If you're a business owner, an agency, or a consulting firm, there's a change you can't afford to ignore, even if at first it sounds counterintuitive: AI is enabling the shift from the software as a service model (SaaS) to the service as software model (SaS). It's being talked about more and more (see Sequoia), it concerns both software providers and, above all, professional services. It can sound like a buzzword. It isn't.
From legacy to cloud
To understand it properly, though, you need to start from the beginning, from the SaaS revolution. About twenty years ago companies had a problem with the software they used: it was monolithic, fixed once installed, support was slow and expensive, and rather than the software adapting to the business it was the other way around, with entire departments depending on these magic black boxes called legacy systems.
Side note: plenty of people built entire careers simply by being among the few who held the know-how of the legacy system.
With the rise of the cloud and the arrival of software providers (think Salesforce or Marketo), rigidity became flexibility. Now the software was hosted in the cloud, it could be configured before use, it was updated regularly. At this point we had:
Companies leveraging SaaS to deliver their own expertise: for example CRM strategy and campaign execution through Salesforce or HubSpot, analytics and reporting on tools like Tableau or Looker.
Software providers leveraging economies of scale, selling the same software to many clients with the promise of customization.
Now it's worth being critical of this model, because it wasn't all smooth sailing. In fact these tools, rather than becoming an instrument through which to exercise expertise, often became the expertise itself, in two senses.
Many of these tools, to be used at full capacity, required custom code and proprietary languages. So whoever used them had to be a market-recognized expert (think of the whole string of certifications). Proficiency with the tool became the real company asset. The extreme case is AMPscript, Salesforce Marketing Cloud's proprietary scripting language for dynamic personalization inside emails and landing pages. To really push the tool, you needed someone who could code in a language that only lives inside that ecosystem. You weren't buying a capability, you were buying dependence on whoever could activate that capability.
The promise of customization was, in reality, configuration inside a software block that was already fairly fixed. And in fact, rather than the software adapting to the company, the opposite was more often true, with the business regularly hearing "that can't be done" or "that's too complex."
The promise of agility had, in effect, turned into a costly and one-note uniformity.
Service as software: meaning what?
Okay, now that we've laid the groundwork, it's time to define SaS and the role AI plays in it.
Right now, when producing code has become a commodity and you can build an agentic system to orchestrate your service, you can genuinely think in terms of infinite, real time customization, matching the need for adaptability. In this sense you no longer have to worry about knowing how to use the software or about its limitations to deliver consulting to a client. You only have to worry about your agentic software being the truest possible representation of how you work.
In effect, your service is no longer delivered by a piece of software: your service is the software itself. Imagine you're an agency specialized in SEO. Now you can build your own system based on your workflows, your data, and above all your way of generating insight and building strategy, and resell it to clients. You're not pitching a piece of software, you're pitching your agency to as many clients as possible, with unlimited room to scale while keeping flexibility and adaptability.
What it actually means to sell service as software
Let's go one step further, because "selling your service as software" can still sound abstract.
In the SaaS model you bought the machine and then hired (or became) the expert who knew how to drive it. The value sat in your ability to bend a generic tool to a client's needs. In the SaS model the value goes back to where it always belonged: your way of working. Except now that way of working no longer lives in the heads of three key people or in a folder of procedures nobody reads. It's codified, executable, and, crucially, sellable.
In practice you're doing three things at once.
First, you're extracting your methodology. The way you run an SEO audit, how you prioritize keyword clusters, how you decide when a piece of content needs a rewrite instead of an optimization pass: all of this stops being tacit knowledge and becomes a workflow orchestrated by agents.
Second, you're packaging your judgment, not just your actions. This is the subtle but decisive difference from classic automation. You're not automating tasks (Zapier already did that), you're codifying how you make decisions. The agent doesn't run through a checklist, it reasons inside your criteria.
Third, you're turning a service that scales linearly (more clients equals more hours equals more people) into an asset that scales like a product. The agency that today struggles to manage 15 clients can serve 150 without multiplying headcount tenfold.
And here's the uncomfortable part worth saying out loud: if you don't do it, the risk is that your own method becomes someone else's product. The commoditization of code cuts both ways.
The benefits (and who they land on)
The first benefit is obvious but worth repeating: margins. You move from a business with high marginal cost (every extra client costs more billable hours) to one with low marginal cost. Delivery stops eating your growth.
The second is defensibility. Anyone can buy a SaaS, which is why your competitive edge there was fragile, resting on certifications handed out by the thousands. Your agentic system, built on your data and your workflows, is much harder to copy, because it replicates know-how that's yours, not the vendor's.
The third is speed of adaptation. In the old world, when a client asked for something outside the software's fence, the answer was "can't be done" or "that needs a development cycle." In SaS, customization happens in real time, because you're not configuring a predefined block, you're recomposing capability at the moment of execution.
The fourth, and maybe the most underrated: you stop being hostage to someone else's roadmap. How many times have you waited for a vendor to ship the feature you actually needed? When the service is your own software, the roadmap is yours.
What changes for each role
This shift isn't just technological. It redraws who decides what inside a company. And whoever doesn't get that now will spend the next few years catching up.
CIOs need to rethink architecture not as a stack of applications but as a capability mesh orchestrated through prompts. It's no longer about which software to buy, it's about composing capability.
COOs have, for the first time, the chance to design operations that are responsive, adaptive, and context sensitive, instead of squeezing processes into whatever space the software allowed.
CFOs need to prepare for AI native costing models, where price aligns to value and to business outcomes rather than to the number of licenses or seats.
Risk and compliance leaders need to invest in real time explainability and guardrails, because now they're governing workflows that may never repeat the same way twice. It's a new kind of challenge: governance no longer looks at a fixed process, it looks at a process that recomposes itself.
And finally business leaders: they finally get to move past "best practice" and define what's right for them, at every moment of execution. Right practice, not best practice. Because best practice was, in the end, a compromise you accepted when the tool wouldn't let you do any better.
SaaS vs SaS: the differences that matter
We're already in it
This isn't a scenario to keep an eye on for the coming years. It's what we're already building at Lorenz, right now, together with agencies and clients who understood before everyone else where the market is headed.
We take your way of working, your workflows, your data, and your way of generating insight, and we turn it into the agentic system that becomes your service. We're not selling you one more piece of software to learn. We're helping you turn your agency into the product.
If you're also ready to stop scaling one billable hour at a time and start scaling like a product, reach out. The right time to build it is before your competitor does.
