For most of the last fifteen years, the rule for business software was simple: pick a good platform, pay the monthly subscription for every member of staff, and let the software company do the heavy lifting. In 2026 the stock market tore that rule up. Some of the best-known names in business software have lost half or more of their value in a matter of months, even as their revenues kept growing. Yet one of the oldest names in the sector, Salesforce, has clawed its way back.

The gap between those outcomes is the most useful lesson we’ve seen in years about where business systems are heading. It isn’t really about share prices. It’s about what customers will still pay for when AI can do so much of the work, and about an option that, for most businesses, simply wasn’t realistic until now: software built around you, that you own.

---

The “SaaSpocalypse” in brief

In January, Anthropic launched Claude Cowork, an AI agent that can work across a computer’s applications on someone’s behalf. Within weeks, investors were asking an awkward question: if one AI agent can do the work of several people, why would a business keep paying for several software seats? By the end of February, roughly a trillion dollars had been wiped off enterprise software valuations. A trader at Jefferies called it the “SaaSpocalypse”, and the name stuck.

Two fears drove the selling.

• Seat compression. Most business software is priced per user, so fewer human users means less revenue.
• Do-it-yourself. AI coding tools have made it far cheaper for businesses, and the developers who serve them, to build exactly what they need rather than renting a one-size-fits-all product.

---

Monday.com: a growing business with a collapsing share price

Monday.com is the clearest example of how harshly the market has judged this. Its shares fell by around half in the first six months of the year, are down more than 70% over twelve months, and sit more than 80% below their peak.

What makes this remarkable is that the business itself has kept growing. First-quarter revenue rose 24% year on year, and the number of large enterprise customers grew even faster. Second-quarter revenue was up 22%. Each time, the shares fell anyway. In August, following a 20% workforce reduction to refocus the company on AI and guidance that pointed to slowing growth, several analysts downgraded the stock.

---

Salesforce: the fall and the comeback

It would be easy to conclude that every subscription software company is doomed. Salesforce suggests otherwise, although it’s worth being precise about what happened.

Salesforce did not escape the sell-off. Its shares fell by as much as 37% this year, hitting a low in June, and by July it was among the worst performers in the Dow Jones. What happened next is what sets it apart. A strong second quarter, with $11.35 billion in revenue and raised guidance, signalled a turn. Its AI agent platform, Agentforce, passed $1 billion in annual recurring revenue and has since moved beyond $1.5 billion. Between the end of July and mid-September the shares rose by around 36%. By late September Salesforce was down roughly 12% for the year, while Monday.com was still down by around half.

---

Why the market sees Monday and Salesforce so differently

On the surface they look alike. Both are cloud subscription businesses, both are mostly priced per user, and both are betting heavily on AI agents. Investors, however, see four important differences.

1. A system of record versus a layer of convenience

Salesforce holds the customer data that large organisations run on: years of sales history, service records, contracts and relationships. AI agents need that data to do anything useful, so Salesforce becomes the place where agents work rather than something they replace. Monday is primarily a way of organising and visualising work. That is valuable, but it is precisely the layer AI is best at recreating, or bypassing altogether.

2. A credible way to charge for AI

If seats shrink, revenue has to come from somewhere else. Salesforce has shown it can charge for agent activity and premium AI bundles, and its Agentforce figures give investors something concrete to measure. Monday is rebuilding its strategy around an AI work platform but has not yet proved that AI usage will replace lost seat revenue.

3. Who the customer is

Salesforce sells to large enterprises with long contracts, complex integrations and very high switching costs. Monday is far more exposed to small and mid-sized businesses and individual teams: the customers who can switch quickly, cut costs fast and are most tempted to try an AI-built alternative.

4. Scale and cash

Salesforce generated around $15 billion in operating cash flow in its last financial year and is returning capital through a $50 billion share buyback. That gives it room to invest through the transition and puts a floor under its valuation. Monday’s reported profits are thin, so investors are far less patient with it.

---

Bespoke is now a real solution

That last point is the one that matters for a business owner. If the value of software lies in the data and the workflow rather than the interface, the obvious question is: why rent someone else’s interface for your data and your workflow?

For years the honest answer was cost. A system built around the way your business actually works was a large, slow, risky project, affordable mainly to big organisations. So smaller firms did the sensible thing: they bought several off-the-shelf tools, paid for every seat, and bent their processes to fit. The result is familiar: data spread across subscriptions that don’t talk to each other, information re-keyed by hand, and a monthly bill that grows every time you take someone on.

AI has changed that equation. The same tools that spooked investors have dramatically reduced the time it takes an experienced developer to build, test and maintain custom software. The parts that once consumed most of a budget are now far quicker, which leaves the time where it belongs: understanding your business and getting the design right. Projects that would have been out of reach a few years ago are now within reach of a firm of ten or twenty people.

It is worth being clear about where bespoke isn’t the answer. If your process is genuinely standard, such as straightforward accounting or email, a good off-the-shelf product is still the right choice, and the best bespoke systems connect to those tools rather than replacing them. Bespoke earns its place where the work is specific to you: inspections, compliance records, client portals, job management, quoting, scheduling, the things that make your business different and that generic software handles badly.

---

What this means for your business

For a business owner, none of this is really about which shares to own. It’s about where to place your trust and your budget over the next five years. We’d draw four conclusions.

• Own your data and your system. The companies that held up best own something hard to replicate. The same principle applies to you. A system you control, with data you can move, is a stronger position than one locked inside a platform whose pricing and priorities are changing.
• Look hard at what you pay per seat. Add up every subscription your team uses and what each one costs as you grow. For many businesses that figure, over three to five years, now compares well with a system built and owned outright.
• Get your systems ready for AI. AI agents are only as useful as the data they can reach. Clean, structured data in one system you control is what lets you put AI to work on your processes, rather than waiting for each vendor to add it.
• Judgement matters more than tools. When anyone can generate software in minutes, the value moves to knowing what to build, making it secure, connecting it to the rest of the business and being accountable when something needs fixing. That is the part AI cannot take responsibility for.

---

Our view at Cirrus

We’ve watched this shift up close. We build bespoke cloud systems for businesses in construction, compliance and healthcare, from geo-located site inspections and warranty portals to CPD management and client portals, each hand-coded around the way that business actually operates. AI hasn’t replaced that work. It has made it faster and more affordable, which means more businesses can now have it.

The lesson of 2026 is not that software is dead. It is that value has moved from renting tools to owning outcomes. The businesses that recognise this early will be in the strongest position when the market settles.

If you’d like to talk about what this means for your business systems, get in touch with the Cirrus team in Poole.

Get in touch See our case studies

Share price figures are approximate and based on publicly reported data as at early October 2026. This article is for general information only and is not investment advice.