SaaS-to-AI Transformation
The Transformation Reality
“SaaS-to-AI Transformation” describes the fundamental change traditional software companies must undergo to compete in the AI era. The defining characteristic: it’s brutal.
Des Traynor (Intercom co-founder) states it plainly:
“A lot of SaaS companies think they’re AI companies, but they’re not. They’re not true deep AI companies. If you’ve not gone through brutal transformation, you’re not there yet.”
What “Brutal” Looks Like
Intercom’s Example
- Founded 2011, grew to $50M ARR in 3 years
- Had an ML team since 2015-2016
- ChatGPT arrived November 2022
- Three years later: “entirely different company, unrecognizable”
The ML team’s reaction to GPT: “This is the moment. A one-way door for humanity. The iPhone moment.”
Filevine’s Example
Ryan Anderson (CEO) uses the Bridge Over River Kwai analogy:
“Don’t become like the general who built a beautiful bridge for the enemy and then committed treason to protect it from being destroyed. Your teams have built incredible things. Some of that code, those architectures, will have to be torn down.”
His framework: Plot systems on a 4x4 matrix (competitive advantage vs. slowing you down):
- Upper right (high advantage, doesn’t slow you): Fortify
- Lower left (no advantage, slows you): Tear down, even if emotionally hard
Why It’s Hard
1. Architecture Changes
You can’t “sprinkle AI on top”:
“ML engineers need to own their data layer and tune it daily. They can’t be going to traditional engineering asking to modify APIs.” — Ryan Anderson
The AI data layer must sit alongside the application layer, not on top of legacy infrastructure.
2. Mindset Shift: Content to Context
Traditional SaaS thinks about content (what’s stored). AI-native thinks about context (what AI needs to take action).
3. Competing with AI-Native Startups
Philip Lor (Personio CRO) feels the pressure:
“AI-native companies are going super fast. Clay was on stage. When you’re a SaaS company, I think it’s your job to become AI-first. I wake up every day thinking we need to go faster.”
4. Talent Acquisition
AI natives want to work where there’s rich data and distribution. You’re competing for them against pure-play AI companies.
The Test
How do you know if you’ve transformed?
Not transformed:
- Added a co-pilot feature
- Put “AI-powered” on your homepage
- Integrated with OpenAI API
Transformed:
- Company is “unrecognizable” from 3 years ago
- Core product fundamentally rebuilt
- AI owns its own data layer
- Revenue increasingly from AI features
- Leadership went through difficult decisions
The Competitive Advantage
For companies that make it through:
“The tie goes to the runner. Customers don’t want to leave their system of record, so you win in a tie against AI-only competitors.” — Ryan Anderson
Your data is the advantage - but only if you shift from thinking about content to context.
The Counter-Argument: Agents as Tool Amplifiers
Not everyone agrees SaaS faces existential risk. Nvidia CEO Jensen Huang argued on CNBC in February 2026 that the market got the SaaSpocalypse narrative wrong:
“These agentic AI will be intelligent software that uses these tools on our behalf and help us be more productive.”
His logic: agents are tool users, not tool replacements. They need systems of record (CRM, ERP, analytics) to store results and retrieve structured data. Nvidia itself is deploying “hundreds of thousands of digital employees” alongside 42,000 humans — and software tool consumption is increasing, not decreasing. Companies like ServiceNow and SAP, Huang argues, will build agents fine-tuned for their own platforms.
The practical resolution: the transformation is real, but it’s selective. Simple automation software faces displacement. Deep systems of record become more valuable as agent adoption drives higher tool consumption.
Timeline
- 2022-2023: ChatGPT catalyst, companies realize the shift
- 2024: Brutal transformation period, early winners emerge
- 2025-2026: Clear separation between transformed and left behind; Jensen Huang counter-narrative emerges
Related Reading
- Des Traynor - Leading voice on transformation reality
- Enterprise AI - The business context
- AI Agents - What transformed companies ship
Expert Mentions
Des Traynor
If you're a SaaS company who thinks you're an AI company and you've not gone through brutal transformation, you're not there yet.
Ryan Anderson
Nothing is sacred. Bridge Over River Kwai - don't become the general who built a beautiful bridge for the enemy and then committed treason to protect it.
Philip Lor
AI-native companies are going super fast. When you're a SaaS company, it's your job to become AI-first.
Sebastian Siemiatkowski
The next thing that's going to hit everyone bad is the switching cost of data. How do I get all of my data from the existing vendor and move it to the new vendor with the help of AI through one click? That brings down switching cost and that's when the real threat to SaaS comes.
Ali Akhtar
Legacy enablement tools get less than 50% adoption. We see customers with close to 100% adoption. Letter AI rebuilt sales enablement AI-native from scratch after pivoting from Tractatus during YC.
Jerry Murdock
The idea of 'I can just bolt on AI to my company' — possibly true, you can maybe get an exit, but being AI native is going to make you a better company. If you're not making your software for autonomous agents today, you're going to be severely challenged.
Jensen Huang
I think the markets got it wrong. These agentic AI will be intelligent software that uses these tools on our behalf and help us be more productive.
Alex Rampell
Three types of SaaS: seats tied to outcomes (at risk), seats as pricing trick (safe), middle ground (nuanced). Public markets can't tell the difference.
Mike Cannon-Brookes
The filing cabinet can do work. Not every SaaS company will thrive, but for Atlassian this is the best thing that's happened to the business — knowledge tools that can act on knowledge.
Steve Yegge
Zendesk example: AI-native companies don't want the UI, they want the API. Platforms that don't expose APIs for agents will be 'producted out of existence.' Innovation shifts to 2-20 person startups.