LumenUs Plans to Use AI for Paperwork After a Death

LumenUs, a grief-care platform founded in 2025 by Sara Tashakorinia, will launch in early access in October 2026 with AI intended to take over the administrative work that follows a death. TechCrunch

The load is large. Settling a loved one's affairs takes an average of about 570 hours of paperwork, and it lands when families have the least capacity to do it. LumenUs is starting with practical help, including finding tax forms, filing insurance claims and shutting down bank accounts.

Tashakorinia, the company's chief executive, started LumenUs from personal experience. Her husband died in a car accident eight years before the October 2026 report. She told TechCrunch that bereaved people receive an honorary badge of a project manager for a project they had no idea about, and said the company is using AI to automate as much as possible to save time and give it back to users.

The first version is narrow on purpose. LumenUs will focus on people in the immediate aftermath of a death, not the full range from anticipatory grief before a loss to long-term grief afterward. Onboarding asks detailed questions about legal and financial resources and state of mind to tailor the experience, then sorts tasks by urgency to separate what needs action now from what can wait.

That sorting matters because delay is not the only risk. Billions of dollars in life insurance benefits go unclaimed because people do not know they exist. Finding those benefits is part of the core job, not an extra.

Tashakorinia comes from cybersecurity. Her co-founders are Sajad Mirzaei, a former engineer at Google and Box, and Larry Keeley. While testing the product, the company is working with clinicians and social workers. LumenUs is a participant in Startup Battlefield 200, TechCrunch's annual showcase of early-stage startups.

The company operates at lumenus.life and calls itself The Grief Intelligence Platform. It says it organizes the paperwork, accounts, benefits and logistics created by a loss, supports families from a serious prognosis through the long arc of loss, and offers its foundation free for every family. It describes itself as founded from lived loss.

The broader context here will be familiar to readers who follow automation. A death sets off a scattered workflow across many separate organizations, with a strict order, occasional requirements for a wet signature on paper, different forms for each institution and careful identity checks. Pulling text from documents and filling in forms is the straightforward part. The difficult part is coordination across banks, insurers, tax authorities and benefits administrators, each with its own rules for proof of authority, timelines and appeals.

Looking at what this means for product design, the choices LumenUs has described fit that problem. Asking about legal and financial standing during intake is a way to establish what permissions and entitlements a person has before making a task list. Sorting by urgency is a form of deadline-aware scheduling. Bringing in clinicians and social workers during testing reflects that mistakes carry a high cost here, and that tone, pacing and asking permission matter as much as speed. A system that prompts a bank closure before the death certificate is accepted, or raises a benefits claim without context, will lose trust fast.

In my view, handling of trust will decide whether this kind of service works. A grief platform holds death certificates, Social Security numbers, account logins, policy details and family information in one place. That concentration makes controls like least-privilege access, where people and software only see the data they need, plus scoped delegation, audit logging and clear data-retention rules, central features rather than compliance details. Worth flagging for enterprise readers, those same controls that would satisfy a chief information security officer will decide whether a grieving executor lets the system act for them. Tashakorinia's security background is relevant here, though how the team executes will matter more than past roles.

There is also a human question that automation alone cannot settle. Grief does not pause for forms. Some users will want to hand off everything, others will want to check each submission. Tailoring based on state of mind suggests LumenUs understands that difference, but keeping a dependable human-in-the-loop, where a person reviews key steps, without rebuilding the 570-hour burden is a careful balance. Anyone who has helped adult children go through a parent's files knows how fast careful work turns into exhaustion.

Looking ahead, if LumenUs gets that balance right, what it makes possible is clear. Time given back to families, fewer missed benefits, and fewer accounts left open to fraud or fees. The underlying technology pieces are not new. Used here with care, they could remove a second source of pain that now follows the first.