Why Enterprise Giants Can’t Build AI That Actually Works
By Alex Turgeon, President, Valere
In past articles, I broke down my perspective on why I think traditional software is becoming yesterday’s news. Today, I’m tackling the follow-up question everyone’s asking: if AI is such an obvious win, why aren’t the big players dominating?
You’d think companies like Shopify, Toast, and ServiceTitan would be crushing this space. They’ve got the data. They’ve got the customers. They’ve got the cash.
Instead, they’re shipping AI features that feel about as revolutionary as adding cup holders to a horse-drawn carriage.
There’s a reason for this. And it’s not what you think.
The Uncomfortable Truth About AI and Incumbents
Here’s what nobody wants to say out loud: true AI automation threatens to make traditional software companies obsolete.
Think about it. Your entire business model depends on humans needing your software. You charge monthly subscriptions because people can’t do their jobs without your platform.
Now imagine launching an AI agent that does the work for them. That’s not a feature upgrade. That’s business suicide.
This creates a fascinating tension. Incumbents need to appear innovative without actually innovating too much. They need AI credibility without AI disruption.
The solution? Chatbots. Lots and lots of chatbots.
The Great Chatbot Cop-Out
Walk into any enterprise software company’s product meeting. You’ll hear the same conversation happening everywhere:
“How do we respond to all this AI stuff?” “Let’s add a chatbot.” “Perfect. Ship it.”
This approach feels safe. Chatbots sit on top of existing systems. They don’t change core workflows. They answer questions about data that’s already there.
But here’s what’s actually happening in the market while incumbents play it safe. Startups are building AI agents that eliminate entire job functions. They’re not surfacing information. They’re taking action.
The difference? Startups don’t have sacred cows to protect. They can build products that make traditional software irrelevant because they have nothing to lose.
Incumbents are stuck trying to innovate without cannibalizing their cash cows. It’s like asking Netflix to build a service that automatically watches shows for you.
Organizational Dysfunction Meets AI
The second killer? Most big companies treat AI like a science experiment.
You’ll find a tiny team tucked away somewhere. Maybe five engineers and a couple product managers. They’re trying to bolt AI features onto products that weren’t designed for it.
These teams don’t have real power. They can’t make major architectural decisions. They definitely can’t reimagine the entire user experience around what AI can actually do.
Meanwhile, the core product team continues business as usual. Building the same workflows. Solving the same problems. Acting like AI doesn’t exist.
Real AI transformation requires rethinking everything from scratch. You can’t sprinkle machine learning on top of a workflow management system and call it revolutionary.
But asking a $10 billion SaaS company to rebuild their product from the ground up? That’s organizational suicide. Too much risk. Too much disruption. Too many stakeholders saying no.
Technical Debt vs AI Requirements
Even when incumbents want to swing bigger, they hit walls that startups never face.
Modern AI applications need ultra-low latency. Voice agents require sub-200ms response times. Real-time automation demands seamless data flow between systems.
Most enterprise software wasn’t built for these requirements. These platforms were designed for humans who can wait three seconds for a page to load. AI agents can’t wait three seconds for anything.
Startups get to build their entire tech stack around AI from day one. They choose databases, APIs, and architectures that support responsive, intelligent applications.
Incumbents are stuck trying to retrofit Ferrari engines into horse carriages. The results are about as graceful as you’d expect.
What “AI Innovation” Actually Looks Like at Big Companies
Let’s examine what major players are shipping under their “AI strategy” banners:
Shopify: AI-generated product descriptions. Nice time-saver. Doesn’t change how merchants run their business.
Toast: Price benchmarking tools. Compares your restaurant’s prices to competitors. Useful data. Zero automation.
Square: Spend insights chatbot. Still in beta. Launching next July. Answers questions about transaction data you already have.
Mindbody: 24/7 AI receptionist. Handles basic appointment scheduling. Getting warmer, but pretty limited scope.
ServiceTitan: Job value predictor. Estimates service call profitability. Helpful for pricing. Doesn’t automate actual work.
Notice the pattern? These features surface insights or handle simple tasks. None eliminate major workflows. None make their core products less necessary.
They’re incremental improvements wearing AI costumes.
ServiceTitan deserves credit for pushing further than most. Their voice agent partnership with Siro shows bigger ambitions. They’re building an AI app marketplace.
But even they face the same fundamental pressure: protect the core business. Don’t disrupt too much.
The API Leverage Game
Here’s where incumbents do flex their muscles: data access.
Building an AI startup for restaurants? You probably need Toast’s data. Construction software? You’ll want ServiceTitan’s API. Healthcare? Good luck without Epic or Cerner.
This creates interesting dynamics. Incumbents might not build great AI, but they can control who gets access to the data needed to build it.
API pricing becomes a competitive weapon. Integration approvals turn into strategic decisions.
Some companies are getting aggressive with this approach. They’ll offer basic API access but charge premium rates for real-time data that makes AI agents effective.
Others slow-roll approvals for competitive applications while fast-tracking partnerships that complement their existing strategy.
Smart incumbents use this time to figure out their AI approach. They might not build game-changing AI internally, but they can control the competitive landscape while they try.
Why This Creates Massive Startup Opportunity
The incumbent struggle opens a window that won’t last forever. While established players fumble with organizational politics and technical debt, startups can capture entire market segments.
But success requires understanding the constraints you’ll face:
Data Dependencies: Plan for expensive or fragile API access. Build relationships with incumbents early. Consider what happens if they cut you off or raise prices.
Incremental Rollouts: End-to-end automation doesn’t happen overnight. Users need time to trust AI with critical workflows. Design your product for gradual adoption.
Service Bundling: Sometimes AI isn’t enough. You might need to offer human services alongside automation to win customers who aren’t ready for full AI workflows.
The companies that crack this formula won’t just compete with incumbents. They’ll make them irrelevant.
Traditional software companies help humans work more efficiently. AI-native platforms eliminate the need for humans in the loop entirely.
That’s not a feature upgrade. That’s a completely different game.
The Innovator’s Dilemma in Real Time
This situation perfectly demonstrates Clayton Christensen’s innovator’s dilemma. Incumbents optimize for existing customers who want incremental improvements.
They can’t justify cannibalizing profitable products for uncertain AI futures. Their best customers aren’t asking for full automation. They’re asking for better reporting and smoother workflows.
Meanwhile, startups target underserved segments or entirely new use cases. They don’t need to protect existing revenue streams. They can afford to build products that make traditional software obsolete.
The result? Incumbents ship safe, incremental AI features while startups go after transformational automation.
Guess which approach users prefer once they experience it?
The Technical Reality Check
Let’s talk about what building effective AI actually requires. Because it’s not just about having smart algorithms.
Voice agents need ultra-low latency networks. Workflow automation requires clean API integrations. Advanced AI features demand computing resources that legacy infrastructure can’t support without major overhauls.
Most enterprise platforms were built for a different era. They assumed humans would be in the loop for every decision. They optimized data storage and retrieval, not real-time processing and response.
AI changes everything about how software needs to work. Response times that felt fast for humans feel sluggish for AI. Data formats that worked fine for dashboards create bottlenecks for automation.
Retrofitting these systems for AI is like trying to turn a library into a race car. The fundamental architecture wasn’t designed for speed.
User Expectations Are Shifting Fast
Here’s what makes this urgent for incumbents: user expectations are changing rapidly.
People are getting comfortable with AI automation in other parts of their lives. ChatGPT writes their emails. AI assistants manage their calendars. Smart home systems automate their environments.
They’re starting to wonder why their business software is so manual and clunky by comparison.
This creates a dangerous gap for traditional software companies. Users are becoming AI-literate. They understand what’s possible. They’re getting impatient with platforms that require endless clicking and form-filling.
The companies that figure out how to bridge this gap will survive. Those that keep treating AI as a feature add-on will find themselves competing against platforms that eliminate the need for their core products entirely.
The API Control Strategy
While incumbents struggle with internal AI development, many are doubling down on their data moats. If you can’t beat the AI startups, control their access to the data they need.
This strategy has some short-term merit. Startups building AI for specific industries often need deep integration with existing systems. That integration requires cooperation from incumbents.
But it’s ultimately a defensive play. Users will gravitate toward the best experience, even if it means switching platforms entirely.
The smartest incumbents are using their API leverage to buy time while they figure out their long-term AI strategy. The rest are just delaying the inevitable.
What Winning Looks Like for AI Startups
For new entrants, the message is clear: this opportunity won’t last forever. Incumbents will eventually figure out how to build better AI or acquire companies that already have.
The time to move is now, while organizational dysfunction and technical debt still handicap the competition.
But success requires more than just building cool AI. You need to solve real problems that incumbents can’t or won’t address.
That might mean automating workflows that threaten their core revenue. It might mean targeting market segments they’ve ignored. It might mean bundling services they can’t provide.
The key is understanding why incumbents are failing and building your strategy around those weaknesses.
The Race Against Time
Incumbents aren’t completely doomed, but their window for meaningful AI transformation is closing fast.
The companies that figure out how to cannibalize their own products before someone else does it for them might survive. Those that keep playing it safe will wake up one day to find their customers have moved on.
For the bold incumbents willing to take real risks, there’s still time. But it requires more than adding chatbots to existing products. It requires rethinking the entire value proposition around what AI makes possible.
The enterprise revolution isn’t just about new technology. It’s about new business models that make old ones obsolete.
The question isn’t whether this transition will happen. It’s whether you’ll lead it or get left behind.
The Bottom Line
Traditional enterprise software companies face an impossible choice. Protect their existing business model or embrace AI transformation that could make it irrelevant.
Most are choosing protection, which creates massive opportunities for startups willing to take bigger swings.
The window is open, but it won’t stay that way forever. The companies that move fast and think differently about what enterprise software can be will define the next decade of business technology.
Everyone else will be playing catch-up.
Ready to unlock the full potential of AI Agents in your enterprise in 2025? Contact us to learn more about how Valere can propel you on your AI journey.
About Valere
Valere is an award-winning technology innovation & software development company, utilizing emerging technology in Machine Learning (ML) and Artificial Intelligence (AI) to enable medium to large enterprises to execute, launch, and scale their vision into something meaningful.
About Alex
Alex Turgeon is President of Valere, a leading AI transformation and software development firm helping enterprises navigate the shift to agent-driven operations. Connect with Alex to discuss how your organization can begin its transformation to the agent era.
