The To-Do List That Updates Itself

01 /The Concept

This platform is still in closed alpha, but the startup building it has already closed its first $5M funding round.

Hoop built an AI agent that automatically assembles a to-do list of tasks assigned to a specific person.

Or as the partner at Index Ventures – the round's lead investor – described it: Hoop is "a to-do list that updates itself."

The AI agent monitors relevant Slack channels and joins Zoom and Google Meet calls to track every message and instruction directed at a given person – then automatically builds and updates a task list from those inputs.

More platform integrations are on the roadmap.

The auto-generated task list resembles an email inbox: each task has a title and a brief description. Users can work through incoming tasks at their own pace – moving some into an active list, leaving others, deleting the rest. Descriptions are editable, and manual tasks can be added directly.

Hoop commits that captured task content will never be used to train its AI models. All task text is end-to-end encrypted, similar to messaging apps like Signal or WhatsApp.

Hoop currently works as a personal tool, but a team version is coming soon – with shared task lists, editing, and commenting. Even for alpha testers, the individual version is paid: $35/month.

02 /Why It Matters

As the Index Ventures co-founder put it: "I use Hoop every day, in every online meeting."

That makes Hoop a passing example of the "toothbrush test" Larry Page uses to evaluate products for acquisition or investment: "Will people use this at least once or twice a day, and does it make their lives better?" If yes, it's worth buying or investing in regardless of current revenue. If no, pass – even with strong financials.

Interestingly, Y Combinator participant Onward was building something very similar last year – but as of a [January review](/review/princip-konkurentnoj-borby-na-shag-blizhe-k-celi), it had stalled out beyond announcements.

Too bad, because the current state of Hoop is arguably just the starting point.

The product is still not quite the "self-updating to-do list" it's been called. Right now it's a tool for accumulating tasks. Truly automatic updates would require the system to also:

- prioritize tasks – sorting them into "urgent" vs. "important" buckets to surface a ready-to-go today list,

- automatically mark tasks complete as they're resolved,

- and notify the people who assigned those tasks that they've been done.

Any tool becomes far stickier when it covers the full cycle of a problem – from capture to closure. And charging more for that becomes much easier to justify.

A [recent review covered](/review/vygodnee-prodavat-ne-instrument-a-rezultat) a startup that built an AI tool for maintaining a database of acquisition targets. Initially it sold access to the database. But the real funding came when it repositioned as an "AI investment bank" – delivering a curated list of companies that match a client's criteria and whose owners are open to a sale, then providing financing for the deal. The database became an internal tool, not a product.

Hoop's missing capabilities could also be assembled from adjacent startups.

Produce8, [covered previously](/review/prostoj-variant-dlja-vzljota) in fall 2022, built a platform that makes the entire team's workload visible. Members connect the apps they use; Produce8 tracks their activity and surfaces it on a shared dashboard alongside current task lists – eliminating the need for recurring status meetings. It raised approximately $4.41M (CAD $6M).

Rize, [covered previously](/review/mnogo-ili-jeffektivno) in February, built an "AI coach" that tracks individual app usage patterns and helps people work more effectively on current tasks – or tells them to take a break before they burn out.

Stitch these individual and team work-monitoring tools together with an intelligent task list, and you'd have something far more complete: a system that doesn't just track what needs doing, but actively helps people get it done.

03 /Opportunities

To-do lists genuinely pass the toothbrush test – millions of people consult them multiple times a day. A version that harnesses modern AI capabilities across the full lifecycle described above is a timely and well-positioned product opportunity.

But the bigger ambition is even more interesting.

The tasks that flow through Slack and surface in meetings don't appear from nowhere. At the top of the hierarchy sits company strategy. Strategy decomposes into department-level action plans. Those plans become individual task assignments for the people who execute them.

Hoop lives at the bottom – the individual task layer. But AI platforms are already being built for the upper layers too.

Elate, [covered here](/review/jeti-proschjoty-nelzja-kompensirovat-uspehami) in April, built a platform for developing company strategy, turning it into execution plans, and tracking implementation. It raised $9.4M.

ProperPlan, [covered here](/review/ubej-biznes-trenera) late last year, built an AI tool that helps solo founders and small teams build development plans all the way down to daily task lists.

Put these together, and the vision becomes clear: digitize and automate the entire task-setting cycle, from strategy formulation through department planning to individual assignment. Done right, that eliminates the need for dedicated status meetings and agents like Hoop – because every task at every level is already queued and distributed dynamically based on real-time headcount and capacity. Bug tickets from Jira would flow in automatically alongside everything else.

Prioritization itself is another candidate for automation – fed by live data from financial systems, CRMs, and KPI dashboards. AI-proposed priority stacks could be overridden by responsible humans, with the system learning from those corrections by asking why a given priority was changed.

It sounds like science fiction. But as Paul Graham said: to build a breakthrough startup, first imagine you're living in the future. Then build the pieces that are missing. This direction fits that frame very well.