Building the methodology layer for communication assurance

Cavefish operates EchoDepth — a validated scoring methodology that tells organisations how high-stakes communications will land before delivery. The go-to-market is partner-led: consultancies and agencies embed EchoDepth as an evidence layer in their own client work.

The structural position we are building toward is a rating agency, not an AI tool. The value is in the methodology, the benchmark data, the governance infrastructure and the partner distribution — four things that accumulate over time and that a general-purpose LLM cannot replicate on day one.

The problem

Every organisation where a communication going wrong carries real cost has no reliable way to know whether that communication will land before it goes live. Earnings calls get challenged. Change programmes fail. Campaigns miss. Sales demos lose. None of it was inevitable — but there was no scored, evidence-based way to know in advance.

The existing tools — focus groups, media trainers, sentiment analysis, general-purpose AI — tell you what people say they felt, or describe what they said. None of them produce the scored, reproducible, auditable evidence that governance, compliance and commercial decision-making actually require.

The structural opportunity

Communication assurance — knowing in advance how a communication will land — is a methodology problem, not an AI capability problem. The limitation is not that AI cannot detect tone. It is that no one has built the validated framework, accumulated the benchmark data or constructed the governance layer that turns detection into defensible evidence.

This is the position Cavefish is building. The analogy is a rating agency: Moody's value was never the formula. It was the methodology, the data, the institutional trust, and the fact that their rating appeared on a bond prospectus and carried legal weight. EchoDepth is building toward the same structural position in communication assurance.

Validated methodology

The EchoDepth scoring framework is calibrated across 6 countries and 14 cultural cohorts with documented accuracy. It is not a prompt — it is a named, citable methodology. When it appears in a board-ready PDF, it carries methodological authority the way a named rating carries authority on a bond prospectus. This cannot be replicated by a general-purpose LLM without the same calibration investment.

Benchmark comparison data

Every scored communication adds to a growing proprietary comparison dataset. What is a normal Trust Score for a CFO earnings call in financial services? What does a high-performing change communication look like in an organisation of 10,000 people? This benchmark context is what makes a score meaningful — and it only exists if you have scored volume. It cannot be replicated on day one.

Governance and audit layer

EchoDepth outputs are consent-documented, timestamped, reproducible and structured for regulatory audit. ICO registered ZB915623. Data Processing Agreement provided as standard. In FCA Consumer Duty, investor relations and legal contexts, the output needs to hold up in an audit — not just be useful in a meeting. General-purpose AI produces conversational output. EchoDepth produces evidence.

Partner distribution

As consultancies embed EchoDepth in their own client deliverables — change management reports, investor communications advisory, research outputs — the methodology becomes part of how those firms do business. Switching cost grows independently of what underlying AI can detect. Each partner engagement is a distribution node that compounds over time.

Every organisation where a communication going wrong carries cost — earnings calls, transformation programmes, campaign launches, regulatory interactions.

Low friction: text and voice inputs, no specialist hardware. Clear commercial outcome. Direct ROI calculable from one avoided earnings call restatement or one successful change programme.

FCA Consumer Duty alone covers approximately 50,000 regulated firms in the UK. All are required to demonstrate active vulnerability identification — not sampling.

Auditable, timestamped evidence is not optional in regulated contexts. EchoDepth is one of very few tools designed to produce it — not describe it.

The global management consulting market exceeds $300bn. Consultancies are buyers of methodologies, not just tools. They embed what gives them differentiated outputs.

Consultancies already have client relationships and budgets. EchoDepth adds an evidence layer they cannot easily replicate. Each partner is a distribution node into multiple client accounts.

Service-led pilots

Organisations submit content and receive scored PDF reports within 5 working days. Pilots from £1,500/month. Revenue from day one — no software integration required.

Partner channel

Consultancies and agencies embed EchoDepth as an evidence layer in client engagements. Partners access analysis credits, white-label outputs and facilitation materials. Each partner is a distribution node into multiple client accounts.

Platform and API

As benchmark data and partner volume accumulate, the methodology moves toward a SaaS layer — API access, dashboard reporting, automated scoring pipelines. Platform revenue compounds on top of the service base.

Traction indicators

Partnerships

Former Director of Digital at The Royal Mint. Director of Digital Performance at Assurant. Behavioural analytics teams across EU, UK and US. Guest lecturer at NYU Stern. MBA from Bayes Business School. B.Eng Computer Systems Engineering.

Strategy Director at AI Wales. Presented EchoDepth at ITEC 2026 (ExCeL London) and ECSO Cyber Solution Days (Cardiff, 2024). IoD Director of the Year shortlisted 2024.