, they don't also sell hardware or something like that. I know everyone's having the conversation. How are we durable against this AI revolution, and how can you be more durable against the AI revolution?
ADAM21:55I would I would say what I have seen. I've heard the argument of, oh, we're the system of record, so that's why we're durable. You need the system of record. Like I kind of push back on that just because these AI applications I'm using, there's something called a dataverse and I can create new tables. So when someone says it's a forecast, I'm saving to that.
ADAM22:18And I have this whole infrastructure I've built out and it's incredible, but I can create my own system of record now in days. So yes, there's truth to that. But like the data, if you can't like, I guess it depends if you can pull out your existing history or not. But the point is, I would say the things that are durable are things like, least for now, cybersecurity.
ADAM22:42You probably don't want to go vibe code your own cybersecurity.
RACHAEL22:45Yeah.
ADAM22:46Probably a bad idea. Yeah. If you have something related to hardware. So like, you know, AWS has hardware behind the scenes that's like you're you're not going to be able to replace that. What else? I I've seen certain product families or certain product groups that are definitely more vulnerable. So something like an IT ticketing system easy to replace or like a Monday.
ADAM23:17Easy. Now if you have other products, highly recommended to secure that from everyone. Just like eventually completely turning off of that. I would say bundle that in certain product groups like have product bundles to wear those vulnerable products. It's now no longer standing on its own in Island. It's just a part of an ecosystem that that just comes with it.
ADAM23:39And those it's bundled with more durable products that can protect that from a self because otherwise you're like, oh, look, I can cancel this and save some money, but I'm still buy this product from you if you have that scenario. I've seen bundling being a good approach, but it has to be a discussion if if no one's looking at it, which I'd be shocked.
ADAM23:58I'm sure everyone's talking about it right now. Then it's then. Yeah, their SaaS will die.
RACHAEL24:05Yeah. And you, you brought up security. And I think that's a great thing to point out to is that a lot of this vibe coding stuff, you know, if the person doing it doesn't understand the security precautions and security measures, there's a lot of I mean, I'm not a coding person, so I can't speak very, very eloquently to this, but from what I understand, there's a there's a lot of security risk that comes with vibe coding and like using proprietary vibe coded builds in a big company,
ADAM24:33Yeah. And I had to go through extensive review before I get implement anything I built at scale security and privacy team was they are strict. So go extensive reviews and a lot of different like it teams digging into the code base and the repository themselves before we can get clearance. There's things that you might not even think of.
ADAM24:55There's little toggles when you're building an agent that says, you know, do you want to grab this from the web? And they have so many issues secure from their security standpoint about doing that, of apparently there's ways that you can just like steal those queries from the web. And, and it's like sending some of your private data out there.
ADAM25:14So there's right ways and wrong ways to do it. So anyway. Like I said, a steep learning curve. But once you get through it the first time then you're really off to the races.
RACHAEL25:22Yeah, that's that's something that I also need to audit on my own AI instances like figure out where my toggles are, my sending information somewhere I don't want to. So what are you building next then?
ADAM25:34So yeah, I'm more excited for the future. I mean, I love the sales command center and everything, but what I want to do next and what I am working on now,
ADAM25:43other than just all the other enhancements, I'll keep doing enhancements for the sales command center. There's more I want to do and tell until I see one day the metro that says no one's logged into Salesforce for six months, you know, then I'll be like, okay, we're good.
ADAM25:56But so other than that, there's other command centers that I'm building out. One is this finance slash ops command center. So we kind of talked about especially in rev ops, I think there's not really great tooling. There might be some out there, but I've never seen anything amazing for for rev ops there. Just again all these glorified Excel files oftentimes that like midsize companies at least maybe larger companies have solved it.
ADAM26:24And then for finance they have planning software. There's things like adaptive and pigment and, and a plan and all these things. Honestly, I hate all of them. I've used them. I've been deep in the systems of back inside of them and they're so hard to configure, which is another huge benefit of AI applications. You can do whatever you want.
ADAM26:44You don't have to wait for their roadmap and hope for a new feature. You just you can do it. Sky's the limit. Yeah. So I'm building out my this is for me personally because I wanted it for me for planning season and everything. And then it turns out so does everybody else. But it's just the dream state for for planning operations in finance to where finance will have their side.
ADAM27:08They'll have access to certain things that ops doesn't have access to whatever, and all the data is going to work together. So all the planning tools, the, the, the comp plans will be in there based on the quota planning and territory planning and headcount and all of that. And you'll be able to work together with the comp plan to output your commission structure.
ADAM27:27And the whole thing is going to work together. And what's cool about that one is I found a way to, on the back end, make it work with like a Excel like interface, which is what everyone's used to. So as you like, edit your comp plan, hit edit. You can use all the same Excel functions. And and so it's enabling the builders to build.
ADAM27:46They don't need to be hyper technical and get a quarter million dollar implementation consulting plan together just to influence a single tool. So that's that's just one. The other ones I'd maybe more quickly mention. Another one is as sales agent. And I guess this is give ideas out there. Everyone's like what's the use case for AI? I guess this is a get the creative juices flowing on on what's possible and some ideas of what you could do.
ADAM28:13So the other one that's definitely possible is that I'm working on is a sales agent that will join meetings as an attendee. And so instead of in sales command center today, it will give coaching off of a rubric and scored sales calls. And it'll say, hey, you talk too much or whatever it's like after the fact. This one, I want it to be live coaching.
ADAM28:38So it's you'll have a screen open during all your sales calls and it will say, hey, you're talking too much. Don't forget to go back to the agenda. Ask these questions. You still have your med pick or your your whatever your methodology is, your med pick fields are weak and so you need to be asking these things. Ask this question to get them talking again.
ADAM28:59So things like that live coaching to where you can actually train the agent off of your top 1% of reps, like your best of the best, and you're increasing the floor of the capabilities of the rest of your sales team to where they're prepping for meetings. Like the best of the reps. They're they're asking the right questions like best reps would.
ADAM29:18And and you're actually up leveling everybody. So I'm super excited about that. That will be live on calls kind of guiding them through the sales process. And then the third and final one is a lot more technical, a lot harder and longer of an explanation. But long story short, as I build out a sales command center, a finance command center, operations and all these different pieces, each one requires its own infrastructure that overlaps a bit.
ADAM29:47But at the very end, when I finish all of those, I am going to then take build a what's called like a harness, like a data harness, but like for a technical term. But basically I'm going to have the back end working with agents and build a chat like interface that gives the power back to the people, to where end users, they could use my sales command center, whatever these might be the structure rigid, the structured and rigid applications, but they can build whatever they want, and it's off the curated, contextualized, cleaned logic and everything like of this system.
ADAM30:26And they can then go vibe, code whatever they want, and trust that they don't have to figure out all the hard stuff that I had to figure out the first time through. So that's a lot to chew on, I'm sure. But yeah, there's big things on the horizon.
RACHAEL30:40Quick pause. Everything we talk about on this show. Diagnosing go to market ops. Prioritizing projects for revenue. Impact processes. Metrics. Insights. Building a predictable go to market engine. We've built frameworks for all of it. They're free and undated on our website. Union square consulting frameworks. The link will also be in the show notes, so make sure you check that out.
RACHAEL31:03All right. Back to the episode.
RACHAEL31:06wanted to ask about the live coaching one. Do you, do you worry that that would cause salespeople to become too reliant on it and maybe lose some of those skills for themselves if, like the risk there for, for, you know, putting the training wheels on and never taking them off?
ADAM31:25Yeah. Interesting thought. I haven't thought of that. I mean, I would be surprised if they're so reliant. I mean, the agent's not going to be speaking for them,
ADAM31:33I mean, it's yet to be seen if that ends up being a concern. But again, I'd be very surprised. I think it's more
ADAM31:38right now, as we do coaching or even as a manager says one on one and tries
ADAM31:43train and coach his
ADAM31:44They're just trying to remember like, oh, this is a better way to do it.
ADAM31:47here's a better example. Let's say with your specific product, there's a common hurdle that is just hard to explain. Or maybe there's a dark patch in the company's history that comes up at you. Google search it a lot, so they have to explain it a lot.
ADAM32:02I do sales reps comes in and he just doesn't know a great way to get to that obstacle.
ADAM32:07The agent can be trained on the best way to explain certain things, and they can just give it little notes so it doesn't have to remember everything. But like, hey, they're talking about this hurdle and stuff like, hey, don't forget, you know.
RACHAEL32:19Mention this thing.
ADAM32:20So I think this is more of like a way to remember and, and speed up the ramp process and the coaching process. No matter what you go into a company, you have to learn all their products and all that. It's just you might be more sales driven than technically minded. So I think it's just a way to, like I said, give superpowers, the AI superpowers to these reps to help them.
RACHAEL32:42Through it. And I love that idea. I wish I had something like that just for even though I don't do sales calls or anything like that, just to have them for, you know, podcast calls or even podcasting is like, oh, hey, you forgot to ask this question, or you should. They can do this more. That'd be so handy to just have that right.
RACHAEL32:57Like, like tooltips essentially of like how to do a conversation better than you're doing it right now or oh man, imagine having that in just everyday life. That would be amazing.
ADAM33:06Oh, I need it for sure.
RACHAEL33:08Yeah. So you have a link to a demo actually, of how you made. I think it was the sale, the sales command center. Right. And I will include that link to the demo in the show notes. So if anyone's listening to this in the car or anything like that where they can't access link right away, just be sure to go to the actual podcast episode page either Spotify, Apple, wherever you're listening to it.
RACHAEL33:32We also have individual podcast episode pages on our website for each episode, and you can find the link to that demo there. Adam, are you going to be creating more demos for the different stuff you end up creating in the future?
ADAM33:45I could I mean, probably it wasn't that hard to create that demo for anyone who watches it. You'll see. Okay, Adam used AI to create the video of the AI app that he created with AI, and so you'll notice that even the narrator's voice from 11 labs is. Sometimes the accent is a little too strong to hear understand what they say.
ADAM34:06So yeah, like it only took me 30 minutes or so to create this short one minute demo video. So in is less on like how it was built. It's more of just like here's examples of walking through and this is all dummy data. There's there's nothing sensitive in there all locked up. But yeah, it gives people the idea of the world of what's possible really.
ADAM34:30But yeah, I mean, I probably should create some demos for anything else that when it gets completed.
RACHAEL34:36That'd be awesome. I like I personally love to watch those and learn more about it, especially, you know, as the world keeps turning towards AI and it's that adapt adapter get left behind thing. Right. You got to learn learn how to do this AI stuff or eventually somebody will get hired over you. Who who knows it, right.
RACHAEL34:55how can this now that we know, like the use cases and stuff and what we can be doing with AI, how can a crew or CFO or other revenue leader or other person wanting to build these things, get their data in a place where they can efficiently and realistically build something like this and have it spit back something accurate and actionable and helpful.
ADAM35:16Right? And that's the right question to ask.
ADAM35:20Just because AI is here doesn't mean that you can use it today. If you don't have the right data maturity and the system and architecture set up, don't think that you can just slap it on. Maybe one day. That would be incredible. I would love that. If it was smart enough, you could slap it on your the super dirty CRM and somehow it magically knows how to.
ADAM35:41That all correctly. It is not today and I have not seen anything close to having that work properly. So there is a right way and a wrong way. It does require significant focus and priority and investment having the right data team and but it depends on the size of the company. Also there's a there's this great book called Revenue Architecture.
ADAM36:08And in this book they talk about different size companies and really where they should focus in the in that stage. So they mentioned if you're a $50 million revenue or below product quality and really nailing the product matters more than like your go to market or your strategy and your analytics and all of that, they then say, if you're 50 to 500 million in that range, product still matters more, but it's a lot more balanced.
ADAM36:35You know, go to market matters a lot more than it did. And this is also not just at the company level, but if you have multiple products, completely different like business units broken out, I would say you should look at this at the sublevel as well. And then if you're over 500 million, that's funny. When it switches and go to market and strategy and execution analytics all matter more than the product.
ADAM36:58By then you've really nailed the product and you're still iterating. But, you know, hopefully you have traction by then. So what I'm about to say like this, this advice, this input is based on the company size. Ensure there in. Also the caveat is this is one way I can give an example of like here's one way you can do it.
ADAM37:19Any tools I mentioned I mentioned power BI a lot. That's one option. Tableau is fine. I'm not a fan of looker at all. I would say that doesn't scale. If you're not looker, I would just get off and go something more mature. But there's a there's at Databricks is is amazing. You don't have to use all these tools.
ADAM37:37So there's more than one way to do it. Building out the data team properly is a really important one. I think it's overlooked that I'll definitely have some continuing views on. But one of the caveat to mention is is it changing quickly? Like I said, if you if we have this conversation in December, it would be very different conversation.
ADAM38:00If we had this conversation next January, it would probably be a bit different. So like it's changing so quickly. So this is as of today, one way that's worked. And I can say confidently that works because it's been successful for me in my career. So the samples out there, so get that out of the way. There's some caveats.
ADAM38:22There's so jumping into like where you can get started. There's something you can look up Gartner's Data Maturity Scale. And it will kind of just show you different levels of your data maturity. Are you able to identify what happened in the past? Can you tell why did that happen? Like diagnostic analytics? Can you are you good at forecasting forward looking like what will happen in the future.
ADAM38:47And then like the max maturity would be? How can you make it happen like your prescriptive analytics. So this is your forecast. How can we take that to next level and actually execute effectively. So and that's not again at the whole company level. That could be certain products, certain functions finance might be at really good at what will happen.
ADAM39:08And rev ops is still stuck in the past of like what happened is the best they can do. So find out where you are today and then you need to figure out who's going to own this, like who's going to get us to the next level. So I believe is investing in a strong data team. If your SMB or startup, this piece is probably less relevant.
ADAM39:31I would still say go with like a consultant or something to get you up and running part time. Something like that for power BI. I guess a shout out would be there's a great consulting company called Jensen By. I've worked with a lot of power BI consultants and I've been disappointed or underwhelmed a lot. Jensen by is the exception.
ADAM39:53They have been absolute experts in that field of building the business logic. Super impressed to where I have to plug their name in here. So. But build a strong data team and this is the part that some disagree with. But I am very firmly confident in my answer. Where they sit matters so much and who you hire. So I believe data goes to it to die and a data team that lives in it.
ADAM40:28I personally haven't seen it to where they sit in it, and all stakeholders and C-suite and the CFO is like, we have data figured out. I'm happy all my needs are met. Like I've never seen them be like super satisfied. Now nothing gets it. But there's different functions of data analytics. I differentiate in two different ways. There's data engineering.
ADAM40:52It is a lot more like a IT function. They own your snowflake data warehouse, all of that. And then there's like the analytics and reporting data science that I firmly believe these functions need to still sit in something like operations. They roll up to the closer that they can roll up to the CFO. That's really fine. But show me where your data team sits, and I'll tell you how much you are prioritizing your data.
ADAM41:18In my opinion, my perspective. So it's just critical. If you're hiring only data shouldn't be your submitting tickets. And these guys are service roles only their support roles, they're seen as such. If you can hire diverse backgrounds of someone from management consulting and go to market and operations and finance and whatever, but also have the strong technical skills, they can be seen as a strategic partner that you wouldn't dream of doing, partner or planning, annual planning or any strategic decisions without them having to see the table because you're so much more accurate and effective with the team by your side.
ADAM42:00So and then the last layer is your data stack. Your tech stack matters. But you know, a lot of details that could go in there.
RACHAEL42:09Yeah. And I'll plug our own content here too. We talk a lot about the different maturity levels of go to market operations, a little bit different from Gartner's data maturity levels. So both are super. You put them both together. And I think that's a super well rounded view of like maturity and how to get from one Richard stage to the other.
RACHAEL42:30The link to our frameworks will be in the show notes as well, but it's at Union Square. Consulting frameworks. We talk about diagnosing go to market ops for all the different channels and motions that you have, and what maturity level they might be at our maturity levels. Talk more about. The things, the processes and systems you already have in place and what that tells you about what what you are capable of doing with your data maturity.
RACHAEL42:57And so, yeah, I think that the Gartner data maturity dovetails really nicely into the stuff that we talk about as well.
RACHAEL43:06when you talk about like building this out and building this data architecture. Do you have an example that has worked really well for you?
ADAM43:17Yes. Again, it's worked really well for me, but not the only way to do it. What's worked really well. And again, going back to the contrarian views here, I firmly believe in a centralized team. There could be rev ops as an analytics person, and HR has one and finance has one. And recruiting, they're all fragmented. They end up being siloed and they end up replicating work and duplicating different forecasts and contradicting each other.
ADAM43:42And and I've just seen that as a mess where everyone has a different definition of everything. Not only is that a hurt to productivity, but it's
ADAM43:50I mean it. You can't get more advanced because you can't get past the basics on some of these calls that you're in. So I believe in centralizing it again under ops or CFO or wherever, but you centralize the single team but also centralize your data model.
ADAM44:04So in power BI, anyone listening who's familiar with power BI or feel free to share this episode with your data team and they'll know what I'm talking about. When you build a dashboard, it automatically will create a data model or they call it a semantic model. Most people do it that way. You have tons of data models for every dashboard you created, so you end up having hundreds of them one day.
ADAM44:26We didn't go that route. There is a way you can create a single data model in power BI that that hosts all the company logic. So go to market product finance in one place and it all works together. It has relationships on how it should communicate with each other. So it's it's things like.
ADAM44:49We we change the definition of what we're going to define pipeline generation at some point. So in order to clean things up we clean it up in the logic in within our data model. And we have a single one that now that feeds all dashboards. So if you need to change the definition of pipeline generation change, instead of finding where it might exist in every place, you're probably gonna miss some change in one place.
ADAM45:14And the Cascades throughout everything. But now dashboards are still critically important. Don't get me wrong, some people might think that it's no longer important, but still mission critical to have dashboards, and I'm happy to explain why. But now there's more data products that can be built more than just a report or a dashboard. Now, the AI applications we were talking about sales command center.
ADAM45:39It's built off of our power BI semantic model, our data models. It's built off that centralized logic. We're not defining bookings over and over and over again of like what the logic should be for that. It's still in one place. So there's a lot of power in centralizing that single source of truth across each function. But again, I think it needs to be said that data visualization, some might think in the world of AI that it's just why would I need that is it's irrelevant.
ADAM46:14And an important point is the way I see is there's two different kinds of dashboards, or at least two, maybe three, but at least there's exploratory reports is what I call them. And management dashboards, explorers and reports are like, here's a single KPI or a related group of KPIs. So we have one for bookings, one for pipeline generation, one for open pipeline whatever.
ADAM46:34And it's just non disputed. And you can go there and you can slice and dice and filter in a million different ways and export the raw data if you want to into Excel. So everyone has visibility into the governance centralized. Confirm that definition of a single KPI. That really helps if something changes or breaks or whatever, like you know about it really quick of and you're able to fix that or adjust those KPIs, get used in what I call a management dashboard.
ADAM47:02So we have one that we call plan coverage. So two hours every Thursday morning, there's a lot of C-suite members and all sales leadership and a lot of people on this call where we review in great granularity things like Causeway bookings compared to target, compared to open pipeline versus the forecast versus your gap to call. And all of these things deal by deal.
ADAM47:27Your top 25 deals. And I'm stuck here and someone can raise their hand and say, hey, I know someone there. Let me give them a call and, you know, kind of collaborate. Everything has a target at the deepest granularity, which is incredible part of that dashboard. You don't want that to change too much. You want to look at that target in that target.
ADAM47:49Exactly. Every time you have that call. Now that's deterministic results one plus one equals two numerical value to every time. AI not only does it need clean data and context, that's going to come from your semantic model, but it has probabilistic outputs. If you ask ChatGPT today, one plus one equals what? That's a simple example. I was probably going to say two the numerical value.
ADAM48:18Two you would hope, but it could say 12, it could return your question and the answer and say one plus one equals two. That variability doesn't work for some use cases. So I want to make it perfectly clear, like the world of dashboards and reports as it is today, is evolving to more data. Products can be built. And I think power BI is going more heavily towards the semantic layer.
ADAM48:42Data modeling side and dashboards are one option. I still think it's absolutely mission critical, and it's the foundation bedrock that AI should be built off of.
RACHAEL48:52That's that's so much to digest, but I feel like I understood it all. And so with all of this capabilities with AI and all this AI stuff happening, a lot of people are worried like, oh, my job, you know, is going to take my job. Everyone's jobs are doomed, essentially, because I can do everything. But what's your case that companies should invest more in their data team in the AI era and not less?
ADAM49:19Yeah, I think that just like SaaS companies are in the adapter die stage, and I would say every single one of them is very aware of that. That's not new news. And they're all talking about it as we hear the word AI thrown out. So much individual people also are in the adapter die stage. If you're not at the cutting edge of what's out there, if you're not like on a million different email subscriptions to like, find out but the latest things from the right people and stay up to date on the latest information.
ADAM49:51You are at risk of becoming irrelevant. But what I built with sales command center and all these different things that are that I'm building out right now, they're really hard like they are. They they do take a steep learning curve. Some people think that's not that hard. I don't need I can do that myself. I'll just do that in 30 minutes.
ADAM50:09Yeah. You could ask cloud code to give you a wireframe mockup in 30 minutes. Probably. You're still going to iterate it if you want to go production ready a lot, but the back end, the data actually working and the AI component is actually working is extremely difficult. And right now a lot of there's not. Like I said before, perfect examples.
ADAM50:32Just follow this recipe, this work every time you need some dedicated resources. So I think it's a pretty simple formula. If a CFO right now or Col whoever, if you or I guess CRO as well, their biggest things like visibility, I would say one of the biggest ones. Do you feel like you have visibility in your data? Is in a good place today?
ADAM50:53If so, you're in a good spot. Don't worry about doing anything. I bet you most aren't going to say that. And if you feel like you're not keeping up on this AI revolution trend and you feel like you need more data, visibility into the pipeline and all these things, look at where do you have a data team? First of all, where do they sit?
ADAM51:14Are they centralized or are they fragmented out? And find out if you can make those changes. Invest more in the right people. Don't let it turn to like an IT service ticket taking function. Hire the right people to work and turn into an actual creative, analytical, thought inspiring, collaborative function. But the people who can adapt themselves to those skill sets, they'll never have a problem with employment.
ADAM51:46And at least in the next, who knows? AI could take.
RACHAEL51:49Over, could change absolutely lot.
ADAM51:52And who knows. But for now.
RACHAEL51:54Yeah, they're running races now. They got them running in races, so who knows. Who knows what. They come next. So we have time for one more question before we wrap. If a revenue leader takes one thing from this conversation and does it Monday morning, what do you think it should be?
ADAM52:12That's a hard one, actually.
ADAM52:15I would say
ADAM52:16either a
ADAM52:17figure out what your priorities are
ADAM52:18and like actual reflection or your needs being met today. So you can determine do you need to pivot, do you need to make a change or
ADAM52:25And then be the data maturity sale. Like
ADAM52:28if you're only able to do just reporting meaning what happened in the past.
ADAM52:33What's your forward looking capabilities and how accurate are you in the data science aspect is very powerful of like the why and what next piece.
ADAM52:47Find those holes. And I think I think it's the the people. The people you have on board is going to be more important than anything if you can get if you have the budget to actually invest in data, don't just go buy more tools. It's not going to give you the bang for your buck that you're looking for.
ADAM53:04And you might turn into like the the wasting token trap. Yeah, we're going to stuck in I would say look at the data team. Where do they sit? Do you have the right people. Hire the right people and let them get you to the next level? I mean, the culture around analytics could be off, but I would say it starts with the people.
ADAM53:27That's your base. That's the foundation.
RACHAEL53:29Absolutely. Well, thank you so much, Adam, for coming on the show. And if anyone wants to follow you, follow along with what you are doing with AI builds. Where can they do that?
ADAM53:39You're welcome to reach out to me. I mean, I keep an eye on LinkedIn messages. I'm not a big like poster on social media or anything, but you know, you're welcome to reach out and ask questions. I can give some more targeted feedback. If anyone is trying to do the same thing, I've done that before many times already.
ADAM53:55So yeah, you can find me on LinkedIn. Adam. Coffee and feel free to reach out.
RACHAEL54:00Awesome. And that's coffee for anyone who's just listening and not looking at the show notes. But we will have a link to Adam's LinkedIn in the show notes as well, if we're able to click there. Awesome. Thank you so much, Adam, for coming on the show. Have a good rest of your day.
ADAM54:16Great. Thanks for having me.
EDDIE54:18Thanks for listening to the episode. If this resonated. Please give us a five star rating and a follow. It helps us reach more people, and you get our latest and greatest content without having to search for it. And if you're looking for hands on help and go to market strategy and or rev ops, please reach out to us.
EDDIE54:32We help our clients with everything from annual planning to improving processes and go to market, implementing systems to support those processes and go to market AI. We're always happy to offer a free consultation to help you identify the best opportunities to improve your go to market engine, with or without our help. You can find us at Union Square consulting.com and the info will be in our show notes.
