How to Build a Legacy Asset With Software (Not Just a Subscription)

There’s a strange thing about the digital life you’ve spent decades building. You don’t actually own most of it.

Your Gmail, your Notion, your Slack history, your Spotify playlists, your ChatGPT conversation history, your Google Drive, your Dropbox — every one of those is rented. The moment you stop paying, most of it becomes inaccessible or gets deleted on some retention timeline. If you die, your kids can’t inherit most of it. If the company gets acquired, the terms of service change. If you switch providers, you can migrate the raw files but not the workflows, the connections, the accumulated context.

We accept this. We barely think about it. It’s the shape of software as a service.

Sam Kececi, founder and CEO of The Sentience Company, is quietly running a company that’s structured differently. His product is a personal AI — a per-user model that captures how you think, decide, and communicate. And he’s built the corporate and legal structure around the product specifically so that users own the resulting model as an asset. Which means they can transfer it. Which means they can leave it to their kids.

The 67-year-old jeweler passing his mind to his daughter

Sam told me a story that made this concrete.

There’s a 67-year-old man named Pat, a jeweler in Newfoundland, who runs a 20-person high-end jewelry business. He uses Sentience every day. His employees consult a version of him before sending outreach — they’ll say “run it by my sentience first, make sure my sentience approves.” He uses it to check decisions, review deals, brief his team.

But the reason he uses it isn’t operational efficiency. It’s succession.

Pat is about to hand the business to his daughter. And what he told Sam is: “I actually have this building source of my knowledge now. If my daughter wants to ask — how do you think about a tough customer, how do you think about advertising — it will be able to pull from actual concrete examples.”

His daughter isn’t inheriting a spreadsheet, a customer list, and a set of physical assets. She’s inheriting decades of decisions. The taste her father developed. The specific way he thinks about pricing a stone, handling a difficult buyer, training a new hire. The stuff that used to live only in his head and would have died with him.

That’s a genuinely new category of thing to inherit. And it exists because Sam structured the company so it can.

Why software isn’t usually inheritable

There are two reasons software you use today can’t be inherited.

The first is legal. Most SaaS terms of service state that your account terminates on death, or that access is non-transferable, or both. The company owns the infrastructure and grants you a license to use it. The license ends when you do.

The second is practical. Even if you could legally transfer a Notion workspace to your child, what would they actually get? A pile of documents they can’t parse without the context you had. A history of conversations they don’t have the relationships for. Raw data that requires you, specifically, to make sense of it.

Both problems get worse for AI products. If your child inherits your ChatGPT account, they get access to a chat interface with your history — but the model itself is shared with millions of other users, so there’s nothing specifically yours to transfer. The AI hasn’t been shaped by you in any meaningful way. It’s a general tool that happened to have your name on the login.

Sentience is engineering around both problems simultaneously. On the legal side, they’ve structured the company so that the personal model each user develops is legally owned by that user as an asset — it can be transferred, willed, or held by a trust. On the practical side, the model actually is unique to the user (through per-user fine-tuning + a proprietary knowledge synthesis system Sam calls the “sleep phase”), so there’s something specifically yours to transfer.

The founder’s own reason: his grandparents

The origin story here is more personal than the pitch. Sam told me his grandparents passed away a few years ago. Both had been entrepreneurs — his grandfather started a real estate appraisal business, his grandmother a dance studio. And when they died, everything they knew — every hard lesson, every decision-making instinct, every story about how they’d handled an impossible customer — vanished into text messages, voicemails, and Sam’s own memory.

“The content of them that lives in my brain is pretty fallible,” he said. It’s a startlingly honest thing for a founder to admit on camera. And it’s the reason Sentience exists. The company is, at some level, a memorial that Sam built in the form of software. The features that let a jeweler pass his mind to his daughter are the same features that would have let Sam preserve more of his grandparents than a few voicemails.

He walked away from a $21 million company to build this. That’s a data point about how seriously the founder is taking the problem.

What this means for the rest of us

Most people won’t buy a personal AI product this year. But the idea Sam is testing — that software can be structured as an inheritable asset, and that the software you use should be structured that way — is worth sitting with.

Ask yourself: which parts of my digital life would I actually want my children to inherit?

Not my Gmail. Not my Slack history. Almost certainly not my Twitter account.

But my thinking? The way I make decisions? The specific arguments I find persuasive, the customers I’ve handled, the reasons I chose the paths I chose? That would be worth passing on. And it doesn’t exist in a form anyone can inherit today.

That’s the gap Sentience is trying to fill. And whether or not you buy their specific product, the question is a real one. The average person will accumulate 40-60 years of digital work over their lifetime, and almost none of it will outlive them. Something worth thinking about.

Especially if you have kids who might, one day, wish they could ask you how you thought about a tough decision — and find that the answer went with you.

---

Frequently Asked Questions

Can you leave software or AI to your kids in your will? Traditionally, no. Most SaaS terms of service state that accounts terminate on death and are non-transferable. However, some newer AI products — including The Sentience Company — are being structured so that the personal AI model a user develops is legally owned as an asset and can be transferred, willed, or held by a trust. This is a new category of digital inheritance.

What is a digital legacy asset? A digital legacy asset is a form of software or digital property structured so it can be transferred to heirs after death. Unlike traditional SaaS subscriptions, which typically terminate on death, digital legacy assets are legally owned by the user and can be inherited. Sentience is one of the first personal AI companies structuring its product this way, letting users pass a personalized AI model to their children.

How does inheriting a personal AI work in practice? The heir gets access to a trained model that reflects the original user’s decisions, communication style, and knowledge. In the Pat-the-jeweler example, his daughter will be able to ask his Sentience how he thought about specific business decisions — like handling a tough customer or pricing a difficult stone — and get answers based on actual patterns from his decades of work. The model isn’t the person, but it preserves specific decision-making context that would otherwise disappear.

What’s the difference between a personal AI model and a chatbot? A chatbot is a general-purpose interface to a shared model like ChatGPT or Claude. A personal AI model is trained specifically on one user’s data — their emails, calendar, notes, meetings, decisions — and outputs responses shaped by that individual. Sentience uses per-user LoRa fine-tuning on top of open-source base models, so each user’s model is genuinely unique to them.

Why isn’t most SaaS inheritable? Two reasons. First, legally, most SaaS terms of service prohibit transfer of accounts and terminate access on death. Second, practically, even if you could transfer a SaaS account, the raw data (documents, chat history) requires the original user’s context to be useful. Inheriting a Notion workspace or ChatGPT history rarely produces something the heir can actually use.

Can I already leave my ChatGPT or Claude account to my kids? No. OpenAI and Anthropic terms of service both prohibit transfer of accounts, and the underlying models are shared with all users, so there’s no personalized asset to transfer. Your chat history exists but the AI itself isn’t uniquely yours in any inheritable sense.

Why did the founder of Sentience walk away from a $21M company to build this? Sam Kececi lost both grandparents a few years ago and realized their accumulated wisdom — decades of running businesses, raising families, making decisions — lived only in a few voicemails, text messages, and his own memory. He wanted to build something that could preserve that kind of knowledge across generations. He left his prior company, Macro, where he was CTO and had raised $21 million, to start The Sentience Company in late 2025.

What is The Sentience Company’s approach to data ownership? Sentience is structured so that each user’s personal AI model is legally their asset. The company holds the infrastructure, but the model — trained on your data — is yours. It can be exported, transferred, willed, or held by a trust. This is unusual for a cloud-based AI product and is intended to make personal AI a genuine inheritance category, not just a subscription.

Who is using Sentience for legacy planning today? Early users include small business owners like Pat the jeweler (67, Newfoundland, 20-person high-end jewelry business, planning to pass the business to his daughter), founders preparing to hand off institutional knowledge, and executives building context archives their teams can access when they’re unavailable. The pattern isn’t broad consumer adoption yet — it’s high-context operators with specific succession planning needs.

What are the alternatives to a personal AI for legacy planning? Traditional alternatives include journals, memoirs, video interviews, and legacy documents like estate letters. These preserve narrative but not decision-making context. Personal AI models like Sentience are the first attempt to preserve how someone thinks in an interactive, queryable form. Whether the technology delivers on that promise is still being tested — Sam Kececi admits Phase 7 of his product’s master plan is “unknown” — but the category is emerging.