Why Per-User Licensing and GenAI Could Sink the Ship

Sat 23 2024

Why Per-User Licensing and GenAI Could Sink the Ship

by bernt & torsten

Software-as-a-Service (SaaS) companies have ridden the subscription wave for years, promising scalability, affordability, and ease of use. But this once-revolutionary model shows cracks, particularly the widely adopted per-user license model. Customers feel they're being squeezed dry. At the same time, SaaS companies themselves risk spiralling costs due to generative AI (GenAI) features they’re shoehorning into their platforms.

The result? Both customers and SaaS companies are staring down the barrel of unsustainability. This article explores why the per-user licensing model is failing, how the explosion of GenAI-driven costs accelerates the problem, and why the entire SaaS industry must pivot toward "agentic" solutions—or risk irrelevance.

But here’s the thing: I think there’s a lot to learn from this for individuals, too. No company or job is doomed to irrelevance because of AI. Sure, there might be a handful of exceptions, but in most cases, your framing and what you do to adapt will largely shape your future. The great news? AI disruption is not inevitable—you hold the last word in whether you or your company thrives. Failing to adapt is failing to survive, but the power to prevent it is entirely in your hands.

Why the SaaS Subscription Model is Failing

At its core, the SaaS model relies on recurring revenue. Users pay a subscription fee—often per user—granting them access to software that’s maintained and updated centrally by the provider. On paper, it’s a win-win: customers avoid the upfront cost of traditional licenses, and SaaS companies enjoy predictable revenue streams.

But in practice, the cracks have become impossible to ignore.

1. The Per-User License Trap

Customers hate the per-user license model, and for good reason. It assumes a linear relationship between a SaaS tool's value and the number of users accessing it. Yet, many organizations experience diminishing returns as user counts grow. A 50th user isn’t necessarily doubling the value of a tool compared to the first.

Meanwhile, from the company’s perspective, the per-user model encourages aggressive expansion tactics. Sales teams are incentivized to sign as many seats as possible, often creating waste as organizations are forced to pay for users who rarely (if ever) engage with the software. This “seat stuffing” has led to a deep distrust between SaaS companies and their customers.

2. Generative AI: The Hidden Cost Bomb

The advent of generative AI features has compounded the problem. SaaS companies are rushing to integrate AI-powered tools, like chatbots, recommendation systems, and automated workflows. But these capabilities don’t come cheap.

Every GenAI query incurs computational costs that scale with usage. Unlike static SaaS features with predictable costs, GenAI systems (like OpenAI’s GPT or Anthropic’s Claude) rely on expensive, cloud-hosted inference models. For some SaaS companies, offering these features for "free" with their existing subscription tiers means their costs are rapidly outpacing their revenue.

Case in Point: The Slack and Salesforce Problem

Take Slack, now a Salesforce subsidiary. Slack recently added GenAI-powered features for summarizing conversations or generating action items. Each use of these features incurs backend API costs from OpenAI. Now imagine an enterprise customer with 5,000 users—if just 10% of those users interact with GenAI daily, costs balloon exponentially.

Salesforce, the parent company, faces the same issue. With their Einstein GPT suite, Salesforce risks losing profitability on lower-tier customers who leverage GenAI but don’t generate enough subscription revenue to offset these new expenses.

SaaS Companies at Risk

Several SaaS companies are particularly vulnerable to these dynamics:

Microsoft (Office 365, Teams)
Microsoft’s Copilot features have made waves, but they are capital-intensive. Imagine embedding AI in Word, Excel, and Teams for millions of users. The costs of serving AI requests at this scale may erode profits even for this tech giant.
Atlassian (Jira, Confluence)
Atlassian, a staple for software teams, has integrated AI features for task automation and documentation. Yet, smaller customers on entry-level plans could exploit these features disproportionately, creating a mismatch between costs and revenue.
Zoom
Zoom’s AI transcription and summarization tools are another example. While AI makes the platform more attractive, heavy usage by large teams could turn into a financial liability if subscription fees don’t cover backend AI costs.
HubSpot
HubSpot’s recent foray into AI-driven marketing automation tools means a similar problem. While AI might drive engagement, the cost-to-revenue ratio could skew dangerously.

The Internal Build Revolution

A growing trend is adding to the pressure on SaaS companies: major enterprises are rethinking their reliance on third-party SaaS platforms. Instead of subscribing to expensive services, they’re building internal solutions, often powered by AI tools that automate development workflows.

Why Companies are Ditching SaaS for Internal Builds

Cost Efficiency
Companies like Netflix, Amazon, and Shopify have already demonstrated the power of internal tools. By investing in development teams and leveraging open-source AI models, they’re cutting long-term costs while tailoring solutions to their exact needs.
Data Privacy
With growing concerns about data privacy, many organizations feel uneasy sending sensitive information to external SaaS platforms. Internal builds give them greater control over their data.
AI Tools Lower the Barriers
Tools like GitHub Copilot and AI-driven code generators mean companies no longer need armies of developers to create sophisticated internal software. AI reduces development time and costs, making the "build vs. buy" decision lean heavily toward "build."

Case Study: Basecamp

Basecamp famously abandoned popular SaaS tools like Slack, building its own chat and project management software. While it’s not a GenAI-first company, its strategy highlights the appeal of avoiding SaaS entirely.

The Pivot to Agentic Software

So, where does this leave SaaS companies? The answer lies in embracing an agentic model. Instead of building software that “does” things for users, SaaS companies must focus on enabling autonomous agents and robust workflows.

What is Agentic Software?

Agentic software empowers users to describe what they want to be done rather than manually executing tasks. Think of it as transitioning from offering tools to providing highly capable virtual assistants. These assistants can autonomously:

Execute multi-step workflows
Interact across different platforms
Adapt to user preferences and contextual data

Why Workflows Matter More Than Features

In the AI era, standalone features will lose their competitive edge. Instead, customers will gravitate toward platforms that offer:

End-to-End Workflow Automation: Seamlessly integrating AI-driven actions across tools.
Adaptability: Solutions that grow smarter over time.
Agentic Enablement: Allowing users to delegate complex tasks entirely.

Example: Canvas Agentic Pivot

Canva is an example of a SaaS company making this shift. Its AI tools for graphic design don’t just create assets—they assist users in ideating, iterating, and exporting designs in various formats, effectively acting as creative agents.

The Risks of Failing to Adapt

SaaS companies that cling to the old model—layering GenAI on top of broken subscription models—will find themselves in trouble. They face several risks:

Customer Backlash
With rising subscription prices and underwhelming value, customers will churn.
Competitive Pressure
As companies build their tools, SaaS providers will lose their market share.
Unsustainable Costs
AI-powered features will bleed companies dry if they aren’t monetized appropriately.

The Takeaway: The Future is in Your Hands

The SaaS subscription model is cracking under the weight of GenAI costs, customer dissatisfaction, and a growing internal-build revolution. If SaaS companies want to survive, they must pivot to agentic software—solutions that empower customers to automate workflows, deploy intelligent agents, and achieve outcomes with minimal input.

But here’s the empowering truth: SaaS companies—and individuals—are not sentenced by these challenges. Adaptation is the key. AI disruption can be an opportunity rather than a threat, and it’s entirely in your hands to turn the tide.

This is the dawn of the AI era. Whether you’re running a SaaS company or planning your career, the ability to evolve will determine your success. If you seize the moment and lean into this transformation, you won’t just survive—you’ll thrive.