Software Isn’t Dead. SaaS Pricing Is.

Looking back over more than four decades in the software industry — from the earliest days of packaged systems in the 1980s to today’s AI-powered world — one pattern has stood out: technology evolves, but business value never dies quietly. What does change, relentlessly, is how software captures that value.

I began my career in 1985, when “software” was a line item on a capital expenditure sheet and buyers expected to install code on metal they owned. From client-server architectures to the browser era, from ERP waves to SaaS subscriptions, every transition reshaped pricing, contracts, and economics. SaaS pricing became conventional wisdom not because it was perfect, but because it worked — for a while.

Today, with the rise of generative AI agents and tools that can automate tasks previously embedded in enterprise systems, we are at another such inflection point. The question isn’t whether software will survive. It is whether the old assumptions behind how we price and monetize software services still hold.

Why the market panic feels familiar but is misguided

Recently, markets reacted as though a new class of AI tools meant the end of enterprise software as we know it. That reaction reflects anxiety over what AI might do to pricing power, not what it can do to underlying platforms. Core systems of record — the financial systems, CRMs, ERPs, and infrastructure that enterprises depend on — do not vanish simply because an agent can review a contract or automate a workflow.

What’s changed is the interpretive lens buyers now have on their software spend. Instead of accepting seat counts and list prices as defaults, procurement and finance teams can point to an AI benchmark and ask: What am I paying for? What actual work is being done? Can this part be done cheaper?

That dynamic doesn’t kill vendors. It forces vendors to defend their pricing — continuously, publicly, and in every renewal.

The real shift isn’t technical — it’s economic

In earlier technology transitions, the value debate centered on capability: Can this system do more? Today’s discussion centers on value for dollars spent. AI doesn’t obliterate the need for enterprise systems. Rather, it reveals that many pricing structures were built on convenience and user counts rather than outcomes and risk mitigation.

When internal teams can automate contract review, basic analytics, or routine workflows with prompt-driven agents, seat-based pricing starts to feel arbitrary. Not because the software has no value, but because the proportional relationship between seats and value breaks down.

It’s not uncommon today for companies to retain their core vendor contracts while introducing AI agents alongside them — not as replacements, but as benchmarks and bargaining chips. This recalibration is neither destruction nor dismissal of software platforms. It is price discipline applied with unprecedented transparency.

Seat-based pricing: an artifact of a bygone era

One of the defining economic conventions of SaaS has been seat-based pricing. It made sense when human capital was the scarce variable. Software, in that context, enabled more productivity per person — and charging per seat tracked reasonably with perceived value.

AI challenges that premise. If a single agent can accomplish the work of multiple seats, the traditional seat metric loses relevance. Buyers are now starting to ask questions that, until recently, were hard to operationalize:

What exactly am I buying?
What portion of this service could be replicated internally with AI tools?
Why should automation drive my bill upward when it drives my workload downward?

These are questions of pricing logic, not production logic. They expose the old model as one of convenience and habit, not economic precision.

What survives and what doesn’t

In every previous technology wave, the forces that survive are:

• Proprietary data that is hard to replicate.
• Deep integration with operational processes that can’t be compromised.
• Risk transfer and compliance guarantees that buyers cannot self-insure.
• Continuity and auditability in mission-critical functions.

These are not threatened by AI — they are reinforced by it. AI may automate tasks, but it cannot guarantee outcomes with the same determinism that regulated enterprises require.

Software vendors who adapt will integrate AI advances into their products, not resist them. Their offerings will remain essential because they combine automation with governance — and that combination commands a premium.

Software that is purely UI-centric or that bundles thin logic into expensive licenses, by contrast, will struggle. It isn’t software that dies. It’s pricing models that cannot justify their economics in an age where alternatives are transparent, accessible, and cheap in comparison.

A new era of accountability in pricing

The transition I am watching today is less dramatic than headlines suggest, but deeper and more permanent. It is a shift of discipline: from price by seat to price by value delivered and outcomes guaranteed.

In earlier decades, buyers accepted seat counts and license tiers as proxies for value because there was no alternative. Today, with AI serving as a continuous benchmark, those proxies are losing their credibility.

The question for software providers is no longer simply, Can we integrate AI?
It is: Can we justify our pricing in an era where automation is ubiquitous and transparency is non-negotiable?

Software isn’t dead.
But the era of unquestioned SaaS pricing is.