Alguna's co-founder and CEO, Aleks Đekić, sat down with host Alex Theuma for an episode of The AI Revolution Show.
The conversation covers why the top 500 AI B2B SaaS companies are now changing their pricing every 90 days, why seat-based pricing is being squeezed by AI agents that do the work a seat used to do, how Alguna landed its first 10 customers by building far more of the stack than it ever intended to, and why "no one owns pricing" is one of the biggest structural problems in B2B software today.
Listen or watch the full episode
- Spotify: The AI Revolution Show
- YouTube: @The-AIRevolution-Show
- Apple Podcasts: The AI Revolution Show
Look for episode 12: "90-Day Cycles, the Sales-Finance Chasm, and Why Your Billing Stack Is Holding You Back" (16 July 2026) with Aleks Đekić.
Key takeaways
- Pricing has become a moving target. The top 500 AI B2B SaaS companies now change pricing an average of 3.6 times a year — roughly every 90 days — according to Kyle Poyar's Growth Unhinged research, a pace Alguna's own data independently supports.
- Billing infrastructure is the real bottleneck, not the idea. A first version of a pricing model can be built in a weekend, but what actually takes months is the underlying foundation that lets sales, finance, and billing stay in sync as pricing gets more flexible and complex.
- Nobody actually owns pricing. Even when a company has a "pricing manager," the real decisions are cross-functional. Finance, engineering, product, sales, and revenue ops all have a say, which is exactly why pricing changes move so slowly.
- Seat-based pricing isn't dying, but it is under pressure. As AI agents take over work a human "seat" used to do, pure per-seat pricing inverts the incentive: succeed too well, and revenue shrinks. Most companies are heading toward hybrid models, a seat or usage allowance up front, with usage-based pricing layered on top.
- Usage-based pricing needs guardrails to work. Alerts before a customer hits their limit, the option to buy more at a discount, and surge protection against runaway bills are what make usage-based pricing feel fair rather than terrifying to a buyer.
- Outcome-based pricing is coming, but not yet. It will become viable once AI agents are reliable enough to make outcomes objectively measurable. Once boards and investors are willing to value usage-based revenue the way they value ARR today.
- Alguna's first 10 customers came from solving one real, painful problem: the "chasm" between what sales sells and what finance can actually bill and from being willing to build far outside their original scope (including a full invoicing and collections platform) because a customer needed it.
- The founder's biggest challenge right now is prioritization — deciding where to invest, resisting "shiny object" distractions, and keeping product and support quality high while scaling a deliberately broad product.
Transcript
Host: Alex, good to see you. Thanks for coming on the podcast today.
Aleks: Thank you, Alex.
Host: So, Alguna — in a sentence, for those who don't know, what is it?
Aleks: Alguna helps B2B SaaS companies with pricing, quoting, and billing across their whole customer lifecycle — from the first quote they send to a customer all the way to the last invoice. Hopefully never the last invoice, but if they don't pay, we'll help you chase it down.
Host: What led you to build Alguna? What were you doing before that influenced this — did you wake up one night and think, "I want to fix this problem"?
Aleks: I wish it was just one night. It was months and years of painfully experiencing the same problem over and over, and it's pretty much true of every company I've been part of. Particularly at Alguna — this goes back to a company called Dojo, where I saw how this can work if it's properly set up from the get-go: your sales team can sell very flexibly and creatively without impacting downstream billing and finance, while finance still has full control over what the sales team does.
And then moving into my previous company, Primer, a payment orchestration platform, we tried to launch new pricing models, new features, PLG, similar things — and over time we learned that to do that properly, we needed to integrate five or six different tools. Based on my experience actually seeing this play out, one day it clicked, and I started researching the space and talking to my now co-founder, Jamie, who I'd met at that company. We decided to try it.
Host: And try it you did. I guess if anyone else wants to try it — the barrier to entry for building a product with AI now is much lower, much easier. I imagine this isn't something you could do on a Sunday morning and get to a first version of the product. How long did it take you to get to a version ready for that first customer?
Aleks: If it's in isolation, for one specific use case, it is possible to do that in a Sunday morning. But the problem comes with scale and flexibility over time. As soon as you add another person — your CEO, your VP of Sales — and they want additional discounts, or credits, or something else, suddenly that can't be done overnight, because of how long it takes to get the foundation in place so that, on a technical level, whatever sales quotes, billing can actually bill.
It takes time because of how B2B has evolved around discount terms, credits, products, usage-based and recurring and hybrid subscriptions — whatever you can imagine to get the deal over the line. But billing usually lags, because the assumption is, "Oh, I can always write that into a spreadsheet," or "I can always add a line item in an accounting system and just send it as an invoice." That's where the moat starts to form — foundations take a lot longer than any VC or founder would like. But once the right foundation is in place, building additional products on top of it moves fast — we've built six or seven modules over just the last six months, which comes in very handy with AI. Must be —
Host: — quite an exciting time for you specifically around pricing and monetization. If we think about the last 20 years of SaaS, it's mostly been per-seat pricing — how exciting can that really get?
Aleks: Maybe it's exciting for some. Often it was just one model that didn't really change. Yes, there was usage-based pricing, and we talked about some of the examples — Slack, Snowflake, Databricks — doing very well with it, but it never got widely adopted. Right now, though, everyone's going through the AI platform shift, and the question is: is seat-based pricing still the model?
Host: Usage-based, token-based, outcome-based, whatever — there are so many questions founders and companies have about what needs to be done. So what are you seeing in the market right now? What are people coming to you for, and how are you helping?
Aleks: It's the most exciting time for us geeks in the front seat of what's going on. For some companies it's become the norm to think about this iteratively, for different reasons — their margins, or how they operate. But I think a lot of companies will inevitably fall into that, and we're seeing that traction.
When we started in 2023, people would say, "Why are you solving this? The problem's already solved. Why are you even building this?" There were five different CPQ tools that came out over two months. But now people are starting to realize this is a real need. One research report found that the top 500 AI B2B SaaS companies have changed pricing on average 3.6 times a year — roughly every 90 to 100 days. It's becoming a mass infrastructure investment across every company.
Host: So that's a data point you have — that the top 500 AI companies are changing pricing roughly every 90 days?
Aleks: Yeah, that's from research by Growth Unhinged, from Kyle Poyar — that's the average they came up with. Our own thesis was "changing every 90 days" from our own research, and it lines up with those numbers when we talk to experts in the industry.
The problem is those numbers are skewed by companies like Anthropic, which have unlimited resources to work on billing and invent new ways of doing things — including [name unclear] and OpenAI — while, in that same top 500, there are still companies that haven't changed their pricing at all. So the average is skewed toward one specific group. But the companies that don't start thinking about this, or figuring out how to tackle it — where everything's still fixed while they add AI products as a baseline and their margins drop from 80% to 50% — are going to be in trouble soon, when they look at their financials and realize something needs to change.
Host: Yeah, definitely — at best, maybe some leading companies change their pricing annually. Many wouldn't even do that, from when they first built the product all the way through the whole decade. It's an interesting data point — obviously these are the top AI companies, but we think this might become the norm: this kind of cadence, looking at pricing every 90 days. I guess a lot of companies don't have a specific person for this, but if you were going to do it, having somebody who owns pricing within a business — I haven't really seen too many companies with that specific role. Usually it's the CEO or CFO who owns pricing, right?
Aleks: Yeah — historically the final say came after six months of hiring a pricing consulting team, who'd come in, look at your usage, look at the benchmarks, do all the research, and figure out what people are willing to pay and how — which works really well at scale in B2C. In B2B, pricing experimentation comes down to whether the customer will get over the line if you offer this particular price. So it's much more art than science.
Generally speaking, though, we've seen that pricing isn't really owned — on paper it's one person, but even when that person's title is "pricing manager" or something similar, it's usually driven by cross-functional collaboration between finance, engineering (especially where infrastructure cost is involved), product, sales leadership, and revenue ops. There isn't really one person who can make that call internally, though the final decision is made by whoever has the authority to make it. That's one of the challenges — no one owns it. Everybody wants input, and the customer on the other end usually suffers, because they're either really benefiting, or, if they're a heavy user, they end up overpaying compared to someone else on the platform.
Host: It's probably a good time for pricing consultancies right now too, right?
Aleks: Yeah, and I think they still need to be involved in the process, because they can save you a lot of time and trouble with that experience. The problem is implementing the decisions. Based on our research and the companies we've talked to, we found public companies that, over 15 years of existing, went through eight pricing changes. Eight in 15 years is just under one every two years. If you're now doing three a year, something needs to change internally — you shouldn't be in a situation where you need to move fast in the market but have to wait six months to actually roll out a new pricing scheme. That's a common thread across a lot of these companies, even ones with 50 engineers on the billing team, which honestly is crazy to me. It comes down to knowing what you want to change, and then having the right infrastructure in place to do it.
Host: I can already see the YouTube thumbnail: "You've got to change your pricing every 90 days." But it's pretty much the case. So — thinking about the seat-based model and the impact AI is having on it: because agents doing the work sort of inverts the per-seat model. If you succeed, revenue shrinks. Is that the crux of it? Is pricing more art than science, or as much science as art?
Aleks: It's a two-way street between the buyer and the seller. That's probably one of the biggest issues — companies want predictable revenue, even if it costs them on margin, if those users start over-using the product. We see this even in our own customer base, where we're still on seat-based pricing: as we add usage components, 80% of customers aren't using as much as they've paid for, but that top 20% is really using up the whole allocation. That's usually the trigger to start changing.
But you can't jump straight into usage-based pricing, because there's a drop in reported revenue that you'd have to explain to your board — just to keep showing growth. So you have to find a middle ground: a hybrid model where you buy a certain amount of usage upfront, and then move into full usage-based pricing once that's used up. Because it's a two-way street, the customer wants certain protections in place too — they want to avoid price surges. So having the right levers in place is something we'd recommend to companies.
Host: Usage-based pricing should track the value delivered, right? But sometimes it can terrify buyers, because it's unpredictable, certainly from a billing perspective. So how do you align price to value while still giving customers something they can budget for?
Aleks: One thing is setting up an allowance layer — either per seat, or as a company-wide shared pool of usage. The second part is alerts and notifications: "Hey, you're about to hit the next tier." That's something that hasn't really been done much in SaaS generally. The next piece is bringing in your customer success team to act almost like a salesperson: "Hey, you're about to hit 80% of your usage — here are your options. You can buy another bulk at a 50% discount." That's great for us as the seller, because we get predictable revenue and margin, and it's great for you too, because you're not going to get hit with a surprise bill, and we don't have to deal with your CFO complaining about a new invoice.
So that's one lever — potentially a really strong one — but it requires infrastructure. A second lever is surge protection — [product name unclear] has that, for example, and it's worked really well for a lot of early-stage startups: if your usage skyrockets and you really go through the roof in month one, month two, or three months in a row, they'll cover that for you, because it means you're benefiting from the product, you're growing, and eventually they'll benefit too. Sometimes you need to give something to get something.
And then there are all the familiar levers: tiered pricing, discounts, minimum spends, maximum spends, specific credits you give customers as they go through the process — almost like milestones — and ramps and phases, so the price changes over time. You want your end customer to see you as a value provider, not just a cost center — someone they're happy to pay because it actually makes sense for them.
Host: What's the most wrong thing you hear people say about AI or pricing in the AI era?
Aleks: It's always a bit more case-by-case. From a trend perspective, the idea that everything will switch from usage-based to outcome-based is something we internally didn't believe in. More and more companies are realizing that may not work beyond certain use cases — customer success tools, like Zendesk and Intercom, and Sierra, are doing really well with it, because it's very easy to tell whether it worked or not. You and the customer agree on the outcome, and it either happened or it didn't.
For everyone else in the industry — moving from a seat-based product to a usage-based one — it's going to be harder to tell whether the outcome is right or wrong in a particular case. So what we've seen is there's not going to be an overnight flip. Companies have to report and grow based on recurring revenue, and a lot of boards, VCs, and private equity firms don't believe in usage-meter reporting. We have customers with consistent usage on these products that we still can't report as ARR, even though it's pretty much consistent month over month — just because of how the world operates right now, based on ARR multipliers. Not many companies have been able to break out of that, except, as mentioned, Snowflake and a few others.
On the outcome side, it will change — and this is our internal theory too — once it becomes very objective, and that objectivity will come once LLMs are good enough to do truly autonomous, agentic work. It will happen. It's 2026 as we're recording this, and I'm sure you're reading the same LinkedIn posts: "outcome-based pricing is the future."
Host: Do you ever see a future where, like in SaaS — which was pretty much all seat-based with a little usage-based mixed in — everything eventually becomes outcome-based? Or is it more likely we'll see a mix of all these different pricing models?
Aleks: It's a bit of a balance, I'd say — especially since my algorithm strangely keeps showing me pricing, billing, and monetization content. I think it's inevitable that we'll end up there in one way or another, especially with autonomous, agentic workflows, once things become very objective. But given how far behind SaaS infrastructure is — and by "infrastructure" I mean multiple tools taped together for every single step in the process — it's going to be very hard for these companies to go back and forth between tools and change their pricing the right way. So it's either a very expensive rip-and-replace that takes a year, or a slow transition as these single-point solutions start adding more features so they can support the new process.
Host: Shifting gears a bit — building a fairly complex solution in an AI-native world, where there are incumbents you're looking to replace — how did you go about getting your first 10 customers? And maybe share a bit about your GTM motion, and how you're growing and scaling the business.
Aleks: From day one, we tried to approach this from first principles. We got into this out of frustration with so many companies doing things in a pretty questionable way — not because they wanted to, but because as they scaled, things just kind of designed themselves. We have CFOs and CROs with the best degrees in the world spending Friday and Saturday nights doing spreadsheet reconciliation between an invoicing platform, an accounting platform, and a payment platform. We know this because our usage logs spike at times of day you wouldn't expect people to be working.
Using that first-principles approach — making sure the data flows from the start of the process to the end as a single source of truth — let us approach companies with a specific problem: not being able to connect their sales and finance teams. The biggest challenge our product solves is connecting the truth the sales team believes in — the contract they signed — with what finance actually invoices. There's a real chasm between the two, and we're trying to bridge it. Because they're using five or six different tools, there are a lot of specific pain points along that journey.
Our first 10 customers came really organically — it's not that I had a big network; I'm a first-time founder, and so is my co-founder. So for us it was about approaching real problems and working directly with customers who felt the pain and trusted that the product would get to a point where they could take a leap of faith, and we'd be there for them when things went wrong. That's pretty much how all our first 10 customers came about: "Can I trust this team to deliver on the promise?" — especially since nothing else had worked out for them already. And once you have that trust with around 10 customers, you get a 360-degree view of what's going on in the industry.
We didn't do the thing we probably should have — the lean-startup approach of doing one thing really well and growing from there. We spread our product way too thin from the get-go, because our thesis was that the only way to solve this properly is to solve it as horizontally as possible from the start. That's how we started building use cases with customers as design partnerships — we literally built products with forward-deployed engineers, building a dedicated product for that customer. That's now covered about 80% of the use cases we see in the pipeline: "We already have this, because we built it for somebody else."
From a GTM side, the approach is similar — reaching out, finding these kinds of problems. But because of the complexity of the sales cycle, no one wakes up in the morning and says, "I'd like to change my billing." It's usually, "Do we really have to do this now, or can we get away with some alternative?" So we're building a reputation in the community — people who work with us can say, "They were there for us when we needed them," especially when something went wrong, like an error. We're there for that immediately, because no one reaches out to a company to say nice things unless there's an issue to fix. So our GTM motion is built around building that reputation, building trust, and following through. Over the last few quarters we've seen significant inbound from LLMs too, because we invested in that early — just to make sure we're spread across multiple channels, so that when customers decide it's time to replace their current setup, we're ready.
Host: What did your first customers need that you hadn't anticipated?
Aleks: We'd always said — my co-founder and I both come from payments — that we would never touch payments, and we'd never touch invoicing, dunning, retries, any of that. We weren't going to go there. But the first customer needed it. The existing solution — a really well-known brand, at least from what we'd seen — couldn't do it the way they needed. So instead of just building a billing engine with usage metering and line items that you send to your provider to collect payment, we ended up building an invoicing platform too, which then escalated into processing payments through a third party, then dunning, then a collections process. Suddenly 30% of the product was something we never wanted to build, because there wasn't a better solution and it didn't make sense to do anything else.
Host: So you built that for the first customer — has it paid off for the other customers too?
Aleks: Yeah, it became one of our biggest competitive advantages, because you get a real end-to-end experience across the board, and it helped us win over finance teams. A lot of those finance teams had struggled with reconciliation between payments products, invoicing, usage metering, and revenue recognition. Because we cover that whole flow, they're suddenly able to click around a few times and close their books within minutes.
Host: Is your ICP the same now as when you first started selling? Have you refined it? And what's the hardest objection you get from the people you're selling to?
Aleks: Our ICP, we knew from the start — we wanted to go after the most complex use cases in the industry, and we built with that in mind from the get-go. Our ideal use case is: a founder signs up, starts a company tomorrow, implements us, and never has to worry about scaling this part of their product again, because that's how we built it. But obviously, as you go through the process, you learn over time.
The hardest objection has always been — when you're a young company that has to define how revenue works, and needs to be reliable at scale — trust. Building trust from the get-go has always been the number one objection, followed by, "Are you going to be around?" — because you're about to invest a lot into our product and our time, and we want to make sure you'll still be here. On the trust side, we have a very clear way of operating during the GTM process, so customers can see what to expect, even though, arguably, we're better once they go live than during the sales process itself.
Host: We need more companies like yours, which will then help us get to the next stage. Can you give some examples of building trust in the GTM process?
Aleks: We try to involve technical teams early on — our product engineers are called "product engineers" because they build the product, but they also support customers on their journey. We bring in their own data as early as the second call, to show them, contextually, how it would work. Most importantly, after that, we ask them to please push the limits of our platform with their own use cases — because we don't want it to be a nice sandbox environment where we solve all the tidy problems, show the finance team how great it is, and then three months into implementation they realize, "You guys don't actually support any of this — why did we sign up for this?" and it falls apart.
We can't afford customers like that, because it's very expensive for us to be involved in the process. So what we try to do is push them to push us, because we know the scrutiny on us is much higher as a startup, compared to companies that have been around for 10 years. We really try to create that trust environment — push what we're capable of, see if we can solve all their cases or edge cases, and then work together on finding that last 10% that's missing, so we can collaborate to get it over the line.
Host: Thinking about yourself here — first-time founder, as you mentioned — what's genuinely the most difficult thing right now? What's keeping you up at night?
Aleks: Prioritization is number one — knowing where resources need to go, where you need to invest more, where you need to invest less, and how you make sure you can keep the quality of support and product high with all of that going on at once.
Host: How do you make decisions around prioritization?
Aleks: Very much pipeline-driven, because of the complexity of the sales cycle — we ended up skipping a few steps and landed on much more complex sales cycles, which determine what we're able to do or not, and who we can work with or not. And there are shiny objects, as a founder — "Oh, it'd be nice to build this feature," or "it'd be nice to land this customer for whatever reason" — even when deep down you know it's not the right thing to do. Shiny objects are always a challenge. So really it's about being disciplined and staying focused: "This is the goal, we're going after that," and putting everything else that's less important to the side. Unfortunately, there are only 24 hours in a day, and you have to sleep some of it.
Host: You do, you do. How much sleep do you get?
Aleks: Next question. [laughs]
Host: Okay, playing it forward — does per-seat pricing survive at all, say by 2028, 2029?
Aleks: I think some shape or form of it will stay forever, because predictability will always be needed. And — this is going to sound controversial, so let's see how this goes — investors who've backed companies like Snowflake or Databricks understand the power of usage-based pricing when it's done properly at scale: land small, then expand, and grow with your customer, so a $50K deal can become a $2 million deal. That's not common for the majority of cases, though, and a lot of companies — especially founders at Seed, Series A, or Series B — depend on how quickly they can show ARR. Over the last 12 to 24 months, we've heard some very creative ways of reporting annualized ARR, or annualized run rate, because there's a benchmark in the industry you're expected to hit.
Companies that have really scaled from a usage base need a team that understands this is an upfront investment that pays off over time. And usually, as a first-time or second-time founder, you don't have the flexibility to go to your board and say, "Can I raise $30 million now — trust me, in three years we'll be doing $5 million rounds." That takes a lot of leap of faith, especially now that everyone's going from zero to a hundred million in ARR overnight. So the thesis is: it'll stick around, but it won't look exactly like it does right now.
Host: Which companies are structurally advantaged by this AI shift, and which are most exposed?
Aleks: The most exposed are the ones with multiple tools stitched together that often don't work well together — CPQ, document design, invoicing, billing, payments, collections. The ones that are well set up are the ones that have had some usage-based component from the get-go, as we've mentioned a few times now — if there's always been a component that required engineering to be involved in pricing, you're automatically ahead of the game, because there was already something flexible built in.
Companies on pure seat-based pricing that were just using one type of tool that only supports that model — or, surprisingly, a lot of them are still doing it through spreadsheets, even with 200, 300, 400 customers on the seat-based side — are going to have a hard time. Although, I'd argue, if you're still on a spreadsheet, it's actually a great time to pick the right infrastructure and go for it.
Host: Final question, Alex — what's the thing about Alguna that you never get asked, but you've always wanted to talk about?
Aleks: I'd say it's that every company now has a new product line they're not aware of, which is pricing. If they don't approach it the way they'd approach building the rest of their product — with a dedicated person, dedicated engineering team, dedicated stakeholders who care about it — and instead treat it the way some older companies have... Think about how mobile apps had the "growth manager" role, where someone's entire job was to find ways to drive adoption — Facebook was famous for that. A lot of companies moving to usage-based pricing are going to need that. They'll need to focus on: if we launch this, how much will it cost us, what will the margin be, how quickly can we move, and how do we iterate on that.
Anthropic is a great example here — they have weekly limits, hourly limits, five-hour limits, different limits per model. All of that is going to become a necessity in some form — maybe not at that scale, but in some form. That's the effort that's needed, so it becomes proper product management instead of a one-off, 12-month project. And by the time things change, it's already too late.
Host: Well, Alex, thanks so much for coming on the podcast. It's been great to hear you speak and share your knowledge on monetization and pricing, which is probably — if not the — one of the most important topics in software right now. Thank you so much for coming on the show, and for paying it forward.
