The discovery your team never has time to run.

The discovery your team never has time to run.

The discovery your team never has time to run.

The discovery your team never has time to run.

echoUX reads the feedback you already have, continuously, in the background, and finds what’s really behind each request, with the evidence to back the call in planning.

The difference

Build for a need someone actually has.

Uncover what’s behind your incoming feedback and requests. What holds up reaches your roadmap. What doesn’t gets struck, and you can see why.

Illustrative

In a single quarter, the same invoice request arrives in dozens of shapes. Three of them:

“Can we add a print button to the invoice?”

“Need to export invoices as PDF for our accountant.”

“Any way to save the invoice so I can send it on?”

Add them up, and one build looks obvious:

↓

The obvious build

Build the print button.

You ship it. Nothing moves.

echoUX reads it differently. It reads for the tension, the nuance, the thing left unspoken, and tests each reading against the evidence, until the need surfaces:

↓

The need underneath

confidence 0.86

As the person who closes the books, I need to hand finance a record they will accept, so I don’t get chased for the PO number and tax details after the fact.

Heard 18 times across support and sales. ✓ holds

So this is what you build instead:

↓

Build for the need

Build the record finance accepts.

This time it lands. And you can show why.

The same read turns “just send me an alert” into “warn me while there is still time to act,” and “add a search box” into “help me find the clause before I sign,” and that’s the depth echoUX surfaces for you.

No adoption needed

Nothing to adopt. It runs in the background.

Nothing to adopt. It runs in the background.

Nothing to adopt. It runs in the background.

Nothing to adopt. It runs in the background.

echoUX plugs into the channels you already have and delivers findings into the tools your team already works in. No new workspace, no daily check-in.

01 / connect

Point it at your channels

Slack, Intercom, CSV upload, or any tool that can send a webhook.

02 / read

It reads continuously

In the background, on its own. There is nothing for your team to run or manage.

03 / deliver

Findings come to you

They arrive in Linear, Notion or Slack, each one with its evidence attached.

How it reads

Every finding is one you can defend.

Every finding is one you can defend.

Every finding is one you can defend.

Every finding is one you can defend.

Every finding earns its confidence from the evidence, not from opinion, so it holds up when someone pushes back.

Illustrative

Raw feedback

“We set up the workspace for two days before anyone could do anything real with it.”

EVD-0244 · intercom · first-person

“I gave up halfway through config and just went back to the spreadsheet.”

EVD-0251 · slack · first-person

→

Problem · PRB-0417

✓ Holds

People give up during setup when configuration blocks their first real task.

confidence 0.86

heard 6 times

✗ Struck

A draft claim said setup takes too long for everyone. No quote supported “everyone”, so it was struck.

It compounds

It gets sharper the longer it reads.

It gets sharper the longer it reads.

It gets sharper the longer it reads.

It gets sharper the longer it reads.

The more of your feedback it reads, the clearer the picture. A theme that was a hunch this quarter is grounded by the next. Next quarter it knows more than this one.

Illustrative

Theme

Saved changes don’t take effect

this quarter

this quarter

in six months

in six months

Times heard

Times heard

6

→

24

Channels it appears in

Channels it appears in

1

→

4

Confidence

Confidence

0.71

→

0.91

Trajectory

Trajectory

● emerging

● emerging

→

● strengthening

● strengthening

Honest note: echoUX does not react to a single piece of feedback. It reports what it has seen repeat, so without a history to read it usually takes months before findings start landing in your tools.

Ask it directly

Connect the AI tools you already use.

Connect the AI tools you already use.

Connect the AI tools you already use.

Connect the AI tools you already use.

Claude, Cursor, or anything that speaks MCP. Sign in with your browser, no API key, choose which topics it may read. From there, it can do three things.

Illustrative

Check a claim

Hand it a line from a roadmap or a brief. One of four answers comes back: supported, not fully supported, contradicted, or nothing verified, with the quotes it rests on.

Claim, from a roadmap

check_claim “Travellers who have paid are left unsure whether their stay is confirmed.”

check_claim “Travellers who have paid are left unsure whether their stay is confirmed.”

↓

✓ Supported

heard 6 times across support and sales · quotes attached

Track an assumption

Stake it once. Every time that topic reads new feedback, echoUX checks it again, and tells you if the answer changes.

Assumption staked

track_assumption “Travellers who have paid trust their booking is confirmed.”

track_assumption “Travellers who have paid trust their booking is confirmed.”

↓ three months later, new feedback arrives

✗ Now contradicted

“I paid and still don’t know if it’s confirmed” — heard 4 times since. You’re told the moment this changes.

Compare a theme

Ask how a theme has moved since a date. You get when it was first and last seen, how many reading cycles carried evidence for it, and each dated point where it strengthened, weakened, or went quiet.

Theme

compare_theme “Saved changes don’t take effect” since 6 months ago

compare_theme “Saved changes don’t take effect” since 6 months ago

↓

first seen 14-03-2026 · last seen 02-09-2026

evidence in 9 cycles

02-05-2026

strengthened

19-06-2026

strengthened

07-08-2026

silent

A faster start

Bring your history with you.

Import the feedback you already have and echoUX reads it at the dates it originally arrived, so the picture compounds in days rather than months.

Pricing

Priced by what it reads, not by seats.

Priced by what it reads, not by seats.

Priced by what it reads, not by seats.

Priced by what it reads, not by seats.

Start with one product area, and add more as you grow. It runs in the background, so there are no seats to buy. The price follows how much it reads.

Starter

£49 / month

For one product area, the one you choose.

Continuous background reading

Every finding carries its evidence

Delivered into the tools you already use

Import your feedback history

250 pieces of feedback, each month

30 AI claim checks a month, and 1 tracked assumption

Team

from £149 / month

For several teams, each with its own area.

Everything in Starter

Three areas, then £49 each after

750 pieces of feedback, across three product areas

90 AI claim checks a month, and 3 tracked assumptions

Each area delivered into its own tools

Organisation

Custom

For the whole organisation, not just product.

Everything in Team

Reading across product, sales and support

Reading volume agreed in your contract

300 AI claim checks a month, and 30 tracked assumptions

Concierge onboarding and support guarantees

SSO and security review

Starter and Team start with a free 30-day trial. Card required, nothing charged today.

belief → knowledge

The discovery your team never had time to run, now runs itself.

The discovery your team never had time to run, now runs itself.

The discovery your team never had time to run, now runs itself.

The discovery your team never had time to run, now runs itself.

You walk into planning with the real findings already made, and the evidence to stand behind each one.

What people ask us

We already use a tool that collects feedback and summarises it. How is this different?

Those tools show you what people said. echoUX tells you what’s actually earned the right to be called a finding, and it won’t show you one that can’t be backed up. It reads the same feedback those tools read, alongside them, not instead of them.

So what does echoUX actually deliver?

A finding takes whatever shape your feedback can support. If it points at a goal, you get a job to be done. If it’s mostly about the solution, a how-might-we. If it’s a clear, well-defined problem, a problem statement. If it’s closer to a need, a need statement. Only the shapes the evidence holds up reach you, each one properly written and with its quotes attached.

Is this a replacement for our user research?

Absolutely not. You should always have UX researchers on your team. As generative research becomes richer, they’re the ones who catch nuance even when no clear signal is present, and no tool can replace that judgment. We built echoUX because we deeply believe many customer-facing companies need some form of generative research, especially when they lack the means, buy-in, or capacity to build a dedicated research function. For those who already have one, echoUX is a complement, never a replacement.

What can it read?

Slack, Intercom and Zendesk, plus CSV upload for your history and a webhook for anything else. More channels are being added.

This means reading our customers’ conversations. Where does that data go?

echoUX reads the channels you connect and keeps the raw feedback for up to 180 days. After that, the raw feedback is deleted. What survives past 180 days is the finding itself, and that’s designed to carry no personal information. Full detail on how each type of data is handled is in our DPA. Disconnect a channel and the reading stops. Close your account and everything is deleted.

How do I know a finding is real?

Every finding travels with the quotes that ground it. If a claim reaches past its evidence, it gets struck.

How is it priced?

By the product areas you are reading, never by seat. Start with one on Starter and add more as you go.