Two Suspended Students Just Exposed Everything Wrong with PropTech
The Cluely Playbook: How to Build PropTech That Users Love
Today's Brief:
When Columbia Dropouts Accidentally Built a $120M Company
Why Real-Time Beats Everything Else
The Open Secret of Product-Market Fit
How 22 People Generated $7M ARR
The Enterprise Discovery That Changed Everything
What This Means for Property Technology
The Bottom Line
When Columbia Dropouts Accidentally Built a $120M Company
The most amazing business stories often begin with failure. In early 2025, Roy Lee and Neel Shanmugam were suspended from Columbia for creating Interview Coder, an AI tool designed to help software engineers cheat on technical interviews. Most students would have apologised, deleted the code, and moved on. Lee and Shanmugam did something different: they raised $5.3 million from Abstract Ventures and Susa Ventures.
Six weeks later, Andreessen Horowitz led a $15 million Series A, valuing the company at $120 million. The product had evolved from Interview Coder into Cluely - an invisible AI overlay that provides real-time assistance during video calls. Through translucent rendering that bypasses screen sharing detection, it operates undetected on Zoom, Teams, and Google Meet.
The technical architecture is deceptively sophisticated. Real-time speech-to-text captures conversation audio while OCR analyses on-screen content. The system then uses GPT-4 and Claude to generate contextually relevant suggestions, all with sub-second latency. Users access the AI through a simple keyboard shortcut (CMD+Enter), receiving instant guidance during high-stakes conversations.
What started as a tool for technical interviews quickly found broader applications. Sales teams use it for objection handling. Consultants pull data during client calls. Job seekers gain confidence during interviews. The controversial "cheat at everything" positioning guaranteed media coverage, but the underlying innovation was genuine: they'd solved the latency problem that had plagued real-time AI assistance.
You’re probably thinking, Zakee, this is a real estate newsletter, what’s the relevance?
Well, stories like this are gold for proptech. Cluely demonstrates how the most successful products often emerge from non-traditional founders solving personal problems. The property technology landscape remains dominated by companies founded decades ago by industry insiders. Perhaps that's precisely the problem.
Why Real-Time Beats Everything Else
Every meeting tool before Cluely operated on the same assumption: help users after the conversation ends. Transcription services, meeting summaries, action item extraction - all valuable, all retrospective. Cluely asked a different question: what if users needed help during the conversation?
The technical breakthrough wasn't just speed but contextual awareness. The system analyses both audio conversation and screen content simultaneously, understanding not just what's being said but what's being shown. When a sales prospect asks about a specific feature while looking at a pricing page, Cluely knows to provide pricing-specific responses rather than generic product information.
This real-time capability fundamentally changes user behaviour. Early users reported that despite initial 5-10 second delays, the ability to get answers while speaking transformed their confidence in meetings. The company has since reduced latency to sub-second levels, making the assistance feel truly instantaneous.
"Every single person who has a meeting or an interview is testing this out"
Roy Lee, after Cluely's enterprise launch
Now. Let’s think about how real estate professionals access information today. A broker on a property tour pulls up comps on their phone while walking through units. An asset manager in a board meeting scrambles through spreadsheets to answer ROI questions. A developer in a planning meeting searches through zoning documents on a second screen. Each context switch breaks conversation flow and signals unpreparedness.
Each context switch breaks conversation flow and signals unpreparedness.
The gap between real-time expectations and real estate software reality has never been wider. Clients accustomed to instant information from every other industry encounter real estate professionals still fumbling through PDFs and spreadsheets. Cluely demonstrates that users will pay substantial premiums for immediacy - their $20/month Pro tier costs more than many real estate CRMs, yet users pay it gladly for the confidence it provides.
The lesson extends beyond speed to presence. Real-time assistance means being present in the user's workflow, not requiring them to leave their conversation to find information. Real estate software that forces professionals to pause client interactions to look up data has already failed the modern user.
The Open Secret of Product-Market Fit
Cluely's founders didn't conduct market research. They didn't interview potential customers or build personas. They built a tool to solve their own immediate problem: passing technical interviews after suspension made traditional job searching impossible. The initial product was exactly what they needed, nothing more.
This approach contradicts conventional product development wisdom, yet it consistently produces breakthrough products.
The founders' personal experience with the problem meant they understood nuances no amount of user research would have revealed. They knew the anxiety of waiting for an interviewer's question. They understood the cognitive load of simultaneously listening, thinking, and responding. They felt the frustration of remembering the perfect answer after the call ended.
As usage expanded beyond technical interviews, patterns emerged. Sales teams discovered the tool helped with product knowledge during demos. Consultants used it to access data during client presentations. Each use case shared the same core need: contextual information delivered without interrupting conversation flow.
In proptech, I’m currently shocked that I see startups emerging this year that are building dashboards, replicating past businesses, and essentially not tackling any real problems that real estate professionals have.
The winning model holds for major software companies. Stripe's founders were frustrated by payment processing complexity. Airbnb's creators needed to pay rent. Uber's founder couldn't get a cab in Paris. Personal pain points scale better than surveyed preferences because the founders maintain an intuitive understanding of the problem even as the solution evolves.
Real estate needs founders who've personally struggled with the inefficiencies.
This point is so important.
Founders who've lost deals because they couldn't access information fast enough. Who've watched clients' eyes glaze over during data-heavy presentations. Who understand that the difference between closing and losing often comes down to confidence and speed.
How 22 People Generated $7M ARR
Cluely reached $7 million in annual recurring revenue with just 22 employees. The math is striking: $318,000 in revenue per employee, compared to typical SaaS companies at $150,000-200,000. This efficiency didn't happen through superhuman productivity but through fundamental rethinking of how software companies operate.
First, distribution. Rather than building a sales team, Cluely released an open-source version that gained 850+ GitHub stars. Developers could examine the code, understand the technology, and build trust before purchasing. The controversial positioning generated "billions of impressions across social media" without paid advertising.
Second, product simplicity. Cluely does one thing - real-time AI assistance - exceptionally well. No feature bloat, no complex configurations, no lengthy implementations. Users download the application, sign in, and start using it immediately. This simplicity extends to pricing: three transparent tiers, no "contact sales" buttons, no hidden fees.
Third, technical leverage. Modern AI infrastructure means a small team can deliver capabilities that would have required hundreds of engineers a decade ago. Cluely doesn't maintain its own speech recognition or language models; it orchestrates existing services into a novel user experience.
Compare this to traditional real estate technology companies. CoStar employs thousands. CBRE's technology division has hundreds of engineers. These companies maintain complex databases, support legacy integrations, and manage enterprise sales processes. The overhead of being comprehensive crushes the ability to be excellent at anything specific.
Over the next 12 months, I believe we’re about to see a massive shake up in proptech - in terms of innovation, depth of the application layer, and consolidation.
The Enterprise Discovery That Changed Everything
Cluely began with consumer subscriptions at $20 per month. The founders understood individual users, priced accordingly, and built features for that audience. Then something unexpected happened: enterprises started calling.
The shift was dramatic. ARR doubled from $3 million to $7 million in one week after launching enterprise features. What enterprises wanted wasn't fundamentally different from consumer needs - they wanted the same real-time assistance, just deployed across teams with appropriate governance.
"The enterprise opportunity emerged from individual users advocating internally, not from top-down sales processes"
Building enterprise features took days, not months. Single sign-on, team management, analytics dashboards, compliance tools - the basics that every enterprise requires. But because the core product was already solid, these additions were straightforward extensions rather than fundamental rebuilds.
The enterprise pricing model remained simple: transparent multipliers of individual seat costs rather than opaque negotiated contracts. This approach seems obvious, yet most real estate software companies still refuse to publish pricing, forcing potential customers through lengthy sales processes to learn basic information.
I’ve personally been through this process and it’s painful - I was looking for a new PMS a couple of years back. It took hours of calls and back and forth before we even had an idea of the pricing, and immediately stopped discussions as it was double our budget. Transparency is important.
Real estate technology could learn from this progression. Start with individual agents, analysts, or asset managers. Make them successful. Let them advocate internally. Then add the enterprise features organizations require without compromising the core user experience.
Instead, proptech typically approaches from the opposite direction: build for enterprise requirements first, then wonder why individual users resist adoption. The enterprise sale might sign the contract, but individual users determine whether software actually gets used.
What This Means for Real Estate Technology
Real estate technology platforms average over a decade old, architected for desktop computers and patient users. The standard workflow remains: gather requirements, search databases, export data, analyse in Excel, create presentations, share via email. This sequential processing mindset is in every aspect of real estate software.
Cluely demonstrates what users now expect:
Instant answers
Invisible complexity
Hidden UI
The disruption pattern is clear. New entrants will unbundle specific workflows from comprehensive platforms, delivering superior experiences for narrow use cases.
Consider investment committee presentations. Currently, analysts juggle between Argus models, CoStar reports, market research PDFs, and PowerPoint. An AI-native tool could provide real-time support during IC meetings, instantly pulling relevant metrics, generating sensitivity analyses, and flagging concerns based on the committee's questions. The entire process could be conversational rather than presentational.
Or broker-client interactions. Instead of "I'll get back to you on that" becoming the default response to detailed questions, imagine AI that knows every comp, every zoning restriction, every market trend, accessible instantly during property tours. Not a database to search, but intelligence that anticipates what information matters in each moment.
The incumbents face an impossible position. They can't rebuild from scratch without abandoning existing customers. They can't match new entrants' development speed while maintaining legacy systems. They can't simplify without removing features some customers depend on. Their comprehensiveness has become their weakness.
Meanwhile, real estate professionals under 35 enter the industry expecting software to work like consumer applications they use daily. They won't tolerate searching through multiple systems for basic information. They expect AI to handle routine tasks. They assume software will work seamlessly on their phones.
Transaction management platforms that require desktop access will lose to mobile-first solutions. Valuation software that takes hours to run scenarios will lose to real-time alternatives. CRMs that require manual data entry will lose to systems that capture information automatically.
The next wave of real estate technology won't be comprehensive platforms but focused tools that excel at specific jobs. They'll be built by small teams using modern infrastructure, distributed through product-led growth rather than enterprise sales, and priced transparently for immediate adoption.
The Bottom Line
Cluely built a $120 million valuation in months, not decades. Their path - from university suspension to Andreessen Horowitz funding - seems unlikely, even absurd. Yet the fundamentals are reproducible: solve a real problem, ship fast, embrace controversy when it serves distribution.
The proptech market presents similar dynamics to pre-disruption interview software. Slow incumbents, changing demographics, and new technical capabilities create opportunity for focused entrants. The winners won't be teams trying to rebuild CoStar or Yardi. They'll be small groups targeting specific workflows with modern solutions.
The playbook is clear: identify a moment where real estate professionals need information but can't access it quickly. Build narrowly for that specific use case. Distribute through individual users who become internal champions. Add enterprise features only after proving individual value. Price transparently. Ship constantly.
Real estate transactions involve hundreds of these information gaps. Each represents an opportunity for focused innovation. The commercial broker who can't instantly access lease comps during negotiations. The asset manager who can't model scenarios during investment committees. The developer who can't visualise zoning implications during site walks.
What's missing isn't technology - AI, cloud infrastructure, and mobile platforms make sophisticated solutions accessible to small teams. What's missing is founders who understand both the problems and the possibilities. Who've felt the frustration of current real estate software and refuse to accept it as inevitable.
The suspended Columbia students who built Cluely had no business creating enterprise software. Perhaps that's exactly why they succeeded. Real estate technology's next chapter won't be written by insiders optimising existing paradigms but by outsiders asking entirely different questions.
Thanks for reading. See you next week.
My startup, Buildable, takes a lot of lessons from this playbook.
Real estate professionals struggle with comparing properties with like-for-like assets. Incentivised buyers manipulate numbers, downloading data from dashboards and analysis them takes hours, and frankly - it’s manual work that should be off-loaded.
I lived this pain whilst working as a real estate developer-operator in the UK.
So, I went all in and built the solution. Narrow use case. Focused to professionals in Dubai to begin with.
Looking forward to sharing more on this as we set to launch asap this month.