SaaS Strategy · AI Era · ISV Survival

The Disruptor
Gets Disrupted.

Sethunath U N
Sethunath U N
Chief Consultant & Advisor, Mavis Dx
May 2026  ·  5 min read
SaaS & AI ISV Strategy Digital Disruption Foundation First

I find myself thinking about this often. How drastically the SaaS business world has changed — and how quickly what once seemed like permanent moats can become liabilities.

Think about where SaaS came from. It wasn't born in boardrooms or strategy decks. It was born from frustration. From real people standing in real places, hitting real problems, and deciding something had to be built.

Someone standing for too long in a check-in or baggage queue at an airport decided there had to be a better way. That frustration became the passenger service system that runs most of global aviation today.
🏥
Someone spent a full day in a hospital — papers in hand, walking between departments, repeating the same information over and over — and saw the burning need clearly. That experience became integrated hospital information systems that transformed clinical care.
🚌
Someone standing with uncertainty at a crowded bus station — no information, no certainty, no comfort — conceptualised an application that gave travellers convenience, confidence, and control over their journey.

Deep domain expertise, hand-in-hand with technology, data, and process — that combination eased problems that had frustrated people for decades. The world was disrupted. Human life genuinely improved. And the companies that built those solutions thrived for years, sometimes decades, on the strength of what they knew and what they built.

The world was disrupted by B2B and B2C applications. Now those same disruptors are facing their own disruption.

— Sethunath U N

And Then, AI Arrived

A new disruptor. One that doesn't wait for frustrated people in queues — it works at computational speed, at scale, and it's getting cheaper by the month.

I hear the scepticism. I share some of it. High training costs, the real need for expert human validation, the gap between AI demos and AI in production — these are genuine concerns, not excuses. The cost of running AI responsibly still often outpaces the gains.

But here is the part that keeps me honest about the trajectory:

280×
GPT-4 level intelligence costs approximately 280 times less today than it did in 2022. That cost curve is not slowing down. What costs a rupee today cost hundreds two years ago. The economics of AI are moving faster than most business strategies can absorb.

So while the sceptics are right about today, the optimists may be right about tomorrow — and the window between those two positions is shorter than most ISVs realise.

The Real Risks for ISVs — In the Next 2 to 5 Years

This is not a distant theoretical threat. These are risks that are active right now, for companies operating in the same domains I have spent 30 years in.

Commoditisation of the Application Layer

The UI and workflow logic that took years to build stop being defensible. AI can replicate them. The product stops being the moat.

Hyperscaler Encroachment

AWS, Azure, and Google are embedding domain AI directly into their platforms. ISVs that sit between the hyperscaler and the customer are being squeezed from below.

🏢

Customer AI Maturity

Larger enterprises are building internal AI platforms that absorb functions previously outsourced to ISVs. The customer is becoming the competitor.

📈

Margin Compression

Even ISVs that survive will face customers expecting deep customisation without proportional price increases. AI has changed the benchmark for what "standard" looks like.

Who Survives — and How

The ownership pendulum is moving toward customers. But not uniformly and not completely. The ISVs that will thrive are not necessarily the biggest or the oldest. They are the ones that make three deliberate shifts:

The Three Shifts That Separate Survivors from Casualties

1

Own irreplaceable domain intelligence

Not just domain software. The code can be replicated. The accumulated knowledge of how an airline actually operates at 3am during a disruption cannot.

2

Operate as trusted AI operators

Not just application providers. The value shifts from what you built to how reliably and safely you run it, with AI in the loop.

3

Shift the moat from code to data + expertise + compliance

Proprietary training data, domain models, and operational trust are what AI cannot simply generate for a customer who goes elsewhere.

But Here Is the Part Most People Miss

If AI is expensive to run and requires expert validation to be safe — and it does, right now — then the companies that will actually extract value from it are not the ones who rush to adopt it. They are the ones who are already built to use it well.

⚙ Clear Processes

AI needs structured inputs. Without defined processes, you feed chaos into the model and get confident chaos back.

📊 Data Discipline

AI models are only as good as the data they work with. Clean, governed, well-labelled data is not an IT project — it is a business prerequisite.

🌟 Domain Expertise In-House

Validation requires people who actually know when AI is wrong. Without domain experts, you cannot catch the confident mistakes before they become costly ones.

💵 Financial Rigour

You cannot measure AI ROI without a baseline. Cost-to-serve models, pricing discipline, and FinOps governance are how you know whether AI is actually paying off.

A company without these foundations will spend heavily on AI, get inconsistent outputs, validate them poorly, and make expensive mistakes. A company with these foundations will use AI as a genuine multiplier — compressing months of work into weeks, sharpening decisions, and accelerating execution without sacrificing accuracy.

AI on a broken foundation doesn't fix the foundation. It automates the chaos faster — at scale.

— Sethunath U N, Mavis Dx

That Is Exactly What Mavis Dx Builds

Our four service pillars — GTM clarity, cloud economics, operational maturity, security compliance — are not just good practice for today. They are precisely the foundations that make AI adoption viable rather than risky, profitable rather than expensive, and strategic rather than reactive.

The person who stood in that airport queue didn't just have an idea. They had the domain knowledge to build something that actually worked at scale. That combination — deep expertise meeting the right technology at the right moment — is what created lasting value then.

It is still what creates lasting value now. The technology has changed. That truth hasn't.

Is your foundation ready for the AI era?

If you are a mid-market ISV navigating these shifts, let's have an honest conversation. Not about AI tools — about whether your fundamentals are strong enough to make them work.

Outcome-based · No-commit · Monthly engagement
You continue only when you see real value. That is our promise.


Start the Conversation → mavisdx.com