The Headless Future: Re-Engineering Your Tech Stack for the AI Era with Ben Edwards | Episode 264

Episode Summary

SaaS isn’t dying, but the way founders use it is about to change. Ben Edwards of CMap joins members to explain where professional-services software is heading: away from logging into platforms and clicking through dashboards, and toward “headless” access where your data lives inside the AI models your team already uses. Ben breaks down what MCP access means in practice, why your PSA and CRM are sitting on operational data more valuable than most founders realize, and how to rethink your tech stack so that value isn’t left trapped.

About the Guest

Ben Edwards

VP of Consulting · CMap

Ben Edwards is VP of Consulting at CMap, a Professional Services Automation platform serving 700+ boutique firms in consulting, architecture, and engineering. With 15+ years in SaaS, Ben leads CMap’s North American consulting practice and helps firms navigate the shift from traditional SaaS dashboards to AI-native operations. He is a Collective 54 member.

Key Takeaways

• What “headless” SaaS and MCP access actually mean for a boutique firm
• Why your operational data is an underused source of value, and how to surface it
• How to evaluate your tech stack when every tool now claims to have AI
• A SaaS insider’s view on which tools are systems of record worth keeping and which are not

Full Transcript

Greg Alexander: Hey, everybody. Welcome to the ProServe Podcast, brought to you by Collective 54. I’m your host, Greg Alexander. For those that might be new to this show, this show is dedicated to helping you do three things that is, make more money, make scaling easier, and make an exit achievable. And this show is recorded and built and designed and delivered exclusively for boutique professional services firms. So if you are in the expertise business, you market, sell, and deliver expertise, this is for you.

So this week we’re going to have an interesting conversation around the tech stack that boutique professional services firms are using. And the reason why we’re doing that is because it is evolving greatly. And things that we have long considered to be best practice SaaS tech stacks, CRMs, PSAs, project management systems, QuickBooks and financial stacks, etc. worked for many, many years, but now how users are engaging them is changing. And it might be a time to reconsider the architecture of your tech stack as we make the migration from the tech-enabled firm to the AI-native firm.

We’re being asked this question a lot by members, so we’re very fortunate that we have a long-standing, well-respected, and well-liked member who lives in that space. His name is Ben Edwards, and he’s an executive with a PSA firm called CMAP. And if you’re not familiar with PSA, that stands for Professional Services Automation. So Ben and I were catching up a few weeks back, and we got on this topic, and I said, you know what, we should turn the recorder on, because this will be valuable to everybody else. So that’s what we’re gonna do today.

We have a lot of new members. A lot of these AI-native service firms have found their way to us, and they might not know you just yet, so would you please provide a formal introduction?

Ben Edwards: Sure. So I’ve been in the SaaS world for 15-plus years, four and a bit with CMAP, who serve 700 boutique professional services firms, largely in the consulting, architecture, and engineering space. And my role is across both sides of the Atlantic even though I do obviously come from England with my accent, I am now based in Orlando and helping our North American client base leverage our solutions and this wonderful, brave new world of AI.

Greg Alexander: Okay, fantastic. So there were a few things that you said to me when we spoke that really jumped out at me, and let me just get into those.

One thing you shared with me is that our members are sitting on a treasure trove of operational data that might be more valuable to them than they realize. And it’s sitting inside of legacy SaaS apps and they might be trapped inside of things like dashboards that maybe these days people aren’t mining as often as they once were. And you explained to me this new concept of the headless system and MCP access as a way to untrap that operational data, and it was a really insightful conversation.

So I just want to throw that over to you, and maybe just have you explain in layman’s terms what that means to everybody as a starting point. Would you please do that.

Ben Edwards: Yeah, so as a starting point, the existing data that we find is under-tapped and under-harnessed by professional services firms. It typically lives in lots of different silos that might be across a CRM, a PSA like ours, a finance system, your general ledger. You might have separate task management tooling, timekeeping tooling, spreadsheets. And there’s a huge amount that can be leveraged by a professional services firm if they could pull it all into a single place.

And that would enable them to get better visibility on how their engagements were performing, both from a time and a revenue or margin perspective. That could then feed back into your pricing, which then might impact how you create your SOWs. It might even start impacting your commercial model. We’ve started seeing a large number of businesses moving from T&M to fixed fee to more outcome-based pricing. Hopefully all of this improves your EBIT and your cash flow.

Where the MCP comes in is in the olden days, that would be like an API-based connector. MCP is essentially what the LLMs can use alongside traditional SaaS to surface a whole range of data points that typically exist in all of these different types of solutions. What that should deliver for a professional services firm is the ability to combine previously siloed data and access it in real time. And you can then ask very meaningful and real questions of that data and be able to harness it.

So instead of, for instance, coming to the end of your month, closing your books, and realizing that cash is not coming into the business as quickly as you wanted maybe a number of those projects that you thought on paper were going to deliver really strong margins have actually experienced scope creep or over-delivery, and that’s shrunk things. That pipeline that you were building actually isn’t as well qualified. On a daily basis, through that combination of data that lives in lots of places, and then the MCPs which live on top of your LLMs, you will get much faster and quicker real-time visibility.

Greg Alexander: So for those that are new to this what does MCP stand for?

Ben Edwards: Model Context Protocol. And it’s a phrase that I actually first saw originated by Anthropic and Claude at the turn of last year, but other LLMs and other SaaS solutions are harnessing those three letters to essentially talk to their customers which in this instance is people in professional services firms about how to layer that data that you’ve already got on top of something like an LLM.

Greg Alexander: Okay, fantastic. The term “headless” seems to be industry jargon that’s getting thrown around quite a bit, so what does that term mean?

Ben Edwards: So again, just a bit of scene setting. In the olden days of software, it was all on-prem so on-premise, you bought access to a solution and you had to install it. Then it went into the cloud, so that data transferred via, let’s say, Amazon Web Services or similar, and you got access to it.

Headless is now where, from a SaaS and AI perspective, the tooling that you’re offering your customers again, in this context, it’s professional services firms are the guardrails and the processes and the data entry points that would capture things like a pipeline opportunity, a SOW, a project, an invoice, data back about whether that had been paid. But it doesn’t have to be expressed in a dashboard. It doesn’t have to be interrogated by logging into your favorite SaaS application and seeing it within the rails that they provided in terms of a user interface, a user experience.

So it’s even more ethereal than the cloud. It’s essentially what you’re looking for is whether your SaaS solution, your favorite tech partner, can provide the robust guardrails and processes that add value. Because essentially the data that you’re adding in it might not be today for lots of C54 members, it might be in 6 months, it might be in 12 or 18 months, depending on their adoption journey but you would use an LLM to pull all of that data together and interrogate it, ask questions of it. You’re still paying licenses or credit-based usage to your favorite SaaS or tech partner, because they still have value they are providing these robust guardrails and processes but you probably just don’t need to be logging into a HubSpot one second, then a Salesforce another second, then a PSA, then a QuickBooks, because the UI piece is probably, over a longer period of time, going to be removed.

Greg Alexander: Yeah. So this takes me to my next question it’s an important one. I work with Collective 54 members trying to help them become AI-native. We have been working with them to stand up a database, and when they come to us they say, hey, SaaS is dying, I don’t need SaaS anymore, I’m gonna turn off all my SaaS licenses. And we say, that’s not true. Those systems are still really important. They’re gonna be the source of the data that gets pulled into this centralized repository. Yes, you’re probably gonna access it through AI tools, whichever one you prefer. But you still have to have those foundational building blocks underneath in order to pull this off.

So, first off, do you agree with that statement, or do you disagree with that statement? Either answer is fine I want to get your perspective on this. And then, since you are a SaaS company, you’re seeing more implementations of this than we are. What are you seeing in the market right now? Is SaaS going away, or is it actually becoming more important because the AI systems rely on it?

Ben Edwards: As often, you have hit the nail firmly on the head, Greg, and I do agree that you need the guardrails and the processes in place. Because that is where you’re going to see the fundamental gains that transfer to what you require from a business perspective. And lots of boutiques have grown exceptionally well through point solutions, spreadsheets, basic plumbing but then it reaches a breaking point, a tipping point, where you need to scale beyond the founder bottleneck or the key few partners. And it’s in that instance where data becomes way more important, because you’ve got to democratize it across your business rather than have it live in a single or a few people’s heads.

So what we’re seeing from a SaaS perspective is that there are fundamentally two major things that have happened in the last 6 months. Valuations on SaaS firms have gone considerably lower because a huge amount of money is obviously getting pumped into AI. And investors whether that be VCs, private equity, or huge funds are saying that there’s a SaaSpocalypse, and that then comes out in the media coverage that SaaS is gonna die, people revert to vibe coding tools to build their own solutions, and therefore on the front of it, it looks like troubling times for SaaS.

Speaking on our behalf, and for lots of other SaaS businesses I know, our investment in AI is unprecedented too. So really, there’s still huge value in a large number of SaaS tools that are out there in the market, but like professional services, we’re going to have a split where a percentage of the market survive and thrive because they pivot exceptionally quickly and they leverage the existing relationships they’ve got, plus the fundamental technology they’re offering is of high enough value. And that would translate to professional services firms who move quickly and adopt this brave new world and figure out some of the challenges that are in front of them right now and continue their growth and scaling.

And then there will be a percentage that don’t, and there’ll be the have-nots. There’ll be have-nots in SaaS, in the same way there’ll be have-nots in professional services, in the same way there’ll be have-nots in AI-first tooling. You know, if you look at hype cycles, we’re definitely if not hitting the peak, we’re just approaching a peak. These things often rhyme and happen time and time again, so I would not be surprised if there is a slight pullback on all the investment that’s gone into AI-first tooling, and it pivots back a little bit more towards SaaS because the valuations have compressed so significantly.

From a SaaS perspective, it’s all around ARR and then your NRR, which is essentially your ability to retain customers. And if those two things are still exceptionally strong, the fundamentals of that business and the value they’re delivering to the customer still exist. It’s down to the market to decide whether everything goes all in on AI, or people pull back a little bit and come back into SaaS.

I would say Q1 and Q2 were absolutely fascinating, because Claude Code in particular, and their ARR growth and revenue growth and user growth, just went absolutely stratospheric. So anybody who was watching that, along with tracking NASDAQ or their favorite tech stock prices, would have gone, well, SaaS is gonna absolutely disappear very quickly. The reality is that’s not the case, because actually a large number of clients still rely on SaaS to input high-quality data and then get high-quality outputs out of it.

Greg Alexander: Yep. So I agree with you. So let’s double-click on this a little bit because I want to give everybody some takeaways.

Many, many, many SaaS firms are going to go away. The reason for that is that they’re not a system of record. The free market is gonna do what it does those that have real value are actually gonna thrive in this environment and become more important to clients. And those that are just neat little nice-to-have tools are going to get killed because they’re not a system of record.

But many of our members didn’t grow up in the IT world like you did, like I did, so they don’t know what the term “system of record” means. So when they’re looking at their general ledger and asking themselves, okay, here’s the totality of my spend on my tech stack which of these systems should I keep, which should go away they have a hard time making that distinction.

And what’s making it worse is there are all these newly formed software companies that claim they’re AI-native from the get-go, and they’re saying, ditch the old guys and come with us we’re the new breed of these systems of record. They get excited about that. These are really compelling sales pitches. And then they sign the deal, and then they’ve got to do the migration, which is not easy. So before you make the decision to ditch the old guys and go with the new guys, you’ve got to understand what it’s going to take to go from point A to point B, and is it truly worth it?

So would you spend some time explaining to the audience what “system of record” means, and maybe highlight a few categories of the tech stack that fall into that category, and offer some wisdom so that our members don’t just fall victim to a great sales pitch?

Ben Edwards: I always think a great system of record has two things. It has some intrinsically built guardrails and processes that take a user from A to B that they wouldn’t have if they were using pen and paper, an existing spreadsheet, or whatever tooling they had. Like, it enables them to build really repeatable processes. Professional services in particular is human capital it’s not like we’ve got a factory that we can churn things out of. Therefore, embedding great process has value.

And then the second thing, I think, from a system of record is intrinsically adding data into that solution. It could be qualitative data, like you get from an AI call recording solution. It could be quantitative data the pipeline value in a CRM, the project value in a PSA, the invoice value on a GL.

I think there’s a larger number of solutions in the market which build on top of that and will take data that has been input somewhere else into a data visualization layer. Power BI, for Want of a better phrase, is an exceptional tool to do that. It’s a great visualization for anyone listening, to compare what is a system of record versus more of a data visualization tool. There are tools like HockeyStack, which is a marketing and sales enablement tool, and many others, which essentially take data from multiple different sources and provide you some insights. Those are the things that can be vibe-coded exceptionally quickly, more of a point solution requirement, or a point in time where you want to get some insights.

Systems of record are what you will build the business on. What is the enterprise value built on? So, for us as a SaaS firm who’ve got private equity investment, our systems of record showcase pipeline value, our ability to win new opportunities, the value of clients once they become a CMAP customer, the ability for us to invoice those businesses on a recurring basis. There is absolutely no chance we would even try to vibe-code or build any of that tooling, because that fundamentally underpins the value of our business and how other investors and our management team see and access that information.

And in a professional services sense, enterprise value is going to be built on your growth rate, your EBITDA which is going to be based on your utilization, your project margin. So things that fundamentally impact the enterprise value. You absolutely can try these days to build your own tooling around that, but to your point, the risk-reward ratio is the one I would question. And the investment in time and energy to build another cottage solution versus just leveraging something that exists already in the market that’s worth considering carefully.

Greg Alexander: You know, one last well, I guess I have two follow-on questions to that, and then we can wrap it up. I appreciate your wisdom here.

Our framework for those that might be new to this is what we call ERA 1, ERA 2, ERA 3. Very simply, ERA 1 is a professional services firm who sells and delivers work through human labor. ERA 2 is a professional services firm who sells and delivers work with humans that are enabled through technology, so they have higher rates of productivity. And then ERA 3 is different people who sell and deliver work, we call it the 80/20 split, where AI is doing 80% of the work and humans are doing 20% of the work. The human judgment and wisdom in that 20% is infinitely more valuable. But it’s different in that the AI tools are actually completing tasks that’s the difference.

So if you just think of that as a fundamental building block the challenge that we’re having is that we have some members that are still ERA 1 firms. They never made the move when they should have to go to ERA 2. Unfortunately, regrettably, that is a fact. So what they’re doing now is going from ERA 1 to ERA 2, and they’re installing legacy systems that are really hard to install, and I think this is a mistake. I think they could skip that step and go to the ERA 3 tech stack, however you might define that, and save themselves quite a bit of grief.

But that’s my opinion not everybody agrees with that. For example, I have one member right now who was going through a brutally painful implementation of NetSuite. And they got upset with me because they’re a half a million dollars into this, and I told them to pull the plug. Because I think it’s going to take another half a million dollars to get to the finish line. And what’s the point? You can probably go to a new solution for a fraction of the cost and be up and running in 30 days. But that’s a very controversial and strong opinion, which unfortunately sometimes I have too many of.

Would you advise those members the same way that I am? Or would you suggest it is the natural way to go go from ERA 1 to ERA 2 first, and then from ERA 2 to ERA 3? What’s your opinion?

Ben Edwards: I think there’s a huge amount of opportunity for firms who truly understand what ERA 3 means to go to that point. I will be honest I see a number of firms employing AI experts from outside of their business on maybe a consultative or fractional basis to pull them to that point, without maybe the senior execs fully investing and knowing what it is that they’re trying to achieve. But I have equally met a really good number of founders who know it themselves, and if that’s the case if you know it yourself and can make that jump then that would obviously be the most logical way to go, because like you say, you miss probably the expense and the time commitment that’s associated with going through 1 to 2. You can go straight from 1 to 3.

There’s probably just a number of businesses who just aren’t at that point, though. And to them, I would say something is better than nothing, and the earlier you start collecting data in whatever form will be of high enough value. Something like going from nothing to NetSuite is huge, and yet it still only puts you at ERA 2.

And the advice would be just to gather as much as possible on everything from quote to cash. And if you feel like you’re missing some data points in that process even if it is pen and paper, even if it is spreadsheets just start capturing it, because where AI will become really powerful is if you have a richness in your data set. Which is actually where this conversation started in the first place. You really need to have a good volume of your own data to provide some differentiated value as you grow and scale your business.

I mean, you will have done this when you were in your consulting firm, Greg. You will have been regimented with people making sure pipeline opportunities were up to date, the value of that was up to date. You did your deal desks where you were doing SOWs, but you weren’t letting salespeople run amok, because you had delivery people making sure the margin was tracked. You were probably hot on resourcing, hot on everything invoicing-wise. You just need to build that muscle memory to embed as much data across your firm as possible. Even small things can be hugely valuable in the next 3 to 6 months, because you could then look back on it and go, wow, those 10 projects we ran we actually have a really good sense that those were the work stages that were highly efficient versus those that weren’t. This is the person that’s really running hot versus this is the person that’s constantly doing things on the side of their desk.

It doesn’t it’s not a silver bullet, unfortunately, and the fundamentals of running a professional services firm are still what they were 5, 10, 100 years ago. But in those days, the data was living in people’s heads, and you could make very quick judgment calls based on what you intrinsically knew. Nowadays, with SaaS and with AI, you have the ability to collect it all and democratize it, and enable the rest of your business to leverage it in ways that are really, really powerful. But without that, you’re kind of running blind.

Greg Alexander: Yeah, I agree. And it’s a good counterpoint. And that is good advice. We are in the AI era, which means data is the gold. And everything you can do to collect all forms of data is the answer. And for every firm, it’s probably a little bit different.

All right, let me end with this last question, and that is somewhat of an unfair question, but I’m going to ask you to look into your crystal ball and maybe look out three to five years which these days could be an eternity. But where do you think all this is headed?

Ben Edwards: I think there are some really interesting niches in AI-first tooling that for professional services firms is unearthing data and insights that were previously lost in the ether. And that is things like your call recording technology, and being able to really understand what your customers are saying, what your team are saying. Having access to that qualitative insight just wouldn’t have been possible 5 or 6 years ago.

So that combined with all of the existing quantitative data that has been around for over a decade through existing tooling. The melding of that, and then the layering on top of your favorite LLM, or benchmarking data like you guys provide, with some of the co-pilot type solutions I think that will be absolutely fascinating. Because to an earlier point, for the haves, for those that are really wanting to succeed, you’ve got absolutely everything at your fingertips.

I do think, unfortunately and this is the same for a lot of businesses, professional services, SaaS, and no doubt others if you’re too far behind the eight-ball, you won’t be thriving, and you will unfortunately be left behind. So I think there’s going to be an ever-increasing polarization of success stories and failures. That’s probably my prediction.

Greg Alexander: Yep. I tend to agree with you. Alright, Ben. Well, as always, this was very helpful. I appreciate you constantly making a contribution to our community, so on behalf of the members, I just wanted to publicly say thanks. And we look forward to the private member Q&A session that we’ll be doing in an upcoming role model session. I’m sure they’ve got a lot of questions for you, so I appreciate you jumping on.

Ben Edwards: Thanks ever so much for the opportunity, Greg. Really cool.

Greg Alexander: Alright, just a couple calls to action. So if you are a member, look out for the invitation we’ll be sending to you we’ll have a private member Q&A with Ben, and you can ask questions directly to him. If you’re not a member, and after listening to this you think you might want to become one, just go to collective54.com, fill out an application, and we’ll get in contact with you.

And then if you’re not ready for either of those two things, and you just want to consume some more of our content, I would direct you to my new Substack that is at Greg Alexander, C54. And that’s where I provide my commentary on the industry. Thank you for your attention. Until next time, I wish you the best of luck as you try to grow, scale, and someday exit your firm.

Related Episodes