We launched an AI assistant for unfinished work. Is the problem clear enough?

Thank you—this captures the problem very well.

The feedback here has made it clear that both “AI continuity assistant” and “unfinished work” are too abstract. We need to lead with the actual pain: open loops, dropped follow-ups, and mentally tracking who owes what.

“Review-first follow-up assistant” and “smart open-loop tracker” are much closer to how users understand the value. We would probably avoid “autopilot,” since Vortyx deliberately keeps the user in control rather than sending messages autonomously.

We’ll make the review-first trust signal much more prominent. Thanks for the direct and useful feedback.

That is the key tension in the design. In our closed-loop testing, we did see false positives, but the filtered review set was still much smaller than the underlying email volume.

The larger value begins after review: once a user accepts an item, Vortyx keeps the follow-up, commitment, or waiting thread visible until it is resolved. We still need to validate with active users whether the review step remains lightweight over time rather than becoming another inbox.

Point taken—the feedback on “AI continuity assistant” has been remarkably consistent. We’ll stop leading with the category label and focus on the concrete outcome: keeping follow-ups, commitments, and waiting items from slipping through the cracks.

That is a fair point. The feedback has made it clear that leading with a new category creates unnecessary friction.

We are now looking more closely at which users experience the highest cost from dropped threads—such as agencies awaiting client approvals, sales professionals managing follow-ups, and founders coordinating important conversations—so we can lead with one concrete workflow rather than a broad category.

Point taken.

Point taken

Had the same reaction. This is a very real problem but "unfinished work" is not the right term.

That is a fair observation. We have not yet tested different landing-page headlines.

The consistency of the feedback here is a strong signal that “AI continuity assistant” is creating unnecessary friction. We’ll experiment with more outcome-led headlines around follow-ups, commitments, and what still needs attention, and measure which one gets users to keep reading and try Vortyx.

To answer directly: the problem is understandable, but 'AI continuity assistant' is a bit abstract as a category label I'd lean toward describing it functionally first ('finds what you forgot you promised to follow up on') and let the category label be secondary. I'd call it closer to a task-management product with an AI-native default, not a general productivity assistant the review-first approach is the differentiator worth leading with, since most tools either autofill your calendar or require full manual entry.

One thing I'm curious about, given the review-first design: once something's confirmed as needing follow-up, is the expectation that the user still writes and sends it themselves, or is there any path to handing that execution off? That gap found it, but still have to act on it is exactly what I'm building TaskRelay around.

The clearest use case today is an important email thread that remains unresolved.

For example, you send a client a proposal and expect a response. Vortyx can surface it as a potential follow-up, show why it may need attention, and—once you confirm it—keep the thread visible until they reply or you mark it handled.

It also works in the opposite direction, such as detecting that you promised to send someone a revised document by Friday.

This comment was deleted a month ago

This comment was deleted a month ago

Good distinction 'recovering context' is actually the harder problem than 'remembering it exists.' A tool can nudge you that something's pending, but reconstructing what was actually said, promised, and where it was left off is the real time cost. That's probably the strongest use case to lead with forgotten customer conversations since the stakes (a real relationship, a missed deal) are highest there, versus something like an abandoned personal doc

That is fair—“unfinished work” is broader than what we mean.

The day-to-day problem that led us to Vortyx was this: we would send an important email, promise to send something later, or wait on someone else, and then have to remember the thread manually across inboxes, notes, and calendars.

So the clearest scenario is not a half-finished document or project. It is an open conversational loop—something sent, promised, or awaited that still needs to reach resolution.

Agreed on the framing 'unfinished work' is broad enough to mean five different things until there's one concrete example anchoring it. Also curious about the founder's answer to your last question the personal daily friction that led to building it is usually the clearest way to explain any category-less product.

Agreed, especially on making review-first loud that's the trust signal most AI tools skip entirely, and it's exactly the kind of thing that differentiates 'helpful' from 'another thing I have to double-check.' Landing pages usually bury their best trust signal in a features list instead of the headline

Strong reframe 'Active Thread Tracker' is a much sharper mental model than 'continuity assistant.' Your LeadPulse example nails the actual failure point too pulling leads is the easy part, tracking who replied/ghosted/promised-an-intro is where it quietly turns into a spreadsheet nightmare. That tracking-and-remembering gap is one thing, but even once you know who needs a follow-up, someone still has to write and send it that execution layer is what I'm building TaskRelay around. Is the follow-up drafting/sending itself something you handle yourself for LeadPulse users, or does that stay entirely on their side?

Thank you for the feedback.

Your distinction between the initial action and the difficult follow-through is exactly what led us to build Vortyx. “AI productivity assistant” feels too broad, while “task manager” suggests that users still have to manually translate everything into tasks. “Active Thread Tracker” and “Open Loop Detector” are much closer to the outcome we are trying to deliver.

We also agree that review-first should be more prominent. Vortyx can surface a possible follow-up or commitment, but the user decides whether it should be tracked, dismissed, scheduled, or acted on. It does not automatically send messages.

On the context question: each source is connected separately rather than merged into one uncontrolled stream. For email, Vortyx examines the relevant message or thread context; for voice, it works from the transcript; and calendar or captured notes provide additional timing and intent signals. It then converts those signals into structured candidates such as a follow-up, commitment, waiting item, recurring action, or next step.

Each candidate retains its source and supporting context, and the user reviews it before it becomes an active tracked item. We also use confidence checks and deduplication so the same thread is not repeatedly surfaced as new work.

We are still improving how context should carry across sources without making the system feel opaque. Your outreach example is especially useful because it gives us a concrete workflow to test: reply received, no reply, promised introduction, follow-up date, and context captured elsewhere.

And thank you for the LeadPulse example—it makes the cost of losing these threads very tangible.

Good distinction 'what needs my attention across multiple businesses' is actually a more useful category than a single-thread continuity tool. Makes sense you'd frame it that way given the care home + FounderFlow juggling you mentioned before. By the way, still up for trying that first routine hand-off whenever something comes up no rush, just flagging it's still open whenever it's useful.

Thanks—this distinction is helpful.

FounderFlow sounds focused on helping founders see what needs attention across multiple businesses, whereas Vortyx is more focused on preserving the thread after an action has been taken: what you are waiting on, what you committed to, and what still needs to happen before something is resolved.

I agree with your point on the headline. “AI continuity assistant” asks people to understand the category before they understand the problem. Several comments have now raised the same concern about the headline, so we’ll change it and lead with the concrete promise instead.

“Never lose track of what you’re waiting on” is much more immediate, and the broader category can come later once the value is clear.

Open loops inbox' is a good concrete first-screen label testable, too, since you can literally ask a new user 'what do you expect to see here' before they connect anything and check if their guess matches reality.

That is very helpful. “AI continuity assistant” may be useful internally as a category, but it is probably asking users to decode too much before they understand the value.

“Open loops inbox” is much more concrete. It immediately suggests a place where waiting items, commitments, and next steps remain visible until they are resolved.

Your point about predictability before connecting email is especially important. We should show clear examples of what Vortyx will surface—and what it will not—before asking users to grant access.

Thanks—this distinction resonates. Vortyx is not a voice or WhatsApp sales agent, so the Swiftsell AI reference may relate to another product, but the underlying trust question is very relevant.

We are exploring whether people value Vortyx primarily for its AI capabilities, or because it gives them confidence that follow-ups, commitments, and waiting items will not quietly slip through the gaps. We increasingly believe the second is the stronger value proposition.