For most of the SaaS era, software companies have sold tools.
You buy the CRM, then your sales team operates it. You buy the marketing platform, then your marketers build campaigns inside it. You buy the support software, then your support team uses it to answer customers.
The software helps humans do the work.
AI is beginning to challenge that relationship.
At SaaSiest 2026, Nathan Latka, CEO of Founderpath, shared nine examples of software companies using AI to increase revenue and improve profitability. But underneath the individual tactics was a much bigger shift.
The fastest-moving companies aren’t simply adding AI to their software.
They’re using AI to do the work their software previously helped humans perform.
And that could fundamentally change what customers are willing to pay for.
Stop Selling Software. Start Selling the Job
One of Nathan’s examples was Flossy, a software company serving dental practices.
Its original proposition looked much more like traditional software. Today, the company sells AI agents focused on specific jobs inside a dental practice, including booking patients, re-engaging them and getting them back into appointments.
That’s an important distinction.
A generic AI assistant might answer a question.
An agent built around a specific industry workflow can produce an outcome.
For a dentist, the value isn’t that an AI model generated text. It’s that another patient ended up in the chair.
That shifts the conversation from features and seats towards jobs to be done.
The question for SaaS companies becomes less about, “What AI functionality should we add?”
It becomes, “What work are our customers currently paying humans to do that our product could start doing for them?”
The Dashboard Is Losing Its Monopoly
For years, the dashboard has been the default interface of SaaS.
Log in. Look at information. Click buttons. Move things around. Export something. Make a decision. Repeat tomorrow.
AI creates a different possibility.
Nathan highlighted companies building what he described as closed-loop systems. Instead of giving users information and waiting for them to act, the software can receive an objective, perform the work, measure the result and adjust.
Imagine an advertising platform where the user doesn’t manually create hundreds of campaigns. Instead, they give the system a budget and an outcome: spend $5,000 testing 1,000 ad variations and optimise towards a target return within 30 days.
The agents perform the experiment. Winners continue. Losers disappear.
The interface still exists, but it is no longer where most of the value is created.
The value is in what happens after the user stops clicking.
AI Changes What You Can Charge For
This also creates an interesting pricing problem.
Traditional SaaS pricing has largely been built around access.
Seats. Licences. Usage. Locations. Features.
But if software increasingly performs the job itself, those units become less obvious.
If an AI agent can complete work that previously required an employee, charging based purely on how many people log into the platform may actually disconnect pricing from the value being created.
One founder Nathan highlighted described this emerging category as “result as a service.”
The phrase matters because it reframes what customers are buying.
They’re no longer simply paying for the capability to do something.
They’re paying to have it done.
That creates an opportunity for SaaS companies to rethink pricing around the economic value of the work rather than the number of humans operating the software.
Your Integrations Are Becoming an AI Strategy
There is a catch.
AI can only perform meaningful work when it has enough context.
That makes something surprisingly unglamorous increasingly valuable: integrations.
Nathan used Rev, a platform serving automotive repair businesses, as an example. The product connects with the systems already used inside a repair shop, allowing it to understand the vehicle, identify the work required, surface repair instructions and help produce the documentation required to complete the process.
The deeper those connections become, the more useful the AI becomes.
That means the integrations page many SaaS companies originally built for product convenience or SEO may become a strategic asset.
Every connected system gives the platform more context.
More context allows the AI to perform more work.
And performing more work makes the platform harder to replace.
The Real AI Moat May Be Your Data
Foundation models are available to everyone.
That makes “we use AI” an increasingly weak differentiator.
Nathan’s examples pointed towards a different moat: proprietary data.
Golf Genius has millions of users and deep domain knowledge inside the golf industry. Vantaca operates across millions of homes and can use that data and context to automate work inside property management. Founderpath has thousands of financial data connections from software companies using services such as Stripe and QuickBooks.
The model isn’t necessarily the valuable part.
The context surrounding it is.
A generic AI model might be extraordinarily capable, but it doesn’t automatically understand the history of your customers, the workflows inside your industry, the transactions moving through your platform or the specialised rules governing a particular market.
Companies that already sit on that information have an opportunity.
Their existing SaaS product may be the infrastructure that allows AI to become genuinely useful.
Trust Determines How Much Work AI Can Do
But access to data requires something else.
Trust.
Nathan described how Founderpath allows software founders to connect financial accounts and request capital through a conversational interface. Behind the scenes, the system can analyse the information, create an underwriting decision and, when approved, move money.
That’s very different from adding a chatbot to a dashboard.
The AI is being trusted with an economically meaningful action.
And the more consequential the action, the more important the customer’s relationship with the underlying company becomes.
This creates an interesting paradox in the AI era.
As software becomes more autonomous, trust may become more important, not less.
Customers have to trust you with their data before you can build context. They have to trust your system before they allow it to act. And they have to trust the outcome before they hand over even more responsibility.
The companies that earn that trust can gradually move from helping customers work to actually doing the work.
AI Could Completely Change SaaS Efficiency
The shift isn’t only changing products.
It’s changing companies.
Nathan highlighted Arcads, which reached around $15 million in revenue with a team of roughly eight people. He argued that AI is pushing expectations for revenue per employee dramatically higher than the benchmarks SaaS leaders have historically used.
That doesn’t mean every SaaS company should suddenly cut its workforce until it reaches some new efficiency target.
But it does suggest that the relationship between headcount and growth is changing.
If AI can write code, analyse customer data, operate workflows, generate content and perform repetitive work inside both the product and the company, adding revenue may no longer require adding employees at the same rate.
For founders, that creates a very different growth equation.
The next generation of category leaders may not just grow faster.
They may grow with fundamentally different cost structures.
The Best AI Companies Won’t Look Like AI Companies
There is no shortage of companies adding AI to their homepage.
But “AI-powered” is quickly becoming meaningless.
What matters is what sits underneath it.
Do you have unique data?
Are you deeply connected to the systems your customers already use?
Do customers trust you enough to give your software access to that data?
Can your AI perform a meaningful job rather than simply generate an answer?
And can you connect that work directly to an outcome customers are willing to pay for?
Those questions are much harder than adding a chatbot to a product.
They’re also much harder for competitors to copy.
The biggest opportunity for existing SaaS companies may therefore not be reinventing themselves as AI companies.
It may be recognising what they’ve spent years quietly building: customer relationships, integrations, workflows, proprietary data and deep knowledge of a specific market.
AI can turn those assets into something much more powerful.
For the last twenty years, SaaS helped people do their jobs.
The next generation of SaaS may increasingly do the job for them.
