The Enterprise Software Con: Why You’re Paying for AI That Doesn’t Exist

Inside the procurement tech industry’s dirty secret: customers funding R&D through three-year contracts for features that may never arrive

Enterprise software companies have perfected a lucrative business model. Sell the vision. Lock in multi-year contracts. Deliver the product later. Maybe.

A recent piece by Antonio Goncalves at vallor.ai exposed this pattern when he highlighted a question from Reddit asking if Coupa’s AI capabilities are real. The community’s response was brutal: “Anti-intelligence... the ultimate oxymoron!”

The issue extends far beyond one platform. Across the procurement tech landscape, enterprises are paying premium prices today for features promised tomorrow. They are funding vendor R&D budgets while their teams wait and manually process contracts.

This is not innovation. This is vapor.

What Reddit Really Thinks

The Reddit discussion revealed what enterprise customers experience but rarely say publicly. One user described Coupa’s AI as “anti-intelligence” and called it “the ultimate oxymoron.” Another admitted they cannot tell if the platform has real AI or if it is “just being plastered onto their product like every other ‘AI’ company out there.”

The skepticism is warranted. One commenter noted that Coupa recently acquired Cirtuo, which has some autonomous negotiation capabilities within limited use cases like renegotiating payment terms or revising liability clauses. But they do not see the company being able to integrate that across their platform. “It will likely stand as an independent solution with very light to no integration just like most of their other modules.”

Another user mentioned that Coupa has something called Coupa Navi. “It’s not as robust as it’s inspired to be though.”

The pattern is clear. Acquisitions get announced. Integration never happens. Features remain siloed. Customers wait.

The Enterprise Playbook

Ashley Dean Smith summarized the strategy perfectly. “Build as we bill is the oldest trick in the book. Most of these platforms are running on hopes, prayers, and PowerPoints.”

Chris D., who specializes in procurement transformation, elaborated on the real cost. “Enterprises are paying premium pricing, with annual uplifts, not for what works today, but for what’s promised tomorrow. In reality, many are unknowingly funding the vendor’s R&D and marketing engines, not their own digital transformation. You’re underwriting the next conference booth, conference sponsorship, ad campaign, and ‘coming soon’ slide deck, while waiting on features that may never fit your business or are not released fast enough.”

He urged companies to flip the model. “Stop paying for hype. Start investing in tools and partners that deliver measurable value now, not just roadmap rhetoric. Demand features and functions faster. Band together at their conferences and press for more, better, faster.”

Simone Ripamonti at SAP Ariba pointed to another major player. “I’m always looking forward the yearly SAP Ariba public announcements of revolutionizing the procurement landscape with their soon-to-be released cutting edge AI features.”

The response from Goncalves captured the industry’s fatigue. “Still waiting.”

The Reality Behind the Roadmap

Vincent Teyssier provided a damning example. “To be honest Coupa has been very poor at delivering new features. I’ve recently seen a big shipping company using Workato to integrate OpenAI to OCR invoices then post them in Coupa. Even such basics can’t be done properly inside Coupa.”

When basic invoice processing requires third-party workarounds, something is fundamentally broken. Customers are not just paying for missing features. They are paying high prices for platforms that cannot execute core functions without additional tools and integration costs.

Achraf Talhaoui at Qando Procurement Strategy & Solutions provided broader context. “Good post Antonio. Summarizes the commercial side of last 10 years of digital procurement. I have seen an entire generation of SaaS platforms, so not only Coupa but several of Gartner Magic Quadrant leaders and challengers in Suites as well as in best of breed solutions. So it is time to buy outcomes, not seats and promises.”

The shift he describes is critical. The traditional software licensing model charges per seat regardless of value delivered. Customers pay the same whether the platform transforms their operations or collects dust.

Goncalves confirmed that vallor.ai takes a different approach. “Absolutely, and we never charged for seats, our pricing has always been linked to outcomes from day 1.”

The Acquisition Shell Game

One of the most revealing Reddit comments explained Coupa’s AI strategy. “We were contracted with Llamasoft while they were acquired by Coupa, Llamasoft AI is where they get the AI from.”

This pattern repeats constantly in enterprise software. Company lacks capability. Company acquires company with capability. Company announces transformative AI features. Integration never happens or takes years. Customers wait.

Lari Numminen at generatemore.ai defended Coupa’s acquisition strategy. “All joking aside, Coupa have strong foundations in AI eg. through the Spend360 acquisition back in 2017. I wouldn’t dismiss their leverage and ability to bring in best-of-breed AI functionality through acquisitions.”

But acquisition announcements and working integrations are different things. Buying companies with AI does not automatically mean customers get usable AI.

The Marketing Problem

Yuri Hembitski captured the core issue. “Pricing for ‘coming soon’ features. Good one. Applies in many areas of professional lives.”

The practice is so common it barely registers as problematic anymore. Software companies routinely sell roadmap items as if they are current features. Sales teams demo capabilities that do not exist in production. Contracts get signed based on promises, not reality.

Humza Rafiq at JPJ Staffing stated the obvious. “Too many platforms are selling promises instead of performance. Real innovation is what you can use today not what they say is coming next.”

Rashad Miller was more blunt. “Lots of claims out there but no receipts.”

Asmaa Gad at Supply Chain AI Pro noted how difficult this makes vendor selection. “As someone still learning procurement tech, this helps me spot the difference between real capability and good marketing.”

The challenge is real. When every platform claims AI capabilities, how do buyers distinguish between working solutions and vaporware?

The Defensive Response

Not everyone agreed with the critique. Vinay K., who works in technology consulting, pushed back hard. “But how you are different from them? Coupa is staple in Enterprise and dont trust a post written by anonymous guy or may be PAID post. If you ever want to know what enterprise customers love about Coupa then reach to me, I will explain what I learned in 15 years in Enterprise SaaS space. Try to hire a great team, build great product, just by showing your competitor in poor light will not make your product good.”

The defensive tone is telling. Rather than addressing the substance of the critique, the response attacks the messenger and suggests the post might be paid criticism.

This is how established platforms deflect. Invoke their market position. Question the credibility of critics. Avoid discussing actual product capabilities.

The reality is that being a “staple in Enterprise” does not mean the product delivers on its promises. Inertia, switching costs, and integration complexity keep customers locked in long after the platform stops delivering value.

What Working AI Actually Means

Goncalves outlined what real AI should deliver. Agents that actually perform contract reviews and redlines. Actionable cost savings opportunities. Insights you can use right now. Not next quarter. Not next year. Now.

Pranav Dhoolia at useOven.com identified the fundamental principle. “This is a genuine pain point. Customers should get value today.”

Vitalii Honcharenko at Yael Acceptic reduced it to a simple mandate. “Pay for what works today, not what’s ‘coming soon.’”

Vincent Teyssier provided historical context that explains how the industry reached this point. “Guys, remember that P2P was never born from business need. It was a way for SAP and Oracle to decouple it from ERPs so that they don’t need to go as high as board approval in order to sell.”

The observation reveals that much of enterprise software architecture exists to serve vendor sales strategies, not customer needs. When platforms are designed to avoid triggering board-level scrutiny rather than solving business problems, customers end up with suboptimal solutions.

The Path Forward

The procurement tech industry needs a reckoning. Customers must stop accepting “coming soon” as an answer. They must demand working features before signing contracts. They must insist on outcome-based pricing rather than seat licenses.

Vendors who cannot deliver should lose business. Those who can should thrive.

Chris D. urged collective action. “Band together at their conferences, large and small companies, and press for more, better, faster.”

The power exists. Enterprise customers write massive checks. They have leverage if they choose to use it.

But that requires courage. It requires willingness to walk away from incumbent vendors. It requires acceptance that switching costs, while real, may be lower than the cost of paying premium prices for platforms that do not work.

The question is whether procurement leaders have the conviction to demand better. Or whether they will continue funding vendor R&D through their own budgets while waiting for features that may never arrive.

Real AI does not live in roadmaps. It lives in production. Everything else is just expensive vapor.

---

Join the conversation on procurement technology, AI implementation, and vendor accountability at Chain.NET, where supply chain professionals share real experiences with enterprise software, debate emerging solutions, and connect at events focused on practical innovation over marketing hype. The discussion happens here.