A structured quality framework built especially for law firms
HarkIQ analyses call transcripts using a structured quality framework developed specifically for the UK legal sector. The framework draws on established best practice from regulated industries including financial services, healthcare, and telecommunications, adapted for the unique demands of claimant law firm client service. Every transcript is assessed consistently against the same dimensions, ensuring that scores are comparable across calls, handlers, departments, and time periods.
Each analysed call receives a HarkIQ Score between 0 and 100. This composite score reflects the overall quality of the call as experienced by the caller. It is calculated from weighted assessments across eight distinct dimensions of call quality, each measuring a different aspect of the interaction. The weighting reflects the relative importance of each dimension to the caller's experience, with communication and listening skills carrying the greatest influence.
70+
Good practice
40–69
Room for improvement
Below 40
Requires attention
HarkIQ assesses every call against eight dimensions. These are not arbitrary categories — they are grounded in regulatory frameworks, academic research into client satisfaction, and operational best practice from sectors that have invested heavily in call quality.
How the call begins matters. This dimension assesses whether the caller was greeted warmly, identified by name, and made to feel they had reached the right place. First impressions set the tone for everything that follows.
Effective call handling requires genuine listening. This dimension measures whether the handler demonstrated understanding, avoided unnecessary jargon, asked appropriate questions, and communicated clearly without dominating the conversation.
Many callers to claimant law firms are distressed, in pain, or dealing with significant life disruption. This dimension assesses whether the handler recognised and responded appropriately to the caller’s emotional state, and whether indicators of vulnerability were identified and acted upon.
Getting the right information during the first call is essential for progressing the matter efficiently. This dimension measures whether appropriate questions were asked, whether sufficient detail was captured, and whether the call was routed to the correct department or specialist.
Callers need to know what happens next. This dimension assesses whether clear timescales were provided, whether costs or funding arrangements were explained where appropriate, and whether the caller left the conversation knowing exactly what to expect.
Not every call can be resolved by the first handler. When escalation is needed, this dimension measures whether it was handled smoothly, whether context was passed to the next person, and whether the caller was spared the frustration of repeating their story.
Law firms operate within a regulated environment. This dimension assesses whether appropriate data protection notifications were given, whether relevant disclaimers were provided, and whether the call met the standards expected under SRA regulation.
How a call ends determines whether the caller feels confident or abandoned. This dimension measures whether a clear next step was confirmed, whether a specific contact name and timeframe were provided, and whether the caller knew exactly who to speak to if they needed to follow up.
HarkIQ does more than assign a number. Each call is automatically classified by type, department, urgency, and outcome. Vulnerability indicators are flagged using a framework aligned with the FCA's four drivers of vulnerability: health, life events, resilience, and capability. Sentiment analysis tracks how the caller's emotional state changed through the conversation, revealing whether the interaction reassured or frustrated them. Callback commitments are recorded and tracked. Repeat callers are identified. These layers of intelligence transform raw call data into actionable operational insight.
All analysis is performed using Anthropic's Claude AI, accessed through a secure closed API. This means that call transcripts submitted for analysis are not used to train AI models and are not retained beyond the processing window. Before any transcript is sent for analysis, personally identifiable information is automatically redacted, stripping phone numbers, email addresses, postcodes, NHS numbers, National Insurance numbers, dates of birth, bank details, and card numbers. The original unredacted transcript remains securely stored and is only accessible to authorised users within the uploading firm. Full details of our data handling practices are available in our data upload terms.
The HarkIQ framework is not static. It is refined through ongoing research into client service best practice, feedback from participating firms, and analysis of scoring patterns across the platform. Dimension weightings and assessment criteria are reviewed quarterly to ensure they reflect current regulatory expectations and evolving standards of client care.