Question 1

Why is manual QA insufficient for FCA Consumer Duty vulnerable customer identification?

Accepted Answer

Manual QA typically reviews 2–3% of customer interactions, selected by sampling rather than systematic risk criteria. This means 97–98% of interactions are never reviewed for vulnerability signals. The FCA's Consumer Duty requirement for proactive identification applies across all interactions — a 2–3% QA sample cannot satisfy a 100% proactive identification requirement. EchoDepth analyses every interaction submitted, generating vulnerability flags in real time without human reviewer bandwidth as a constraint.

Question 2

What does EchoDepth identify that manual QA reviewers miss?

Accepted Answer

Manual QA reviewers assess what they can consciously observe in an interaction recording — primarily the words used and the surface tone. EchoDepth detects involuntary physiological signals that are not visible to human reviewers: specific facial Action Unit combinations indicating suppressed distress, voice prosody changes indicating cognitive load, and text-emotion divergence where the content is neutral but the underlying emotional signal shows vulnerability. Reviewers also bring inconsistency — the same interaction may be flagged by one reviewer and missed by another. EchoDepth applies consistent signal thresholds across every interaction.

Question 3

Does EchoDepth replace compliance staff?

Accepted Answer

No. EchoDepth replaces the screening function of manual QA — identifying which interactions warrant human review — not the review function itself. Compliance staff spend their time reviewing and acting on flagged interactions rather than manually scanning recordings looking for vulnerability signals they might miss. This dramatically increases the effective productivity of compliance resource while improving coverage and consistency.

Question 4

What is the cost comparison between EchoDepth and manual QA?

Accepted Answer

Manual QA has two cost components: the reviewer time cost (typically £25–40k per FTE per year) and the cost of missed vulnerability flags (FOS referral costs, FCA enforcement exposure, customer redress). EchoDepth has a per-interaction cost that scales with volume. At a contact centre handling 50,000 interactions per month with a 2% QA rate, EchoDepth covers 50,000 interactions versus 1,000 — while typically costing less than the equivalent reviewer capacity for the 1,000 interactions, and avoiding the regulatory exposure from the 49,000 uncovered interactions.

Capability	Manual QA	EchoDepth
Interaction coverage	2–3% (sample)	100%
Vulnerability flag timing	Hours or days post-interaction	Real-time, during interaction
Detects suppressed distress	Rarely — reviewer sees surface	Yes — involuntary signal
Consistency	Variable — reviewer dependent	Consistent signal thresholds
FCA audit trail	Incomplete (97% unreviewed)	Full interaction-level log
Outcomes monitoring MI	Manual compilation	Automated, continuous
Scales with interaction volume	Requires proportional headcount	Linear API cost only
FCA proactive ID requirement	Does not satisfy	Satisfies

EchoDepth vs Manual QA
for vulnerable customer identification.

Head to head

EchoDepth doesn't replace compliance staff. It changes what they do.