Bereaved families inherit fragments.
Voice notes, chats, photos, documents, and stories are scattered across devices when families most need continuity, comfort, and context.
Investor preview | Germany-first | Working MVP
Digital Afterlife
Consent-aware AI memorial avatars that help families preserve a loved one's voice, stories, and personality, turning bereavement into guided, dignified remembrance.
Memorialization, not resurrection
Not a therapy replacement
Persistent AI disclosure
A private memorial avatar.
Families can chat, speak, and see a lip-synced avatar respond from real memories while clearly labeled as an AI simulation.
Anticipatory grief first.
Families can capture a richer, consented archive while the person is alive, creating better data, cleaner rights, and partner distribution.
B2C plus B2B2C.
Subscription for families, introduced through funeral homes, hospices, and insurers. Germany first.
Why Now
Three forces make ethical AI memorialization investable now.
Multimodal AI
Memory-grounded conversation, voice generation, speech-to-text, and lip-synced video now fit inside one product, feasible only in the last ~18 months.
EU AI Act, Article 50
Since August 2026, AI and synthetic-media disclosure is mandatory. Built-in disclosure and consent are the entry ticket.
Care networks
Funeral homes and insurers already buy digital bereavement products: grievy runs in 100+ German funeral homes, and Willow sells memory books through Vienna funeral homes.
Investor appetite is proven
Empathy Series B (2024): bereavement support, B2B2C, with strategic insurer participation.
TechCrunch sourceSolace Care pre-seed (August 2026): legacy planning sold through insurers.
FinTech Global sourceWillow Stories pre-seed (2025): digital memory books via funeral homes.
AIN source
Product Experience
Guided memory capture, grounded conversation, and controlled interaction.
The product is not trying to convince anyone a person is alive. It is a private, consent-aware memorial layer for preserving a loved one's words, voice, stories, and personality cues.
01
Capture: collect portrait, voice samples, chat exports, documents, and memories.
02
Construct: create a per-avatar memory store, personality summary, and voice reference.
03
Reflect: interact through text, push-to-talk voice, or lip-synced video.
04
Control: export data, rebuild personality, regenerate media, or delete the avatar.
Ethics And Safety
Safety is the category entry requirement.
Grief-tech fails if it feels manipulative, opaque, or clinically overreaching. Digital Afterlife is positioned as memorialization and guided reflection.
Consent-first creation
Prioritize living consent and explicit voice/image rights. Family-created avatars require stricter authorization and narrower controls.
Persistent AI labeling
The avatar identifies as an AI simulation across text, voice, and video, not as a living person or consciousness transfer.
Mental health boundaries
No grief-therapy claims. Crisis language should trigger support guidance, escalation, and careful limits on interaction.
Data control
Per-avatar isolation, export, deletion, source-material control, and clear handling of voice, image, and personality data.
Usage limits
Caps and friction points help reduce over-reliance while supporting healthier remembrance patterns.
Partner review
Hospices, funeral homes, counselors, and privacy reviewers can help shape acceptable workflows before scale.
Prototype Status
Functional MVP across text, voice, and animated video.
The product has been validated technically. The next milestone is validation with real families, partner channels, safety reviewers, and paid cohorts.
Memory-grounded chat
Responses are grounded in the person's real messages and documents through retrieval, rather than generic chatbot output.
Voice interaction
Push-to-talk flow with speech transcription, in-persona response, and multilingual speech playback.
Talking avatar
Lip-synced animated video streamed from a GPU-backed service. Positioned as premium or metered because of cost.
GDPR-oriented controls
Users can export all data, delete avatars, and manage generated personality/video artifacts.
Live Demo
See the working MVP in a live session.
Text, voice, and lip-synced video are running end to end today. We walk investors through a real avatar session in about 15 minutes.
Technical Feasibility
Built on production-ready components with a clear cost watch-item.
Backend and persona
Cloud-hosted API, retrieval-grounded persona engine, vector memory store, and per-avatar data isolation.
Voice and video
Speech-to-text, multilingual voice generation, lip-sync animation, GPU-backed video processing, and real-time streaming.
Mobile experience
Cross-platform iOS and Android app with Google and Apple sign-in. The app builds and runs on device.
Economics
Text and voice are the low-cost retention layer. Animated video is the premium tier or metered add-on.
Go-To-Market
Germany-first consumer AI grief-tech with B2B2C distribution.
Germany is the first market because the annual bereavement need is large and trust-sensitive distribution can be built through local end-of-life partners.
Initial wedge
Anticipatory grief families working with hospices, legacy-planning providers, and care networks.
Consumer revenue
Subscription for text and voice interaction with clear usage limits and premium media features.
Partner channel
Funeral homes, hospices, palliative teams, and insurers introduce families to a consented memory-capture workflow.
Partner bundle
Co-branded memory capture for pre-need and post-loss workflows. Premium talking video is priced to GPU cost.
Market
A recurring need: about 1 million deaths a year in Germany, 4.8 million across the EU.
deaths in Germany in 2025. Each one leaves a family deciding how to remember.
Destatis sourcecancer deaths in Germany in 2025: about 23% of deaths follow a known diagnosis, with time to capture voice and stories with the person's own consent.
Destatis sourcefuneral firms in Germany, plus ~1,500 outpatient hospice services, ~270 hospices, ~340 palliative units, and 400+ specialist palliative teams.
DHPV source1.01M deaths
77.8k families: cancer deaths x 34% who find AI memorials acceptable
4.81M deaths (Eurostat 2025)
238k deaths (Eurostat 2025)
Bottom-up sizing: one family per death at a hypothesis price of EUR 120 per family per year, which we test in pilots.
Competition
No one in the EU sells a consented, talking memorial avatar yet.
Our October 2026 scan of public sources: the US players that sold direct have shut down, pivoted, or met public backlash. EU players sell memory books, memorial pages, and grief support through funeral homes and insurers. They are potential partners, not threats.
Consent-first
Built with the living person, not from leftovers: the answer to the backlash after-death apps get.
Text, voice, and video
Generated in one app, not pre-recorded clips or a book.
Sold where the moment is
Through funeral homes, hospices, and insurers, not ads.
Built for EU rules
AI disclosure (AI Act Article 50), GDPR, and local language.
Validation
The risk isn't the tech. It's adoption, so we test it through partners first.
of Germans find an AI recreation of a deceased person acceptable vs. reject it. A split market rewards consent and restraint.
Kaspersky survey, 2024German funeral homes already use grievy's digital grief support: the channel buys digital bereavement products.
Startbase source, 2025EU companies selling a consented, interactive memorial avatar today, from our October 2026 scan of public sources.
What we test next
Investor Case
Raising EUR 750k for 18 months of Germany-first validation.
The round is designed to prove partner distribution, paid family demand, emotional safety, consent and rights workflows, and video unit economics before broader EU scale.
Use of funds
- 35% product and engineering: onboarding, memory quality, mobile UX, export/delete hardening.
- 25% Germany pilots and go-to-market with funeral home, hospice, care, and insurance partners.
- 15% AI and video infrastructure: GPU cost benchmarks and premium video packaging.
- 15% privacy, safety, legal, and rights: consent, voice/image rights, GDPR, emotional safety.
- 10% operations and contingency.
- Target outcome: 5-10 pilot partners and 100-250 paid family beta users.
Team
Technical founders who ship, and who already pivot fast.
Mihai-Alin Wolff, co-founder, operations and go-to-market. Leads operations, project management, and Germany-first go-to-market: partner distribution and the customer-facing motion. LinkedIn
Radu Georoceanu, co-founder, engineering. Software-engineering lead behind the multimodal platform. Owns architecture, the persona and retrieval engine, and technical diligence. LinkedIn
Co-founders of Micronest, a SaaS venture. Previously built DestLens, an AI chat-based travel assistant, and pivoted decisively.
Investor FAQ
The hard questions should be answered before they are asked.
Is this creepy?
It can be if framed badly. The product deliberately avoids resurrection language and centers AI disclosure, consent, deletion, usage limits, and emotional-safety boundaries.
Is this grief therapy?
No. It is memorialization and guided reflection. Clinical claims should only come after proper validation, oversight, and regulation.
What is defensible?
The workflow: consented capture, memory grounding, voice/video orchestration, privacy architecture, safety controls, and partner distribution.
What is the biggest risk?
Trust and emotional safety first; video unit economics second. Both are visible early and can be tested before scaling.
How is consent handled?
The first wedge is anticipatory grief, where living consent is clearest. Family-created avatars need stricter authorization, voice/image rights checks, and takedown controls.
Can users delete or export data?
Data control is core to the product: users should be able to export source material and generated records, delete avatars, and understand retention before public pilots.