There’s a popular thesis circulating through analyst desks and investor conferences right now. It goes something like this: artificial intelligence will devour the traditional software business model. AI agents will replace enterprise applications. The great software-as-a-service franchises that minted billions in recurring revenue over the past two decades are finished.
It’s a clean narrative. It’s also probably wrong.
As The Motley Fool recently argued, the consensus view that AI will destroy incumbent software companies fundamentally misreads how enterprise technology adoption actually works. The bear case treats AI as a replacement for software platforms. The reality is something more nuanced and, for investors willing to look past the panic, far more profitable: AI is becoming a feature embedded inside existing software, not a substitute for it.
Consider the numbers. Shares of major SaaS companies have been under sustained pressure since late 2024, with many trading at significant discounts to their five-year average multiples. The logic behind the selloff isn’t complicated. If an AI agent can write code, generate marketing copy, manage customer service tickets, and automate financial workflows, why would enterprises keep paying for Salesforce, ServiceNow, or Adobe? Wall Street’s version of creative destruction has software incumbents cast as the next batch of Blockbuster Videos — too slow, too bloated, too legacy to survive the transition.
But here’s the thing about enterprise software that armchair disruption theorists consistently underestimate: switching costs are enormous, data moats are real, and procurement cycles move at glacial speed.
Large enterprises don’t rip out their CRM systems because a startup demo’d an impressive chatbot at a conference. They don’t abandon platforms where years of proprietary data, custom integrations, and trained workflows reside. The friction of enterprise technology replacement is measured not in months but in years — and often in tens of millions of dollars in implementation costs. This isn’t a consumer market where users can download a new app overnight. It’s a world governed by security reviews, compliance requirements, IT governance committees, and multi-year contracts.
The Motley Fool’s analysis highlights a pattern that should be obvious but somehow isn’t to many on the Street: the largest software incumbents aren’t sitting still. They’re embedding AI directly into their existing products, turning what bears see as an existential threat into a revenue accelerator.
Microsoft is the most visible example. The company has woven its Copilot AI assistant across the entire Microsoft 365 product line, from Word and Excel to Teams and Outlook. It hasn’t replaced these applications with AI. It has made them stickier. More valuable. Harder to leave. And it’s charging a premium for the privilege — Copilot licenses add $30 per user per month on top of existing Microsoft 365 subscriptions. For a company with hundreds of millions of commercial Office users, even modest adoption rates translate into billions in incremental annual revenue.
Salesforce has pursued a similar strategy with its Agentforce platform, which layers AI agents on top of its existing CRM infrastructure. ServiceNow has integrated AI across its IT service management and workflow automation tools. Adobe has built generative AI features — branded as Firefly — directly into Photoshop, Illustrator, and its broader Creative Cloud offering. In each case, AI isn’t cannibalizing the core product. It’s enhancing it and creating new pricing tiers.
This is the part the disruption narrative misses entirely.
Software companies with massive installed bases, deep enterprise relationships, and proprietary datasets are uniquely positioned to deploy AI in ways that startups simply can’t replicate overnight. An AI model is only as good as the data it’s trained on and the workflows it’s integrated into. And incumbents own the data. They own the workflows. They own the customer relationships. A startup with a clever foundation model still has to convince a Fortune 500 CFO to hand over sensitive financial data, pass a six-month security audit, integrate with seventeen legacy systems, and train thousands of employees on a new interface.
Good luck with that.
None of this means incumbents are guaranteed winners. Complacency kills. Companies that treat AI as a marketing buzzword rather than a genuine product transformation will lose share. But the blanket assumption that AI equals death for software stocks reflects a fundamental misunderstanding of how technology transitions actually play out in large organizations.
History offers useful parallels. When cloud computing emerged in the late 2000s, the prevailing wisdom held that it would destroy traditional enterprise software vendors. Oracle, Microsoft, and SAP were supposedly doomed — too tied to on-premise licensing models, too slow to adapt. What actually happened? All three eventually became major cloud platforms. Microsoft’s Azure is now a $100 billion-plus annual revenue business. Oracle’s cloud infrastructure division is growing at over 40% year-over-year. The transition was painful and took longer than anyone predicted, but the incumbents that invested aggressively survived and, in many cases, thrived.
The cloud transition also produced enormous new companies — Salesforce, Workday, ServiceNow. So it’s entirely possible that the AI era will mint its own class of winners that don’t exist yet or are barely visible today. But the idea that existing software giants will simply vanish requires ignoring everything we know about enterprise technology adoption patterns.
There’s also a financial dimension that the bear case conveniently overlooks. The leading SaaS companies generate enormous free cash flow. Salesforce produced over $12 billion in free cash flow in its most recent fiscal year. Microsoft’s free cash flow exceeds $70 billion annually. These companies have the capital to acquire AI startups, invest in proprietary model development, and fund massive R&D efforts — all while returning cash to shareholders through buybacks and dividends. Startups burning through venture capital don’t have that luxury.
And the venture capital environment itself has shifted. After the initial frenzy of AI startup funding in 2023 and 2024, investors are growing more selective. The realization is setting in that building a foundation model or an AI application is expensive, that margins in AI-native businesses are often thinner than in traditional SaaS due to compute costs, and that customer acquisition in the enterprise market requires exactly the kind of sales infrastructure that incumbents already possess.
So where does this leave investors?
The Motley Fool’s thesis suggests that the current pessimism around software stocks has created a buying opportunity. Companies trading at depressed valuations relative to their cash flow generation, competitive positioning, and AI integration progress may be significantly undervalued. The market is pricing in disruption that, while real in some segments, is unlikely to materialize as broadly or as quickly as the bears assume.
That doesn’t mean every software stock is a buy. Differentiation matters. Companies with weak competitive moats, commoditized products, or limited AI strategies face genuine risk. Point-solution vendors in categories where AI can fully automate the underlying task — think basic data entry, simple report generation, or template-based content creation — are more vulnerable than platform companies with broad functionality and deep integrations.
But the Microsofts, Salesforces, and ServiceNows of the world? They aren’t victims of the AI transition. They’re architects of it. And Wall Street’s failure to distinguish between software companies that will be disrupted and those that will do the disrupting has created a pricing anomaly that patient investors may look back on as one of the better opportunities of this cycle.
The narrative that AI kills software makes for a compelling conference panel. It makes for a less compelling investment thesis. Enterprise technology doesn’t work that way. It never has. The companies that own the data, the workflows, and the customer relationships will capture a disproportionate share of AI’s economic value — not because they invented the technology, but because they control the surfaces where it gets deployed.
Wall Street loves a clean disruption story. Reality is messier. And in that mess, the incumbents are quietly building the future while the market prices them for obsolescence.
WebProNews is an iEntry Publication