I asked Grok to model a venture called deathOS — an integrated platform for dying, death, and bereavement. The market it surfaced is real and underserved. The product it sketched to win that market is the same product every AI sketches when you ask it to design a platform.

The prompt that started it

The framing was deliberately broad. Take a fragmented, high-stakes, emotionally loaded domain — estate planning, digital legacy, end-of-life logistics, grief — and design a single integrated business around it. Name it deathOS. Tell me what the market looks like, what the product is, who buys it, and what the numbers are.

I picked death because it is the rare category where three things are simultaneously true. The buyer is motivated in a way no other consumer-software buyer is. The existing tooling is bad. And the buyer is in a cognitive state that makes them easy to fail.

I wanted to see what Grok would do with a domain where the right answer requires taste, not market-size arithmetic.

The shape of the conversation

Grok came back with a clean business overview. Target market in three segments — end-of-life planners (50+, tech-savvy), caregivers (35-65), the bereaved (all ages). Core products: a digital vault for documents and memories, AI-driven wills and estate tools, grief counseling and peer networks. Freemium at $10-50/month, $100-500 for legal services, referral fees on top.

It pulled a growth forecast for the adjacent digital-will services market: $2-3 billion today, 15-20% CAGR through 2030. The total US estate planning market is $18 billion. Online is still under 20% of legal services. The numbers hold up. The category is expanding.

Then it went sideways into the legal layer. Cross-chain digital wills as a patentable invention. Whether an MIT or GPL license matters more for a death platform. Trademarking the brand. The patchwork of electronic-wills law — Colorado and Florida allow electronic wills under UEWA, Germany still requires holographic or notarized, India doesn’t recognize electronic wills at all. AML/KYC if any asset transfer happens. GDPR’s deletion right against blockchain’s immutability.

The beats worth keeping:

  • The market is real and underserved. A $2-3B category growing 15-20% a year, with the dominant incumbents being LegalZoom and a handful of trust mills. None of them have built anything a grieving person would actually want to use.
  • The buyer is in a state most products can’t design for. Grief, cognitive load, and urgency are the buying conditions. A freemium onboarding funnel is not the right shape for this user.
  • The legal layer is genuinely hard. Electronic-wills law is state-by-state in the US, country-by-country in the EU, and functionally absent in India. Any death-tech product that touches actual asset transfer is a compliance operation with a UI on top.

The insight worth keeping

The opportunity is real. The framing Grok defaulted to — freemium SaaS, AI-driven legal tools, community support, referral fees — is the AI-template answer. It describes a product. It doesn’t describe the right product.

The right product in death-tech is not the one with the most features. It’s the one that absorbs the most friction from a person who is in no state to absorb friction themselves.

That means: not a dashboard. Not a vault they have to organize. Not a will wizard they have to walk through. The design constraint is that the user is a person whose parent just died, or who just got a diagnosis, or who is watching a spouse decline. They can fill out a form. They cannot make forty decisions.

Whoever wins this market is going to build the product that makes fewer decisions, not more. The estate planning industry is organized around forms because forms are how attorneys scale. A death-tech product organized around forms is just a cheaper attorney. The category that’s actually open is the one where the product does the work the attorney’s paralegal does — gathers the documents, identifies the accounts, files the notices, tracks the deadlines — and leaves the human to sign where a signature is legally required.

That’s the wedge. Not a vault. Not a chatbot. An executor’s operations layer.

What I’d push back on

Three places Grok’s analysis felt sanded-down or overstated.

The team slide was filler. CEO, CTO, Legal Advisor, future marketing lead. That slide is in every AI-generated business plan because every AI has read every business plan. The team that wins death-tech is not a grief psychology expert and a cloud security specialist. It is someone who has been an executor, someone who knows the probate code cold in three states, and an engineer who can build the document pipeline. The competency mix is operations-heavy, not research-heavy.

The financials were a fantasy. Year 1: $2M revenue, $1.5M expenses. Year 3: $10M revenue. Funding: $1M from VC/grants. These numbers are rounded until they lose meaning. A death-tech product with real legal-tooling ambitions burns well past $1M on compliance alone before it sees its first dollar of recurring revenue. The legal review cycle for a single state’s electronic-wills statute is a six-figure line item. Grok treated the regulatory burden as a risk to note, not as a cost line that shapes the whole business.

“Pending patents” as a moat is the wrong moat. Grok flagged AI personalization and pending patents as the edge. Patents on cross-chain digital wills will not stop a competitor who builds a better executor-ops product. The moat in death-tech is trust and distribution — funeral homes, hospice networks, estate attorneys, hospital social workers. Whoever owns that referral graph wins. No one is patenting their way out of that.

The market Grok surfaced is the part worth keeping. The product it sketched is the part worth throwing away. The space between those two is where an actual business would have to live.

Filed under /ventures. Part of the Grok series — conversations where the model did the work and I did the judgment.