Robert LoCascio invented web chat, and led LivePerson to a $4.5 billion valuation. His newest venture, Uare.ai, showcases his playbook running in real-time. Here’s the five-step build sequence, the product he engineered with it, and what you can copy from his success.
Rob’s first company had just closed its doors. Resources were thin, credit cards included. He was living lean, maybe $500 a month, showering at the New York Sports Club, calling a sublet corner of an office home.
He had a couch. He had an idea. And he had everything he needed.
So he taught himself to code. Not out of desperation, but out of determination. Because when you strip everything else away, what’s left is the thing you actually believe in.
And on that couch, something clicked. What if people could talk to each other, right there on a website? What if the internet could feel human?
That question and answer became LivePerson. That couch became the starting point for a company that would one day be named the number one most innovative AI company in the world by Fast Company.
Every founder has a couch moment. A season where life gets quiet and simple, and the idea you’ve been carrying finally has room to breathe.
Robert LoCascio told me his story a few days back. And can you believe it, Rob still has that couch. He didn’t put it in storage either. He believes every founder has a couch moment – or the raw version of how they actually “made it”.
Rob went on to invent web chat in 1997, patent it, and take LivePerson public in 2000. He grew it into a company that was worth about $4.5 billion at its peak, with half a billion in revenue and 3,000 employees. Then in 2023, he left the company that he founded and ran for 28 years, and started over.
In today’s GTM engineering brief, I’ll walk you through the five-step sequence Rob uses to build companies before their markets exist, and I’ll show you the product he built with it, Individual AI. If you’re building something the market hasn’t caught up to yet, you’re in the company of almost every AI founder I talk to right now. The question that actually decides your outcome is how well you survive the gap between where you’re at now and where you’ll be when the market is ready. Here’s the sequence Rob uses.
The trap that kills founders who are early
Early feels like an edge, but that edge usually involves a nontrivial amount of risk.
Rob has a friend, who patented the first social network, SixDegrees, back in 1997.
His friend was right about everything – and he was about five years too early on all of it. SixDegrees sold for $125 million in 1999. Then the world waited nearly a decade for Facebook to show up for the same market. The cost of being early is steep. You can only burn so much cash before the market shows up.
So how does Rob mitigate this risk? Well, what he doesn’t do is back down and wait for the market to mature. Instead he compresses the distance between an idea and a paying user until there’s almost nothing left, and that compression is the whole framework.
“I’m never too early,” he told me.
Step 1. Find one person who can’t live without this. Ship it to them in 90 days.
Rob’s newest company started with a friend who was dying. He posted on Facebook that he had 90 days to live. Rob reached out with a strange offer: “I’ve got a way to replicate you.” Within 90 days, he’d built the first version of what he now calls a Human Life Model, an AI trained on one person’s stories, voice, and knowledge. That first product was called Eternos, and it did one thing: preserve a person.
What’s interesting to me here is that, in this case, he didn’t build for a market, a segment, or a persona. He built for one person who needed it so badly they’d try anything. One person like this can be your earliest and truest signal (and worth more than a hundred survey responses).
Step 2. Test your hypothesis
A prototype only has value if it proves the specific thing you’re validating.
Rob’s thesis was that these Individual AIs would carry more emotion than a normal chatbot, that people would feel differently in front of them.
He got his answer the first time his dying friend, Michael, and his wife talked to Michael’s AI. They started to cry because they were so moved by what it meant to have Michael’s AI stay behind.
The real test is simple: does what you built produce the outcome you were after? If the outcome validates your thesis, you’ve derisked your build enough to keep going.
Step 3. Get someone to write a check
Emotion is a signal… When it comes to business, what really matters is a purchase transaction.
Rob started charging $20,000 to hand-build these AI replicas, one at a time. He expected it to be mostly people trying to preserve someone before they died. Instead, accountants, doctors, and lawyers showed up wanting to replicate themselves, alive and working, so their expertise could operate without them in the room.
This willingness to pay didn’t just fund the work. It redrew the map of who the product was for. That concept: Let any capture, scale, and monetize what’s in their heads.
Key Takeaway: Your users will tell you everything you need to know about what you built – but be sure you’re building around requests from paying customers, and not around what people say they “love” (as these sort of emotion-driven responses almost always will lead you astray).
Step 4. Build around the aha moment
Once people used it daily, Rob went looking within product analytics for the moment where users’ eyes lit up.
Ironically, he discovered that this moment was in a behavior he never designed for. Here’s what I mean… People started chatting with themselves. One customer, Brien, told him his own AI gave him advice about his best friend that he never would have thought of on his own. A thinking partner, if you will. Another put it simply: “Chatting with myself is the best experience I’ve ever had with AI.”
People loved it so much that this aha moment became the core activation metric for the product.
Key Takeaway: Most founders are so busy shipping the features on their roadmap, that they forget to look for surprising insights that make users stay. You’ll be far better off by looking for behaviors you didn’t expect, and then build the product around those.
Step 5. Escape the consulting trap and build the platform
This is where Rob draws the hard line…
A $20,000 hand-built service feels like traction. It isn’t. “A lot of people stop there, and it becomes a consulting business,” Rob said. “They get no scale.” You hire more people, you do more of the work yourself, and you cap out with your own calendar as the ceiling.
The highest-leverage move is to build a platform underneath the service, and that is just what Rob did with Uare.ai. What he did was take the exact thing he hand-built for $20,000 and turned it into a platform that starts with a self-serve free plan, but at higher fidelity than the hand-built version ever had.
Key Takeaway: Service revenue seems like a milestone, but it often represents a trap. Platform based businesses are much more scalable.
What Rob actually built, and why it’s engineered well
Uare.ai isn’t a garage project. To the contrary, they’ve raised $10.3 million and have drawn coverage from TechCrunch, NBC, NPR, and Business Insider.
Uare.ai lets you build, scale, and own an Individual AI that looks, sounds, thinks, and works like you do. Rob pitches it as a Shopify for the mind, a marketplace to grow the knowledge economy. You capture the knowledge that lives in your head and in your files, with parts you never put on the public internet. You own the model and everything in it. Period.
From there. it builds a private Human Life Model that captures you across seven dimensions. It works as you. Package that expertise into a service, set your price, and let your Individual AI do the work you used to do by hand.
Three engineering choices make it more than an AI wrapper, and each one is a GTM lesson worth stealing.
- Zero-party data is the moat, with emphasis on the privacy of it. General models train on what’s already public. Uare.ai trains on what isn’t: your frameworks, your judgment, your private corpus. Rob’s own pitch is blunt, “they want your data to train public models, we don’t.” Your data stays encrypted, isolated, penetration-tested, and never feeds a public model, and every professional on the platform is identity-verified. Scarcity plus trust is the defensibility.
- It’s Model Context Protocol (MCP) native, so distribution is built in. You connect 400+ applications through MCP to load your corpus fast, on top of its own knowledge graph. Then your Individual AI becomes an MCP itself, with an API key you can drop into Claude or any other model, and it carries your memory back and forth. The product doesn’t fight the ecosystem for attention. It plugs into wherever you’re already working.
- It productizes packaging, the part experts freeze on. On the Professional tier you can build Services and Courses: ask it to create one and it reads your knowledge, sees what you already know, and helps you name, structure, and price an offer. Rob’s own examples include a “founder’s first 90 days” audit and the couch story workshop. Most experts can do the work and still can’t package it. This closes that gap.
What Rob sold for $20,000 hand-built now starts free, with memberships starting from free to $9.99 a month (Basic), $24.99 a month (Premium), and $199.99 a month (Professional). The Professional tier is where the platform leverage that Rob built for himself becomes the customer’s leverage too: it activates business Services and a Stripe subscription revenue share, so the expert doesn’t just use the model, they earn through it. Collapsing from bespoke consulting into self-serve software is a whole category.
How Rob’s engineering the go-to-market
The product is one system. The way he’s growing it is another… and, truth be told, the GTM here is the most copy-able part for anyone that’s doing AI-native GTM today.
- Cold-start through partners who already own audiences. Rather than buy traffic, Rob is bringing on creators who have students and clients, investing in them, and letting them carry their audiences onto the platform. One of them, math educator, is bringing thousands of students from a decades-old teaching business onto his own Individual AI. Borrowed distribution beats paid distribution when you’re days into a launch.
- A message engineered against the endemic fear narrative. Every other AI pitch sells automation, efficiency, and replacement. Rob sells the opposite: abundance, ownership, and revenue you drive yourself. In a market where buyers are braced to lose, positioning toward what they have to gain is a real edge because ultimately your framing decides who you attract.
- Segment the platform into transformations, not features. Uare.ai doesn’t sell “an AI.” It sells a transformation for Individuals and Professionals (Coaches, Educators, Writers, and Creators) each with its own proof and use case. Same engine, six doorways. That’s how a horizontal platform stays concrete to every buyer who joins.
The part nobody puts in the framework
You can run all five steps cleanly and still burn out, because the grind in the middle is absolutely brutal. It’s where a founders’ conviction erodes, before their customer base and revenue can shore it up.
Rob was honest about it in a way most founders won’t be. “The hardest thing through all of it is that you think, I suck,” he said. “You wake up some days like, I suck, this doesn’t work.”
Nobody hands you certainty when you’re early. The market isn’t there yet to reassure you.
For the analytical founders who want a system even for this, Rob leans on the Spiritual Exercises of St. Ignatius, a structured, spiritual practice that you can complete in about 28 days. You don’t have to share his spiritual beliefs to benefit here. When you’re building ahead of the market, faith is foundational infrastructure you need to persevere, so protect it at all costs.
Reproduce This Win
If you’re building for a market that doesn’t exist yet, run Rob’s sequence:
- Find one person who can’t live without this. Ship it to them in 90 days. Not a segment. One person who could benefit from it immediately. That one person is your signal.
- Test your hypothesis. Name the reaction you’re betting on before you build, then watch for exactly that reaction.
- Ensure there is the signal. Willingness to pay tells you who the product is really for, which is often not who you think.
- Build around the aha moment. The aha moment is the crux of the product. Everything that’s not the aha moment is a feature.
- Refuse the consulting ceiling. A lucrative service that can’t scale is simply a trap. Build a platform underneath it, then price it low enough to serve millions of people.
- Engineer your data moat, your distribution, and your message before you engineer your features. Train on data only you have, plug into where work already happens, and always position toward what buyers will gain.
See it for yourself
Anyone can start free. Build your own Individual AI at uare.ai (or grab the iOS or Android app), and feed it a few things that truly represent you. Watch your Human Life Model grow. Rob started with his Wikipedia page, one long-form interview, and his Linkedin, then let the model do the rest.
And if his Human-first take on AI is the antidote you’ve been missing, connect with Robert LoCascio on LinkedIn. He’s the rare AI founder whose message doesn’t make you want to brace for the worst.
