What is the difference between EchoDepth and Culture Amp?

Culture Amp is a survey platform: it collects what employees choose to write in response to structured questions and open-text prompts, then benchmarks and trends those responses over time. EchoDepth is an emotional AI platform: it analyses the emotional signal beneath the written response, detecting where the language is positive but the underlying emotion is negative. EchoDepth does not replace Culture Amp — it adds the emotional intelligence layer that Culture Amp cannot provide.

Why does Culture Amp miss retention risk?

Culture Amp scores what employees write. Employees experiencing early-stage disengagement, quiet quitting or pre-exit emotional withdrawal tend to write neutrally — neither strongly positive nor strongly negative — because they are managing their professional presentation in the survey. This masked dissatisfaction is invisible to any survey platform regardless of how sophisticated its analytics are. EchoDepth detects it because it measures the emotional signal beneath the text, not the text itself.

Platform comparison

EchoDepth vs Culture Amp — what the gap between your scores and your retention rate is telling you.

Q: Can EchoDepth work with existing Culture Amp data?

Yes. EchoDepth can ingest exported Culture Amp open-text responses and apply emotional AI analysis to the data you already hold — without any additional participant burden. You run your Culture Amp survey as normal. You export the open-text responses. EchoDepth scores them for VAD (Valence, Arousal, Dominance) and produces a text-emotion divergence analysis showing where the stated sentiment and the felt sentiment are misaligned.

Q: What does EchoDepth produce that Culture Amp does not?

EchoDepth produces: text-emotion divergence scores (where stated and felt sentiment misalign), masked dissatisfaction identification (employees writing positively while feeling negatively), Employee Culture Index (ECI) emotional health scores, a Trust Risk Register for leadership review, and a board-ready culture health report with specific, prioritised recommendations. These outputs are not available from any survey-based platform.

Culture Amp measures stated employee sentiment. EchoDepth measures felt sentiment — the emotional signal beneath what employees write. The gap between the two is where retention risk lives, and where all survey-based platforms, including Culture Amp, Glint and Peakon, are structurally blind.

By Jonathan Prescott, Cavefish Ltd · Published April 2026

The situation that brings most HR leaders here

Your Culture Amp score is 7.1. Three people from your strongest team hand in their notice next month.

The survey gave you no warning because it measured what employees were willing to write — not what they actually felt. Culture Amp, Glint, Peakon and every other survey platform share this structural limitation. It is not a product failure. It is the fundamental constraint of asking people to describe how they feel in a professional context.

What it measures	Culture Amp (and Glint, Peakon, Qualtrics)	EchoDepth Insight
Primary data source	Structured survey responses and open-text comments	Emotional signal from open-text responses + optional facial coding sessions
What it captures	Stated sentiment — what employees choose to write	Felt sentiment — the emotion beneath the written response
Text analysis	NLP sentiment scoring of written words	VAD (Valence, Arousal, Dominance) emotional signal independent of word choice
Masked dissatisfaction	Cannot detect — invisible in word-level analysis	Primary capability — text-emotion divergence is core output
Retention early warning	Trend monitoring over time — warns late	Emotional risk scoring identifies pre-exit condition before it becomes visible
Benchmarking	Extensive — industry and role benchmarks	Internal — ECI score against prior periods
Board reporting	Dashboard and trend reports	Structured board-ready culture health report with prioritised recommendations
Works with existing data	N/A — is the data source	Yes — can ingest Culture Amp exports without additional participant burden
Investment	Per-employee per-year subscription	From £3,500 (proof of concept on existing survey data)

Why Culture Amp cannot detect masked dissatisfaction

Culture Amp and equivalent platforms — Glint (Microsoft Viva), Peakon (Workday), Qualtrics Employee XM — operate on written language. Their NLP and sentiment analysis layers score what employees write. This is valuable data, but it has a structural ceiling.

Employees experiencing early disengagement, quiet quitting or pre-exit emotional withdrawal do not write strongly negative survey responses. They write neutrally: professionally managed language that neither triggers concern nor generates positive signal. They have learned — from experience — that survey responses persist on record and shape how they are perceived by management. The result is a systematic underrepresentation of negative emotional experience in survey data.

EchoDepth detects this masked dissatisfaction because it measures the emotional signal independently of the words chosen. When a response is written positively but the underlying emotional state is negative, the VAD scoring of the text will diverge from the surface sentiment of the language. This text-emotion divergence — a pattern invisible to all word-based analysis — is the primary signal that a team member is emotionally withdrawing before their behaviour makes it visible.

EchoDepth does not replace Culture Amp — it adds the layer surveys cannot

Culture Amp provides things EchoDepth does not: benchmarking against industry norms, longitudinal pulse tracking, manager dashboards, onboarding and exit feedback flows, 360-degree feedback tools, and large-scale quantitative engagement measurement. For the breadth of your people analytics programme, Culture Amp or an equivalent platform remains essential.

EchoDepth adds a single capability that no survey platform can provide: the emotional intelligence layer on top of the data you already collect. You run your Culture Amp survey as normal. You export the open-text responses. EchoDepth scores them for emotional signal and returns a text-emotion divergence analysis, an Employee Culture Index (ECI) score, and a board-ready report identifying the specific teams and themes where the gap between stated and felt experience is widest.

The engagement model is straightforward: send us the anonymised open-text data from your most recent survey. We apply the EchoDepth emotional analysis layer and deliver a structured report within five working days. No new surveys, no participant burden, no change to your existing methodology.

The ECI — what Culture Amp's engagement score cannot tell you

Culture Amp produces engagement scores and eNPS ratings — both based on stated response. EchoDepth produces the Employee Culture Index (ECI): a composite emotional health score derived from VAD analysis of open-text responses. The ECI captures the emotional dimension of the workforce that engagement surveys cannot reach.

A healthy organisation typically shows a negative-to-positive valence ratio below 1.2:1. Ratios between 1.2:1 and 1.5:1 indicate structural friction. Above 1.5:1 signals systemic strain with elevated voluntary turnover risk within 12 months. A ratio of 1.8:1 or above — the pre-exit condition — represents a workforce that is functionally performing but emotionally exhausted. This signal cannot be detected by any platform that analyses only what employees choose to write.

The practical workflow

Run your Culture Amp survey as normal

Export anonymised open-text responses and send to EchoDepth

EchoDepth applies VAD emotional analysis — 5 working days

Receive board-ready report: ECI score, divergence map, recommendations

Common questions

Is EchoDepth a replacement for Culture Amp?

No. Culture Amp provides benchmarking, pulse surveys, manager dashboards, onboarding and exit flows, and longitudinal tracking that EchoDepth does not replicate. EchoDepth is an emotional intelligence layer — it analyses the open-text data Culture Amp collects and adds the felt sentiment dimension. The most effective people analytics programme uses both: Culture Amp for breadth and benchmarking, EchoDepth for depth and early warning.

Can EchoDepth analyse our existing Culture Amp data?

Yes. EchoDepth can ingest exported open-text responses from Culture Amp, Glint, Peakon, Qualtrics or any culture survey platform — and apply emotional AI analysis without any additional participant burden. You run your existing survey as normal. You send us the anonymised open-text export. We return the emotional analysis within five working days.

Why does our Culture Amp score look fine when people are leaving?

Because employees experiencing early disengagement write neutrally in surveys — professionally managed responses that neither trigger concern nor generate strong positive signal. They have learned that survey responses persist on record. Culture Amp cannot detect this masked dissatisfaction because it measures word choice, not emotional state. EchoDepth detects it because it scores the emotional signal independently of the words used — identifying text-emotion divergence before the behaviour becomes visible in turnover.

What does EchoDepth produce that Culture Amp does not?

Text-emotion divergence scores, masked dissatisfaction identification, Employee Culture Index (ECI) emotional health scores, a Trust Risk Register, and a board-ready culture health report with specific prioritised recommendations. These outputs are unavailable from any survey-based platform including Culture Amp, Glint, Peakon and Qualtrics.

How much does it cost to add EchoDepth to our existing culture programme?

A proof-of-concept analysis of your most recent survey open-text data starts from £3,500 — a scoped project with a defined deliverable. Annual subscription from £12,000/year for ongoing emotional intelligence analysis alongside your regular survey cadence. All projects include a written proposal with fixed fees before any work begins.

The masked dissatisfaction mechanism explained in full — why engagement scores look stable before resignations arrive.

Read →

EchoDepth for culture

Find out what your Culture Amp data is not telling you.

Send us your most recent survey open-text export. We will show you what the emotional AI layer surfaces — including whether masked dissatisfaction is present, where it sits, and what it predicts.

Start the conversation Culture measurement