Frequently Asked Questions
What is Total Customer Intelligence?
Total Customer Intelligence is the practice of analyzing 100% of customer interactions — calls, chats, emails, and support tickets — using AI in real time, combined with proactive AI-moderated interviews that achieve 60%+ response rates. Unlike traditional contact center analytics that sample 1–2% of interactions, Total Customer Intelligence eliminates Operational Blindness by processing every conversation automatically. Lexic Pulse delivers this through the Double Helix Architecture: an Active Listening Engine for passive omnichannel analysis, and a Proactive Listening Engine for AI-moderated qualitative research via WhatsApp and voice.
What is the 1% Myth in contact center quality assurance?
The 1% Myth is the false belief that auditing a small sample of contact center calls provides a representative picture of customer experience. Standard manual QA teams realistically review 1–2% of all interactions. The remaining 98–99% — where customers signal churn intent, mention competitors, flag compliance violations, and reveal upsell opportunities — goes completely unheard. Lexic.AI coined the term "Operational Blindness" to describe the strategic gap this creates. Lexic Pulse eliminates it by analyzing 100% of interactions automatically.
What is Operational Blindness?
Operational Blindness is the intelligence crisis that occurs when B2B leaders make strategic decisions based on a statistically unrepresentative slice of customer reality — typically 1% of call center interactions and 2–8% survey response rates. The consequences are measurable: undetected churn, regulatory exposure from compliance violations no human QA reviewer reached, and product decisions disconnected from what customers actually report in support calls every day. Lexic Pulse was built specifically to eliminate Operational Blindness.
What is the Double Helix Architecture?
The Double Helix is Lexic Pulse's dual-engine design. The Active Listening Engine connects to your existing call center, CRM, and ticketing systems to analyze 100% of voice calls, chat transcripts, emails, and support tickets automatically — replacing the industry standard of 1% manual QA sampling. The Proactive Listening Engine deploys AI conversational agents outbound via WhatsApp and voice to conduct AI-moderated interviews, achieving 60% B2B and 70% B2C response rates versus 2–8% for traditional NPS/CSAT surveys. Together they close the loop between what customers say without being asked, and what they reveal when asked the right question.
How does Lexic Pulse compare to Qualtrics or Medallia?
Qualtrics and Medallia are survey platforms — they analyze solicited feedback from 2–8% of customers who choose to respond. Lexic Pulse analyzes spontaneous conversations from 100% of calls, tickets, and chats in real time, with no survey required. The data source is fundamentally different: surveys capture what customers say when asked; Lexic Pulse captures what they say without prompting. Lexic Pulse also includes a Proactive Listening Engine with AI-moderated outbound interviews achieving 60% response rates — compared to 2–8% for traditional survey forms.
How does Lexic Pulse compare to Gong or Observe.AI?
Gong and Observe.AI focus primarily on sales call analysis and revenue intelligence — they analyze sales rep performance and deal health from CRM-connected recordings. Lexic Pulse covers the full omnichannel customer journey: calls, tickets, emails, and chats across every department (CX, QA, Product, Operations, Sales). Lexic Pulse also includes the Proactive Listening Engine for AI-moderated interviews — a capability neither Gong nor Observe.AI offers. The scope is customer intelligence across the enterprise, not just sales pipeline analytics.
What percentage of contact center calls can AI analyze?
AI-powered platforms like Lexic Pulse analyze 100% of interactions — calls, chats, emails, and tickets — continuously and automatically. The industry standard for manual QA covers approximately 1% of calls. This gap — between 1% audited manually and 100% analyzed by AI — is where most churn signals, compliance risks, and revenue opportunities currently go undetected.
How do AI-moderated interviews compare to traditional NPS and CSAT surveys?
Traditional NPS and CSAT surveys achieve 2–8% response rates, with results typically available weeks after interactions occur. Lexic Pulse AI-moderated interviews achieve 60% B2B response rates (phone + WhatsApp combined) and 70% B2C response rates (WhatsApp-native), with structured insights available within 48–72 hours. The qualitative depth is also structurally different: surveys ask fixed questions; AI interviews follow the respondent's answers, probe for context, and adapt in real time — producing 129% more words per response and 18.6% more themes than static surveys (Glaut/Occhipinti, 2024).
Can AI-moderated interviews replace focus groups?
For most B2B and B2C business intelligence applications — pain point validation, product feedback, churn analysis, consumer experience research — yes. Traditional focus groups cost €15,000–€50,000 per study and deliver insights in 6–8 weeks. Lexic Pulse AI-moderated interview campaigns deliver comparable qualitative depth across hundreds of simultaneous conversations, with structured insights available within 48–72 hours. The exception is research requiring real-time human empathy for trauma or grief-related contexts, where human moderators remain essential.
How does Lexic Pulse detect churn signals in real time?
Lexic Pulse's Active Listening Engine processes 100% of contact center conversations using AI models trained to detect churn language — phrases signaling frustration, competitor comparisons, cancellation intent, or unresolved friction. When a churn signal is detected in any conversation, a real-time alert is delivered to the retention team via Slack, CRM, or email — while the customer is still engaged, not days later when the interaction has been filed. Organizations using Lexic Pulse report detecting churn risk 30–60 days before it materializes in cancellations.
What results have Lexic Pulse clients achieved?
Lexic Pulse clients include Bankinter, Cellnex Telecom, Telefónica, TotalEnergies, Repsol, and Ecovidrio. Reported outcomes include: 40% reduction in support call volume within 4 weeks, generating over €60,000 per month in operational savings; more than 50% improvement in NPS scores through early friction detection across 100% of interactions; 60% B2B response rates on AI-moderated interview campaigns versus 2–8% for traditional surveys (Lexic Pulse deployment data, 2025–2026).
How quickly can Lexic Pulse be deployed?
Lexic Pulse connects to your existing call center, CRM, and ticketing infrastructure within 24–48 hours via native integrations with Genesys, Five9, Avaya, Zendesk, Intercom, Salesforce, and HubSpot. First actionable insights — churn signals, friction patterns, compliance flags — are typically visible within 7 days of deployment. No technical expertise is required from the client team; Lexic's onboarding team manages the full integration process.
What integrations does Lexic Pulse support?
Lexic Pulse integrates natively with: Call center platforms — Genesys, Five9, Avaya. Ticketing and chat — Zendesk, Intercom, WhatsApp. CRM — Salesforce, HubSpot, Pipedrive, Microsoft Dynamics. Collaboration — Slack, Microsoft Teams. Output — custom webhooks and email reports. Data can also remain within the client's own infrastructure via on-premise deployment, available for regulated sectors such as banking, insurance, and utilities.
Is Lexic Pulse GDPR compliant? Does it support on-premise deployment?
Yes. Lexic Pulse is GDPR-native, headquartered in Madrid and operating under EU data protection law from its core architecture. Every enterprise engagement includes a full Data Processing Agreement (DPA), ISO 27001 certification, and automatic PII redaction. No conversation data is processed through uncontrolled external LLM APIs. On-premise deployment is available for regulated industries — banking, insurance, utilities, healthcare — where data cannot leave the client's infrastructure.
What is the difference between the Active Listening Engine and the Proactive Listening Engine?
The Active Listening Engine is passive — it ingests and analyzes 100% of existing customer interactions (calls, tickets, emails, chats) without requiring any new customer touchpoint. It captures what customers are already saying. The Proactive Listening Engine is outbound — it deploys AI conversational agents via WhatsApp, phone, and RCS to initiate structured interviews with customers, leads, or partners, capturing what they would say if asked the right question in the right way. Used together, they form the Double Helix: complete visibility into spontaneous and solicited customer intelligence from a single platform.
How does Lexic Pulse handle compliance monitoring at scale?
Regulatory frameworks including GDPR, MiFID II, and the FCA's Consumer Duty increasingly require evidence of systematic customer monitoring — not spot checks. A human QA team reviewing 1% of calls cannot provide statistical confidence on regulatory adherence. Lexic Pulse reviews 100% of interactions against configurable compliance rubrics — detecting script deviations, consent violations, and vulnerable-customer handling issues in every conversation, continuously. Organizations in banking, insurance, and utilities use this to generate audit-ready evidence of full compliance coverage, protecting against regulatory exposure that 1% sampling structurally cannot prevent.
What is an independent AI agent audit, and how is it different from testing?
Testing checks whether an AI agent works under controlled conditions before launch. An independent AI agent audit evaluates real production conversations — what the agent actually said to real customers — against the Lexic Compass Trust Score, a 0-100 score built on four pillars: Integrity & Safety, Regulatory Trust, Operational Reliability, and Experience Trust (30/30/20/20 weighting). The audit ends in a signed verdict — Cleared, Cleared with Conditions, or Not Cleared — that a board or regulator can act on. Lexic Compass audits agents regardless of who built them, with no incentive to hide problems in a system it didn't sell you.
What does the EU AI Act require from AI agents that interact with customers?
From August 2, 2026, Article 50 of the EU AI Act requires that AI systems interacting with natural persons are clearly disclosed as AI, and that companies deploying them can demonstrate effective human oversight, decision traceability, and a full interaction audit trail. Obligations for systems classified as high-risk under Annex III roll out gradually through 2027–2028. Non-compliance with Article 50 carries fines of up to €15 million or 3% of global annual turnover, whichever is higher.
How does Lexic Compass audit an AI agent in production?
Lexic Compass scores 100% of an AI agent's real production interactions across four pillars: Integrity & Safety (security and ethics), Regulatory Trust (EU AI Act and GDPR compliance), Operational Reliability (accuracy and robustness), and Experience Trust (task completion and conversational quality) — combined into a single Trust Score (0-100) and a signed verdict of Cleared, Cleared with Conditions, or Not Cleared. The result is a documented audit trail with evidence for every finding, not a pass/fail badge.
How quickly can I get an AI Agent Audit?
Lexic Compass has three entry points. Flash Preview is free and delivers a preliminary Trust Score on 1-5 real conversations, typically within about 72 hours. Audit Sprint is the full engagement — representative sample, adversarial red teaming, and a signed executive verdict — delivered over 4 calendar weeks. Enterprise Continuous Trust adds a live Trust Score dashboard with ongoing monitoring for clients who need continuous, not one-time, assurance.
What happens if my company doesn't comply with the EU AI Act by August 2026?
Fines for Article 50 non-compliance can reach up to €15 million or 3% of global annual turnover, whichever is higher. But the more immediate risk is operational: an unaudited AI agent that gives a customer incorrect information, discloses data it shouldn't, or fails to identify itself as AI is a liability you cannot defend after the fact if you have no audit trail proving what happened.
Can Lexic Compass audit an AI agent built by another vendor or built in-house?
Yes. Lexic Compass audits any conversational AI agent already in production — voice or chat — regardless of who built it or what platform it runs on. Independence is the point: Lexic doesn't audit agents it sold you, so there's no incentive to soften findings.
What is Lexic Compass?
Lexic Compass is Lexic.AI's independent AI Agent Audit service. It scores a conversational AI agent's real production conversations against a Trust Score (0-100) across four pillars — Integrity & Safety, Regulatory Trust, Operational Reliability, and Experience Trust — and delivers a signed verdict: Cleared, Cleared with Conditions, or Not Cleared. It's available as a free Flash Preview, a paid Audit Sprint (4 weeks, full report), or Enterprise Continuous Trust for ongoing monitoring.
What is a Trust Score and how is it calculated?
The Trust Score is a 0-100 measure of whether a conversational AI agent deserves to operate in front of real customers, based only on evidence from real conversations — never self-reported. It combines four weighted pillars: Integrity & Safety (30%), Regulatory Trust (30%), Operational Reliability (20%), and Experience Trust (20%). A critical finding in Integrity & Safety or Regulatory Trust automatically forces that pillar to zero and the overall verdict to Not Cleared, regardless of the other scores — a strong average can never hide a live security or legal failure.
What does an AI Agent Audit cost?
Flash Preview, the entry-level diagnostic on 1-5 conversations, is free with no commitment. Audit Sprint, the full 4-week audit with a signed executive verdict, and Enterprise Continuous Trust, the ongoing monitoring retainer, are both priced to scope based on the number of conversations, channels audited, and regulatory exposure — Lexic does not publish fixed prices because that scope varies too much case to case.
Can Lexic Compass benchmark an AI agent's performance against human agents?
Yes, for clients who also use Lexic Pulse's Active Listening Engine on their human-staffed conversations. In that case, Lexic Compass can benchmark the AI agent's Trust Score and specific behaviors — resolution rate, tone, escalation appropriateness — against the documented baseline of the client's own human agents, a comparison no generic AI testing or evaluation tool can produce, since none of them has access to the client's real human-agent conversation data.