Stop Your Revenue Leakage Now

‍Jake Aaron Villarreal:
We’ve seen that, in this size of company, anywhere between two to three percent of revenue is actually leaking. You’re missing the fact that the discount you had for this three-year contract was only supposed to last for one year, but that has carried through.

I’m Jake Aaron Villarreal, born and raised in Silicon Valley, and here to take you behind the scenes to share what it’s like to be a startup founder: the journey they’re on, the problems they face, and the problems they solve in an effort to make our lives better.

I’m excited to have with us today Miguel Vasconcelos, the CEO and co-founder of TechTorch. Miguel, welcome to the show.

Miguel Vasconcelos:
Thank you, Jake. Great to be here.

Jake:
I’m excited to have you.

Just a little bit more on Miguel: he has over 20 years of experience in strategy, operations, and digital transformation. He’s built a company that’s helping private equity firms and their portfolio companies harness AI agents to supercharge revenue operations.

Miguel, before we jump in, where are you calling in from today?

Miguel:
I’m calling in from Portland, Oregon, although I’ve spent quite a bit of my career down where you are, Jake, in the Bay Area.

Jake:
Very cool. I love Portland. Lots of breweries, lots of food, lots of people, and lots of opportunities to innovate there, too. I think you’re in a good spot. What was it that pulled you there?

Miguel:
Initially, my career. I spent a few years in Silicon Valley working with a lot of the larger companies down there. Then I got a call from Nike to come and join them up here in Beaverton, which is just outside of Portland, to drive the big transformation they were doing from being a company that primarily sold through Dick’s Sporting Goods and other chains into selling direct.

I led all of that transformation up front, on the front end with the app and the commerce-enabled website, and then on the back end too. It’s been 11 years since I moved here, and I spent seven of those doing that for Nike.

Jake:
Really cool. Before we talk about your organization and what you’re doing today, let’s go back a little bit farther. Where are you from originally, and what were the things that drove you to get into technology?

Miguel:
That’s a great question. Going far back, I’m born and raised in Portugal. I’m from the north of Portugal, from Porto. I did my undergraduate studies in economics in Lisbon. After that, I’ve been a little bit all over the world.

With my background being in economics, the first job I took was in London doing investment banking. But it turns out that was the beginning of this all. The industries that I covered at the time were TMT: technology, media, and telecom.

From the very beginning, I spent three years at J.P. Morgan in London. That’s where I started to research and understand what made those companies work, who the winners were, and which companies were not doing so well.

Then my interest pulled me toward wanting to do more than just research and trade on these companies. I wanted to actually do the work behind the scenes to help them be more profitable or more successful.

So I called back Bain & Company, which had made me an offer out of undergrad, and I said, “I’m ready to join now,” three years later. I started in Madrid at that time, and that’s basically what I kept doing.

I did a lot of work on the telecom side and later on the technology side for a couple of years in Madrid, and then in Australia with Bain, in larger transformations.

I always had it as a personal and professional objective to come to the U.S. to do my MBA. I came to Chicago, and it was a great experience. I met my now-wife there, ended up moving to the West Coast after that, and continued that career, very much in the Bay Area, in San Francisco, working with larger and medium-sized technology companies.

I just continued to love that. I’m not an engineer by training, but I love technology. I love to learn about technology, and I love to learn about the business side of technology.

Jake:
There’s a lot going on right now in the business side of technology and the transformation of AI all over.

You started in that space, and now you launched TechTorch five years ago with your co-founder. When you did that, what was it that you saw? What was the gap in the market that made you convinced there was an opportunity to tackle? What was the main focus when you began, and where is it today?

Miguel:
I’ll give maybe 30 seconds on my co-founder. We met at Bain when we were doing larger work with big clients in the Bay Area, and then our careers diverged. I came to Nike to continue doing that transformation. He was still down there in the Bay Area, where we’re headquartered, working at the time at Cisco doing the same thing.

What we saw, Jake, was something that I think now is pretty evident. There is this large technology services industry. We believe it’s about a trillion dollars of work in a big market that hasn’t really changed in over 20 years.

The big changes in that industry have been: how do we start to rely on work outside of the U.S.? How do we outsource, offshore, or nearshore some of that work into India and then into South America?

But the way the work got delivered was still very similar. It was all about a very custom approach. What is so special about your company that we need to design for it and then implement a new system, or implement some of the SaaS systems out of the box?

What we believed was that, especially for midsize and smaller companies, they cannot afford the big dollars you have to pay the large implementers, or the time that often takes. Now, me as a small business owner, you need the results now. You needed a system yesterday.

The idea of, “We’re going to design this for two months, and then we’re going to start implementation, and it’s going to be a year before you can see any of it,” was the problem.

We saw an industry that hadn’t changed much. We saw the opportunity for smaller and medium-sized companies that needed to do that transformation, that needed to change the way they work, and we thought there had to be a different way.

Our approach was one where we believed in preset recipes at the beginning. This was pre-AI, or pre the big explosion of AI. We started defining very clear recipes on what it really takes to implement Salesforce or to implement a data warehouse.

As we approached smaller and medium-sized companies, the idea was: how can we do that faster, more efficiently, and at the end of the day, with clear results for you?

That was the genesis of it all: how do we disrupt this large market that hasn’t really changed and make that digital transformation, that digital change that a lot of companies are still going through, available to a broader market?

Jake:
We talked a little bit a few weeks back, and I love the space you’re in. You’re very specialized. Talk about the market you’re really focused on today.

Miguel:
As we started and evolved, one of the first things we wanted to understand was where our experience, what we knew well as a business and as a workflow, could really help us differentiate, and where there was a clear impact from the changes, either through systems or now AI, that you put in place.

We landed on the go-to-market space. Think about go-to-market revenue management: what a company does from getting leads and business in on the marketing side, to putting a quote or offer out to a client, closing that, and collecting the cash. Obviously, there’s work you have to do in the middle, but that was the area we really honed in on.

We both had experience in it. We knew how manual, in some cases, and how difficult, in others, it was. There was still the opportunity there to have a clear impact, either on increased revenue or on a more efficient workforce that could run that part of your business.

So we’re really focusing on that space. As we started doing the work, a lot of that involved how you improve and optimize the work you have in Salesforce or HubSpot, or some of the big SaaS companies that operate in that space.

We ourselves started using AI from the beginning of our platform. We initially had an internal platform to capture those recipes and have a clear step-by-step for our practitioners to get that done.

We then started moving into how you, as a client, can also leverage what’s out here in terms of large models, how ubiquitous they’ve become, and how they can really help with some of the manual work.

The way we’ve done that is by creating agents very focused on these workflows that we believe are 80% ready. They are working pieces of software or agents, but as you deploy them in your environment, they are connected to your Salesforce, your data warehouse if you have one, your ERP, and your marketing or lead sourcing system.

They can start to take some of that manual work away, make some of that work more consistent across sellers or sales support teams, and deliver that faster and more efficiently toward the client and their client.

We’re very focused on that space. What has helped us become relevant and have impact quickly is that we’ve always had the approach of: how can you make this faster? How do we disrupt an industry that was time and materials?

We wanted to do it faster and get out of there, but you also have to understand what we call the mess, or the spaghetti factory, that often lives in these companies with the current systems. We understand that, and then we build not around it, but knowing it, while abstracting from it and creating simpler ways for those sales leaders and sales support teams to interact with the data and the flows that exist today.

Jake:
There are a lot of services out there at the Big Four that are driving a lot of transformation. What’s unique about your value proposition and your service? Is it software in addition to service? Is it agents only? Paint that picture a little bit more.

Miguel:
We are squarely at the cross of those two: a traditional SaaS business that sells subscriptions and a traditional services business that sells a project and an implementation fee.

The benefit we have, and I think what’s making us special, is that we started and grew up with that paradigm already very present. We’re not like one of the Big Four, where for 40 years they’ve been selling services and they have a very efficient model. They are very good at that, and now they’re caught in the middle of, “What does that mean for us? How can we be more efficient? How do we really respond to what we’re seeing in the market?”

We grew up with, “This is the reality. We have to support these companies, and the way we’re doing it is by deploying a lot of agentic workflows or agents that we’ve created.”

But we do that in a way where, once that agent is deployed, we believe that agent is the intellectual property of our client, and they will go with it. We are deploying that and doing that upfront work.

In some cases, there are clients that tell us, “That doesn’t work. It’s a lot of cash up front,” or, “I want to really have your skin in the game.” That’s another thing we’re pretty proud of. We do a lot of that based on outcomes.

Why don’t we do this as we’re seeing the outcomes in changes of what you can automate, how much time you can save, and how many dollars you can save? Then you get to pay us an additional fee.

To be frank, I think a lot of the industry is still trying to figure out how to charge for this. What is the right pricing model? We’re experimenting with a couple of those.

But I’ll tell you what we believe and what we’ll continue to push on: tying those fees or subscriptions to a clear outcome. That’s what got us started, and that’s what I think will continue to move us forward.

Jake:
I love that because it really ties your risk and their risk together. If you can prove that you can deliver and you can show the outcome, it’s worth the payment. Many times, it has to be three, four, or five times what their investment is, in terms of the outcomes you’re providing to them.

The ROI is there, which is really what a lot of companies are looking for. We’ve all read the MIT reports about the 95% of POCs and projects that have millions of dollars invested and no clear outcomes or benefits.

I think that really has changed, and we’re seeing a little bit more of this. As a company, we talked about this a little bit as well. We had to change our model to adapt to what companies are looking for on the talent side.

It’s the same way now. We’re looking at it as more outcome-based. The risk starts when someone starts working for you, not prior to that. So don’t pay us upfront when they start. Pay us as they hit milestones that you deem as milestones of success.

From that perspective, it just feels better. At the same time, you really build a trusted partnership for the long term. That’s really what we’re seeing too. I think we’re both aligned in the same sector.

Miguel:
Very much so.

Jake:
A little bit different, in that you’re delivering product and service, and we’re delivering people to products and services.

You said your sweet spot is PE-backed companies in the $150 million to $300 million ARR range. Why is that segment such a strong fit for your model?

Miguel:
If you take a step back and go to what our big beliefs are as a company, and what I think initially got us here, it was that focus on outcomes and speed. There’s a word I like to use: pragmatic. Not being purists. You don’t have to design the perfect system. You have to design the system, or in this case the agents, that really work for that company.

Having that approach, when we initially started five years ago as a pure platform company, one of our first clients, who is still engaged with us today, was private equity-owned. They loved our approach and the fact that we delivered on the results they needed quickly.

That word of mouth carried through, and we realized there was very strong alignment between our beliefs as a company, the size of those companies, and the challenges of those companies in terms of how quickly they have to turn their investment into clear ROI.

That allowed us to start developing an offering and a solution that was very tailored to this segment of the market. That was initially how we found it, and then that carried us through to good word of mouth by delivering and doing a good job, but also by getting very concrete and specific solutions to the challenges of that market.

We all like to think that we have the better recipe, but what we have is really something that works for this segment, for what they need to do, the challenges they’re facing, and the speed at which they need their solutions.

Jake:
What are the most common operational pain points you see inside portfolio companies when it comes to revenue, quoting, billing, and things of that nature?

Miguel:
There are a few. If you start from the overall view, we call something — and it’s a term you’ll probably see in the industry — revenue leakage.

If I’ve contracted something with you, how the elements of that contract end up getting captured in what you invoice and then in what you collect as revenue can create gaps.

We’ve seen that, in this size of company, anywhere between two to three percent of revenue is actually leaking. You’re missing the fact that the discount you had for this three-year contract was only supposed to last for one year, but that has carried through. Or you missed that there was an upsell on the contract, but the bundle was not being charged the right way.

There’s a component of revenue leakage that materializes when the information you had on the contract was not captured the right way initially in your system. Then it didn’t get passed the right way to the following system. The client is not being charged for that, and you don’t even know. The client is not happy or sad about that. They’re just saying, “Okay, that’s the same price I was paying last month. Great.”

The other one is the fact that a lot of these steps still involve manual work. Sellers like to be selling. They don’t like to be entering information into Salesforce or HubSpot. Then you need to bring in a third party, or in this case another team in the company, to enter that information.

Those handoffs are not perfect. There’s always going to be information that gets dropped or is not complete, and that impacts what you’re really able to do. At best, you can do all this well, but you’re going to have to add more cost to it. You’re going to have to have a team just to cover the information and put it in the right spot.

Finally, the other one is somewhat related to all of these issues. You have manual interactions, multiple handoffs, and multiple teams. The data you end up getting in the form of a report at the end of the week, in a sales flash or pipeline flash, is not data that an executive can really trust, or that the investor can really trust.

That’s another piece we find ourselves supporting quite a bit: how do we have trusted and clean data so that the investor and the executive team can make the right decisions?

This is a problem as old as time, but with all the systems and all the manual work, it has become even harder. As companies are growing fast, you’re probably not thinking about, “Do I have the perfect information?” You’re thinking about, “How do I make that next sale? How do I get to that next client?” This becomes an afterthought.

Companies that have grown fast and are now at a point of scaling need that information. Or as they are acquiring other companies and private equity firms are bringing them together, companies are very different. Now you need to normalize that data.

There’s a component of trusting what you can see so you can manage that company properly.

Jake:
I love the space you’re in.

I used to work at Oracle, and when I was there, there was a gentleman by the name of Ray Lane who was the president of Oracle. It was a time when Oracle almost went under because they had booked a lot of contracts and recognized the revenue within the year they booked it, even though it might have been a three-year contract.

They got into some financial issues, and Ray Lane was brought in to help build and scale the company. Long story short, when I joined, they had 8,000 employees. Today, I think it’s 150,000 or more — maybe much more now.

Anyway, he’s got a venture capital company called GreatPoint. They had an event I went to a few weeks back in San Francisco, and they were asking, in a fireside chat, what areas AI is having the most impact in. One of them was private equity. It was one of the pillars where companies within the portfolio are realizing that AI is a real transformative technology and system that can help organizations be more efficient, more profitable, and all that.

It sounds like you’re in a really good spot when it comes to opportunities. How important are sponsorships or the customers you have to continuing to grow your customer base in terms of your distribution channel?

Miguel:
Very important. For us, there are two things behind that statement.

One, private equity has grown, but it’s still a close-knit network. There’s a lot of trust. You talked about being a trusted partner to your clients. We talk about being a trusted adviser. How do they trust us? How do they know that when they come to us, when they pick up the phone, we can deliver and do the right work?

That word of mouth and trust is foundational to them saying, “Okay, I’m going to bring you into this next problem. I’m going to bring you into this next company.”

Associated with that is pattern recognition. I know I’ve done this type of work and I’ve seen these results. Then I see the same issue and I know I can get the same or similar results.

That known quantity or pattern recognition is very important. They see us as a team where, in some cases, things are not going to go perfectly. But we’re not just going to walk away. There’s that implicit relationship that we’re going to make it right. I think they value that too, and we value that too.

It’s more than a transactional relationship. It has to be a longer-term relationship to get their trust and confidence that we will be the right guys to do the job.

Jake:
Trust is delivering what your value is. You talked about that outcome-based model. Talk about what that really looks like.

You have a new customer. You look at 30, 60, 90 days. What could they expect from you in that sort of window? Maybe it’s a longer window, but with those numbers, what does that really look like for them?

Miguel:
The window will depend on the work, but I’ll give you an example of a company in the Bay Area where we did this work. We finished it in March, but we started in January. Those windows, from when you start to when you finish, have also shrunk. That’s one important component.

It was all about improving their deal desk. If you’re familiar with this concept, for some of the larger, especially SaaS or more complex types of businesses, the deal desk provides help to sellers on drafting the contract, putting a quote out, and closing that out.

They had a pretty large team doing that work manually, creating a quote, understanding what the right price is, and where the upsell or cross-sell opportunities were.

With one of the agents we’ve now developed and deployed across a number of companies, we feel pretty confident about what we can automate and the level of that automation.

That automation means time reduction. Instead of Miguel having to spend two hours to go through the details, I can now do a quote that looks and feels the same. In fact, we’re seeing that it’s actually better for both: the context of the client they’re trying to sell to, the context of the price book, the complexity of their product, and how you find something you can potentially sell more of, sell at a higher price, or sell on a different timeline.

We have ranges where we’re feeling pretty confident about the time we can cut out of that process.

The outcome-based approach we have is: let’s look at the problem. We spend a week really understanding what their problem is and where the agents we have can solve that problem.

Then we align on, “We believe this is a 50% reduction.” In some cases, we’re actually seeing 80% and 90% reduction in the time it takes to do some of these tasks.

We agree that we’re going to deploy it. You can start with a subset of the business, or you can start with the entire business. Again, it depends on the size and appetite for change.

But we define what that outcome is, with a clear reduction that is meaningful: 50%, 60%, 70%. Then we say, “When are we going to measure this? Is it two months after implementation? Three months after? One month after?”

For this particular client, we measured at the end of May, about a month after. In their case, the reduction in time was about 60% to 70% for certain tasks that the team was doing, and they were able to reduce the team as a result.

Those savings and efficiencies were captured. That’s another bigger topic, but for you to capture those, there needs to be focus on their side too.

You quoted that Stanford study earlier, and we see that. You can put AI or cloud in everyone’s hands. You can have folks doing interesting and productive things. But is that productivity translating to the top or bottom line of the company? Or is it translating to Miguel’s efficiency in doing certain tasks? Or is it more of a curiosity and internal development that someone in the company is doing?

The key in capturing that and aligning the gains in productivity and what AI can get to the bottom line has to be captured with hard numbers. We are committing to this. We are going to do this by then. We are going to reduce the team, shift the team to a different area, or have higher revenue targets.

That’s what we see as an outcome: a clear metric. In this case, it’s time spent doing a task, and then that translates into what that means for the number of people you need doing that task, or the expectations you have for the volume of sales, cross-sales, or renewals you get out of that team.

Jake:
That’s fascinating. When you go into a company, who are you selling to? Is it the CRO? The CEO? The CXO? Who wants your attention, or whose attention do you want?

Miguel:
It’s a combination of those. For our clients, I would say the majority of the buyer, or the person we’re selling to, is still the CFO.

Think about $200 million or $300 million companies in ARR, private equity-owned. The CFO is critical in that operation. They are also very interested in the efficiency of those processes.

Organizationally, sometimes sales support actually falls under the finance team. In some cases, it falls under the sales team. In that case, we’ll be working directly with the CRO and their teams too.

But the CFO plays a critical role in the ROI and value discussion, so they tend to be our buyer. In some cases, it’s the CEO too. Smaller companies, or companies that have a clear mandate on this change they’re making for the business, often have the CEO personally leading it or involved in that work.

Jake:
I’m curious: when you go into a company like that, they have a problem or an opportunity to make something more efficient and more profitable. Is your model set up where you’re starting with advisory, like a paid advisory service, to understand the lay of the land and then come back with, “Here’s what we see, and here are areas we think we could improve”?

Or is it different and more about, “Look, we understand the space you’re in. Here’s what we know technology can do. Here are the patterns we’ve recognized. Let’s get right to how we can help you and show results now”?

From the outside looking in, every CEO we talk to who is not technical is saying, “I need to figure out how to use AI to generate revenue, reduce cost, improve headcount,” whatever it is. But they’re not technical, and they don’t have the roadmap.

There is time it takes to get in there and understand how they operate first, so you understand what the problem is before you actually start presenting and solving a problem. That takes your time too. What does that blend look like?

Miguel:
It’s a little bit of both. I’m not a fan of the answer, but it depends.

I’m an economist by training, and there’s that joke about wanting a one-handed economist. I’ll give you the two flavors.

Flavor one is what you mentioned: “We just don’t know enough. We don’t know where to start. We are feeling all those pain points. Help us understand where the root cause is coming from, and then we can define what the best solution is.”

It might not be AI, or it might not be AI right away. We do that, Jake, and in a perfect world we would love to do that.

That’s how we started too. This type of work, for us, used to be: “If you want to do it well, this is two months of work. Let’s do a really in-depth look and assessment of your processes and systems. Then, in those two months, let’s design what the perfect world looks like and put a roadmap in front of it.”

We’re now doing it in two or three weeks. We’ve cut that part down.

Independently of whether we do the full assessment or diagnostic, or just a very targeted diagnostic, when we go into the deployment of an agent or technical solution, as part of that timeline, we always spend normally two weeks, but at least a week, getting very specific on what problem we’re solving.

Do you know enough about that problem to say this is the right solution, or is there more to it?

That is changing too. We don’t want to just be a point solution. We don’t want to have the same solution for every problem you have, because that’s not going to be a good long-term option for us.

But the companies we work with are also pretty anxious sometimes. They want to know how to get results faster.

The more consultative part of what we do is: how do we get a clear quick win and value in the short term? Then we can think about the longer term: what changes do you need to make, and what is the timeline to get to those?

A lot of what’s happening now is that the time is getting compressed, or the companies already have a decent understanding of what they need, and then we’re going straight to solving that problem.

If we do it well, what ends up happening is, “Jake did a great job. Now what about these other two things that are holding us back? How can we get after them?”

That focus on value and capturing that is important to us, more than having the perfect picture of everything happening at the company.

Jake:
Every company is talking about agents today, and the foundational models are making it easier to deploy agents. Now, you might not know exactly what you’re doing and how they work, but companies are getting smarter. The talent inside those companies is getting more educated. They’re setting up sandboxes, trying stuff, automating workflows, and really taking that on.

Now you hear Anthropic and OpenAI are building their own service companies and FTE deployment teams that are, in some ways, supporting this growth and, in other ways, competing with other service firms.

What’s your position on that?

Miguel:
I think they’re seeing a couple of things that we’re seeing too.

There is real opportunity here. You mentioned a lot of CEOs or executives not understanding where to start, and more importantly, what the right application for AI really is. Is this flow, or this part, a real application?

I think they’re seeing that too. In a way, they are expanding their TAM or their market by saying, “The more we can cleanly show and tell what the right applications are, the bigger our market will be, the more tokens, or the bigger the bill will be.”

The other reality, and one of our core and big beliefs, is that the services and software sides are coming together. This is already a reality. It has been a reality for many years. Even off-the-shelf software like Salesforce always required some sort of configuration or support to get it to work and get it to work right.

I think the paradigm is shifting now, where maybe you don’t even need us as a layer here. It can all be customized and built for you. But who’s going to do that? Is it all internal? Are we just going to have 10,000 different companies creating their own CRM?

Or is there some equilibrium here on what is already developed, robust, and easily available, that you can pick from and then build on top?

I think they’re seeing those two dynamics: how do you seed and grow the market by showing specific use cases and applications with tangible value, but also how, for this wave to continue and for adoption to continue, you’re going to need some sort of forward-deployed engineers or support in getting that to the broader market.

With private equity, you’ll see the partnerships or investments some of the big labs are making with private equity firms. I think they’re seeing that, especially at the smaller company size.

We are right on it. We see it every day, and we believe it will be direct competition for us. But it’s also showing that there is big potential here.

Jake:
It seems like there are going to be a lot of acquisitions of talent as well, and teams that have large organizations doing the consulting.

If the product is out there in AI and the adoptability or usage of it isn’t at the level it needs to be, there have to be people on-site implementing it and making sure that it is working and that the function it’s supposed to perform is actually happening.

I think you’re in a great spot.

Let’s shift gears a little bit here to people. How big are you today, and what’s the model that’s worked for you? You talked about international teams and different parts of the world where your teams are located. What’s been the biggest opportunity for you there, and also what’s been the biggest challenge?

Miguel:
We’re about 160 to 170 people globally now. We operate as a global company and have presence in a number of geographies.

The majority of our people are actually in the United States and distributed. We are a virtual-first company. We’re starting to have interesting clusters of talent and people, but we’re still virtual-first.

We have a big presence in Poland. That’s how we got started, more on the technical side. We do have a physical office there, but it’s primarily a virtual culture.

The challenge is that, like probably every CEO tells you, I believe I have the best team. We do have really good talent, Jake.

The problem we’re trying to solve and some of the tools now available are making us able to get the right people and the right profile of talent aligned with our values.

But the biggest challenge continues to be: how do you identify that talent in a competitive market?

The AI or full-stack engineer focused on AI continues to be an interesting and hot market. People are reinventing themselves, which I think is the positive side of some of the changes happening, but it continues to be a challenge.

Out of those 160 people, we have a recruiting team that is healthy for our size because, if we don’t have the right talent, we’re not going to be able to deliver.

Continuing to get the right talent and dealing with the fact that we are all remote is beautiful because it gives us more opportunities to get that right talent, but it’s also harder to build the culture and make sure we are aligned and supporting each other when challenges come.

I think that continues to be the challenge: how do we continue to grow in this virtual model? How do we bring in the right talent? And how is that talent bar changing so quickly?

The folks I was hiring two years ago, who were doing a fantastic job, if they have not reinvented themselves in the last two years with what has happened with AI, it becomes very hard for them to play at the same level as some of the other technical talent we’re bringing in that we’re calling AI natives.

One of the core values we have is being AI-native. You don’t have to know AI better than anyone, but you have to be willing to learn it, explore it, and drive it with our clients and with the products that we have.

I think you’re probably seeing the same thing. Good talent is still hard to get, but then it’s also: how do you raise the bar for the internal talent? How do you continue to upskill the talent you have and keep doing that, because things are changing really fast?

Jake:
How do you upskill your talent internally?

You have engineers you hired years ago, and you have new talent coming in. What’s the way to help transfer knowledge or improve where people are at? Everyone may be at different levels.

Is there formal training that you set up internally? Is it an innovation day where everyone gets together and talks about what they’ve learned? What’s worked for you?

Miguel:
Exploring, but there is still formal training. I think folks still appreciate two days of training, but there are sessions we maintain.

We’ve actually been hearing that more and more from our talent, Jake: some guidance is very helpful. It’s great that we’re using this tool, or we’re using Claude or OpenAI, but how are we using it? What are we using it for?

That is basic. But more importantly for us, now we have our own agents. How are we going to continue to evolve them if I was not the one who built them? If we’re not interacting with code the same way?

So there is some formal training and more experience-sharing sessions. We have very active teams and Slack channels on what’s working: “I tried this. Look at this.”

The other thing, more on the leadership side, is that the leaders I have across the different groups have it in their OKRs to raise that internal talent bar by upskilling and making sure there is clarity around the expectations for those teams.

Holding the leaders accountable makes us do more of the trainings and more of the experience sharing.

In some cases, the other piece that I believe is working pretty well is the pairing or pods we’re putting together, specifically on the agents. That cross-pollination and cross-training: “I’m working with Jake on this specific agent or this specific client.”

There is still that apprenticeship and mentorship piece, and what we hear is that it’s still working pretty well for discovering what is possible or how you do some of the work with the new tools you have available.

Jake:
That’s great.

There’s this term that has been floating around recently called AI hoarding. Essentially, what it is is people are learning AI within an organization, but they’re not sharing it because it’s helping them elevate themselves and, more importantly, have more value for the organization. But they’re not necessarily sharing it because maybe they feel like they’re giving some value away that, from a competitive perspective, they might want to hold on to.

Miguel:
Interesting.

Jake:
In our company as well, I’ve seen it. We’re all learning AI ourselves. Some are very forward-thinking and curious. They’re learning quickly and being able to deploy their own workflows and agents. It’s great, but it’s not really shared until we had a conversation, put it on the table, and said, “This is what we’re seeing. How can we get better as a company?”

We’re all trying to win, so let’s have the conversation. It’s across the board. We’re hearing this more and more.

I’m sharing it with the listeners because it might be coming up. The way we solved it was that we had an open-floor discussion and said, “Here’s what we’ve seen companies do that’s worked, and for companies that haven’t done this, it hasn’t worked.”

It’s about communication and education. Then we put a formal plan in place to start talking about our breakthroughs. What are we learning? How are we doing it? Then having people show value and get accolades out of it.

It’s been a journey and a process, but at the same time, there’s so much opportunity. I’m really excited to see where things go in AI in particular.

You mentioned the people you have, but also your recruitment team. There’s a strategy for a lot of companies: do they build their own recruitment teams? Do they use their own networks? Do they use outside services? What direction do they go?

Talk a little bit about your recruitment team. How big is it? Is everyone a recruiter? Do they just go into their network? How does it work for you?

Miguel:
A little bit of everything. We actually reward referrals too, and we believe good talent brings good talent.

There has been some good growth out of references and referrals from current team members.

We have four people whose job is to do sourcing and recruiting. Like I said, it’s a pretty healthy team for a company of our size.

To grow at the speed we’re growing, we needed continuity and someone internally who understood our values and the profile we’re looking for.

I think we’ve nailed it in a couple of areas for the profiles we hire the most, but we still rely on partners too when we need specific talent. You talked about AI, or some architects within that space that are harder to get, where you really need the right help.

Even senior team members are still a hard part to recruit internally because you do probably one or two of those a year. You’re not going to hire a head of sales every month — hopefully, otherwise something is really wrong — or a COO or one of your core practice leads.

For those, we still believe that having the right partner, who shares that value and trust with us, is important.

Jake:
I always like to leave a little space at the end of the show to talk about the roles that you have in front of you that you’re looking to attract or hire for. Candidates might hear this, see it, and connect with you directly.

It’s really just a free way to say, “Talk about what your growth looks like over the next 10 to 12 or 18 months, what that looks like, and where people can find you if they want to apply or work with you.”

Miguel:
Thanks for that.

We’re still growing a lot, and we’re very positive on the outlook of the market we’re working in.

There are roles that are almost going to be evergreen for us. Think about this type of talent we describe as a forward-deployed engineer. That’s the common term, but in our case, it’s someone who actually understands the functional aspect of what a go-to-market business looks like and then has the interest, or has already developed a good AI mindset, when it comes to working with our own agents or adapting and deploying our own agents.

To me, that’s the unicorn we’ll continue to hire. You understand what a good go-to-market flow is. You have expertise in that, or you’ve done it for a year or two. Now, how do you bring that to other companies with the technology that we have?

Think about folks who were either consultants before, or practitioners of RevOps, but have a strong bend into the AI or technology side.

The key and core AI engineers and architects will continue to be critical for us as we grow the team and grow the agents we have deployed.

The other one we discussed even previously is on our own go-to-market, because this is such a new market. Are you selling services? Are you selling product? It’s neither, and it’s both.

So folks who are interested, either as a BDR or a true salesperson, in understanding how to sell in this new market and this new paradigm of services and product combined — those are roles I still see us hiring for.

We mostly post them on LinkedIn. There are always a good number of those roles posted on LinkedIn, but feel free to reach out directly to me or Rebecca Allman, our COO, who oversees the recruitment team, on LinkedIn. I’m happy to pass those contacts along.

Jake:
Very cool.

Miguel, I’m really happy to have you back on in a conversation for the podcast and the show. I think this was very educational. You’re in such a great space. I wouldn’t be surprised if you’re getting offers from OpenAI or Anthropic at some point to say, “We’re going to give you 10x for your company, but we need all your people with us.”

We’re seeing some pretty cool stuff happening out there.

If anyone wants to find Miguel or TechTorch, what’s the website?

Miguel:
TechTorch.io is where you can find us.

For me, LinkedIn. I don’t have a very common name, so it’s kind of easy to find me. There are a few of us with the same name, but if you search Miguel Vasconcelos and TechTorch, you’ll definitely find me on LinkedIn.

Jake:
Very cool, Miguel. Thanks for coming on.

For the listeners, thanks for listening. It means a lot to me that you spent your time with us today.

I’m your host, Jake Aaron Villarreal, signing off for now, but I can’t wait to catch up with you all on the next episode. Until then, go change the world. Take care.

If you like what we’re doing, don’t forget to subscribe, leave a review on Apple Podcasts or wherever you listen, and follow us on YouTube, where we go behind the scenes to learn what it takes to be a startup founder.

‍