Joe Jerome spends his days connecting AI to everything. His biggest challenge now is convincing people it isn't "just fancy Google."
Sajeel Qureshi has a habit of comparing billion-dollar technology shifts to pizza ordering, Mario Kart, and MTV and somehow it always makes perfect sense.
If you don't want to watch the UNwatchable Episode 2, then read the unreadable blog.
Why AI Is Becoming the New Interface for Business Software
TL;DR
- SaaS companies are treating AI like a bolt-on feature, the same mistake BlackBerry made with the iPhone.
- Customers are starting to open Claude before they open their CRM, accounting tool, or marketing platform. The AI is becoming the interface, not just an add-on inside the interface.
- Traditional software competes on dashboards and menus. AI competes on context, memory, and the ability to take action.
- Most businesses are still using AI as a glorified search box. The real advantage comes from connecting AI to real workflows and data.
- There are four levels of AI adoption: search, context, skills, and agents. Most companies are stuck at level one.
- The winners of the next decade will not be the SaaS companies with the most features. They will be the ones that become the infrastructure behind an AI-driven interface.
Introduction
Almost nobody believed Steve Jobs when he introduced the iPhone. People laughed at the idea of a phone with no physical keyboard. Executives at the time dismissed it outright. Today that reaction sounds almost impossible to imagine, but it happened, and it happened inside boardrooms full of smart, experienced people.
Many SaaS companies are making the exact same mistake right now with AI. They are treating it like another feature to bolt onto an existing product. Meanwhile, customers are beginning to treat AI as the interface itself, the place where work actually starts.
Every Industry Has Its BlackBerry Moment
This pattern is not new. Kodak invented digital photography and still lost to it. Blockbuster had the chance to buy Netflix and passed. BlackBerry owned the smartphone market and then refused to believe the market wanted a touchscreen.
"People at the corporate jobs at that time thought, 'iPhone? No one's going to use it because it's touch screen.'" - Sajeel Qureshi, CEO, Computan
The pattern is always the same. A dominant company keeps improving the thing that made it successful, while the market quietly moves on to something the company never took seriously. The question SaaS leaders should be asking is not "how do we add AI to our product?" It is "what happens if AI becomes the product interface, and our software becomes something else entirely?"
AI Is Replacing How We Use Software.
Instead of opening a CRM, an accounting platform, an email client, a marketing tool, and an analytics dashboard separately, people are increasingly opening one thing: Claude.
"Claude has become like the web browser." - Joe Jerome, Chief Revenue Office, Computan
That comparison is worth sitting with. The browser never made websites disappear. It became the gateway to them, the single layer people passed through to reach everything else. Claude, and tools like it, may be on the same path with business software. The individual platforms still exist and still hold the data, but the starting point for the work shifts to the AI layer.
SaaS Companies Are Still Competing on Features. AI Competes on Context.
Traditional SaaS has always sold dashboards, reports, menus, and buttons. More features, more tabs, more configuration options. AI does not compete on that axis. It competes on context, memory, conversation, and the ability to actually take action on someone's behalf.
Computan is a working example of this shift. Joe explained that their internal AI setup is connected to HubSpot, email, sales calls, proposals, accounting, and work delivery. Nothing is siloed.
"Our Claude has that entire workflow and it's fully educated. It misses nothing now." - Joe Jerome
The advantage here has nothing to do with the AI model being smarter than a competitor's model. It comes down to context. An AI system that can see the whole workflow will always outperform one that only sees a single disconnected piece of it.
Most Businesses Are Still Using AI Like Google
This is one of the more uncomfortable observations from the conversation, and it is worth sitting with.
"Most people are still using AI as a question and answer box... a glorified Google." - Sajeel Qureshi
Today, most users ask AI to summarize a document, write an email, or generate an image. That is level one thinking. Tomorrow's users will ask a fundamentally different set of questions: Which deals need my attention right now? Which customers are likely to churn this quarter? Prepare tomorrow's sales meeting. Build my proposal.
That is not a search relationship with AI. That is a working relationship, and it requires a completely different kind of setup behind the scenes.
The Four Levels of AI Adoption
One of the clearest frameworks from the conversation is a simple ladder that shows where most companies actually stand today.
"Level one is Googling... Level two is context... Level three is skills... Level four is a completely agentic system." - Joe Jerome
Level 1: Search
AI replaces a Google search. Quick answers, no memory, no connection to real business data.
Level 2: Context
Projects, memory, and connected documents. The AI starts to know something about the business instead of starting fresh every time.
Level 3: Skills
Reusable workflows and repeatable expertise. The AI can execute a known process instead of being re-explained every time.
Level 4: Agents
AI works independently. It advises, acts, and monitors without needing to be asked at every step.
Most businesses are still sitting at level one, using AI as a smarter search box, while the actual competitive advantage is waiting at levels three and four.
The Real Opportunity Isn't Building More Software
This is where the conversation moves from observation to strategy. Most businesses do not need another CRM, another dashboard, or another analytics tool. They need to connect the tools they already own so an AI layer can actually work across all of them.
"Software is fundamentally going to change." - Sajeel Qureshi
The opportunity is not in adding more software to the stack. It is in making the existing stack legible to an AI system that can act across it.
Why SaaS Companies Must Become AI Infrastructure Companies
Not every SaaS company is at risk. Products that own structured data, own workflows, own integrations, and own execution will not disappear. They become infrastructure. The interface on top of them becomes AI, but the platform underneath still has to hold the data and do the work.
Joe pointed to companies like Databox and Day AI as examples of platforms already moving toward this infrastructure-first model, positioning themselves to be the system of record and execution layer that an AI interface calls on, rather than the interface itself.
The Next Winners Won't Have the Most Features
BlackBerry did not lose because it stopped innovating. It lost because it kept innovating on the keyboard while the world stopped caring about keyboards. Today's SaaS companies are standing at a similar crossroads.
"They're like radio DJs in an MTV world." - Sajeel Qureshi
The companies that win the next phase will not simply bolt AI onto their existing product. They will rethink the product around AI, treating it as the primary interface while their software becomes the intelligence and infrastructure working behind the scenes.
Frequently Asked Questions
What does it mean for SaaS to be having its "BlackBerry moment"?
It means SaaS companies are at risk of repeating BlackBerry's mistake: continuing to improve a familiar interface (dashboards, menus, feature lists) while customers are moving toward a fundamentally different way of interacting with software, one led by AI as the primary interface.
Is AI actually replacing business software like CRMs and accounting platforms?
Not directly. The underlying platforms still hold the data and still do the work. What is changing is the entry point. Instead of opening each platform separately, people are starting to open an AI assistant first and letting it work across connected systems.
What are the four levels of AI adoption?
Level one is basic search style question and answer use. Level two adds context, memory, and connected documents. Level three introduces reusable skills and repeatable workflows. Level four is a fully agentic system that can advise, act, and monitor independently.
Why does context matter more than features for AI tools?
Because AI's usefulness scales with how much it knows about a specific business: its CRM data, its email history, its proposals, its workflows. An AI connected to the full picture will consistently outperform one that only sees an isolated piece of it, regardless of how many features the underlying software has.
What should SaaS companies do to avoid becoming obsolete?
Focus on owning structured data, workflows, integrations, and execution, and build toward becoming the infrastructure layer an AI interface relies on, rather than trying to remain the primary interface customers interact with directly.
.jpg)