EchoDepth Signal
Detect deception in written statements.
Linguistic patterns reveal what people try to hide. EchoDepth Signal analyses written statements for indicators of deception, scoring credibility and flagging specific concerns for human review.
The Science
Forensic Statement & Linguistic Analysis (FSLA)
When people construct false narratives, their language betrays them. Decades of forensic linguistics research has identified specific patterns — hedging, temporal gaps, sensory omissions, pronoun shifts — that correlate strongly with deceptive statements.
EchoDepth Signal applies this research at scale. It analyses written statements for 30+ linguistic indicators, weights them by context, and produces a credibility score with specific flagged passages.
This is not a lie detector. It is a prioritisation tool — helping investigators, claims handlers, HR teams and legal professionals focus their attention on statements that warrant deeper examination.
- 30+ linguistic deception indicators
- Real-time credibility scoring (0-100)
- Passage-level flagging with explanations
- Trend analysis across statement sections
- Actionable recommendations for follow-up
Example deception indicators
Temporal gap
"After I arrived... the next thing I remember..." — Missing time in critical sequence
Sensory omission
No visual, auditory or physical detail — genuine memory typically includes sensory context
Pronoun distancing
"The car was damaged" vs "I damaged the car" — avoiding ownership of actions
Excessive hedging
"I think maybe... I'm not sure but... I believe..." — uncertainty where certainty expected
Narrative inconsistency
Level of detail varies dramatically between sections — suggests fabrication in sparse areas
Platform Output
What Signal reveals.
Every analysis produces a structured report with credibility scoring, flagged indicators, and specific recommendations for follow-up investigation.
Statement Analysis
Insurance Claim — Vehicle Damage
347 words · 12 indicators detected · 4 high-weight flags
Deception Intensity (Start → End)
Intensity peaks during description of the incident itself — sparse detail where specific recall expected.
Flagged Indicators
Critical time period unaccounted for between parking and discovering damage.
"I parked the car and went inside. When I came back out, I noticed the damage."
No visual, auditory or physical detail during incident description. Genuine recall typically includes sensory context.
"The damage must have happened while I was away."
Avoids first-person ownership when describing the damage state.
"The vehicle was found with a dent on the rear panel."
Recommendations
Request specific timeline with timestamps between parking and discovery
Ask for witness statements or CCTV evidence from the location
Compare damage pattern with claimant's stated timeline
Applications
Where Signal applies.
Claims Verification
Analyse claimant statements before payout. Flag high-risk submissions for investigator review. Prioritise your fraud team's attention.
Learn more → HRRecruitment Screening
Screen candidate statements, application narratives and reference letters. Identify inconsistencies before they become hiring mistakes.
Learn more → InvestigationsWorkplace Investigations
Analyse witness statements and incident reports. Provide HR teams with objective analysis to support fair decision-making.
Learn more → LegalLegal & Evidence
Support solicitors and legal teams with witness statement analysis. Identify statements requiring deeper scrutiny before trial preparation.
Learn more →Common Questions
Understanding EchoDepth Signal.
Does Signal prove someone is lying?
No. Signal identifies linguistic patterns that correlate with deception. The output is a risk score and specific indicators — a prioritisation tool for human investigators, not a verdict. Any findings should be corroborated through traditional investigative methods.
What types of text can Signal analyse?
Any written narrative statement — insurance claims, incident reports, witness statements, job application narratives, reference letters, legal depositions. Signal works best on statements where the author is describing events or making factual assertions.
How is Signal different from a polygraph?
Polygraphs measure physiological arousal during questioning. Signal analyses written text — no live subject interaction required. Signal can process thousands of statements at scale, prioritising which warrant human investigation. It's a screening tool, not a real-time interrogation device.
Is Signal suitable for legal proceedings?
Signal provides investigative support, not courtroom evidence. The analysis helps legal teams identify statements requiring deeper scrutiny. Any findings should be independently verified before use in legal proceedings.
See what Signal reveals in your statements.
Whether you're verifying claims, screening candidates, or supporting legal analysis, we can demonstrate Signal on your use case.