What AI Actually Does During an Intake Call (Plain English)

What AI Actually Does During an Intake Call (Plain English)

There is a lot of confusion about what AI does and does not do during a legal intake call. Some of it comes from AI companies overpromising. Some of it comes from firms that are skeptical of technology and imagine something more autonomous and intrusive than what actually exists. Most of it comes from a lack of plain-language explanation.

This article describes exactly what eNZeTi’s AI does during a live intake call, in plain English, without the marketing language. It also explains what it does not do, which is just as important.

First: What the Coordinator Still Does

The coordinator handles the entire call. They answer the phone, greet the caller, ask the questions, listen to the answers, handle objections, explain fees, and close for commitment or a scheduled appointment. None of that is automated. None of that is removed from the coordinator’s responsibility.

The caller interacts only with the coordinator. They do not know the AI system exists. Their experience is of a human conversation with a human being who happens to be exceptionally well-prepared and calm.

The AI does not speak. It does not interrupt. It does not generate responses that the coordinator reads aloud from a script in real time (though prompts can suggest language). It listens, detects, and surfaces information. The coordinator decides what to do with it.

What the AI Actually Does: Step by Step

Step 1: It Listens

The AI listens to the call audio in real time. It is transcribing the conversation as it happens, converting speech to text. This transcription is happening in the background, invisible to both the coordinator and the caller.

The transcription is not stored in any way that would compromise client confidentiality under applicable professional rules. This is a compliance matter that eNZeTi handles at the infrastructure level, not something coordinators need to manage manually.

Step 2: It Detects Patterns

Against the live transcription, the AI runs pattern detection. It is trained to recognize specific linguistic patterns that indicate specific call states:

• Price objection patterns: language like “I can’t afford,” “how much does this cost,” “I’m worried about money”
• Hesitation patterns: language like “I need to think,” “I want to talk to my husband,” “I’m not sure”
• Qualification gaps: tracking whether key qualifying questions have been asked and answered
• Urgency signals: calendar language, statute of limitations concerns, court date mentions
• Emotional state signals: voice tone analysis, speech rate, language that indicates distress or fear

The pattern detection is probabilistic. It identifies situations where a particular response type is likely to be useful, not situations where it is certain. The coordinator’s judgment mediates every prompt.

Step 3: It Surfaces Prompts

When a pattern is detected, a prompt appears on the coordinator’s screen. The prompt is brief, specific, and actionable. It does not deliver a paragraph of coaching. It surfaces what the coordinator needs to know or say right now.

Examples of what a prompt looks like:

• “Price objection detected. Contingency fee explanation: [scripted language appears on screen]”
• “Hesitation signal. Try: ‘Is there something specific giving you pause? I’d rather address it now.'”
• “Missing: statute of limitations check. Case date was [X]. Deadline: [calculated date].”
• “Key qualifier not asked: insurance status of other party.”

The coordinator sees this prompt on their screen. The caller does not see or hear it. The coordinator reads it, makes a judgment about whether and how to use it, and proceeds.

Step 4: It Scores the Call in Real Time

As the call proceeds, the AI updates a running quality score. The score tracks how many key criteria have been addressed: opening quality, empathy demonstration, qualification completeness, objection handling attempts, close attempt, and others. The score is visible to the coordinator as the call progresses.

A coordinator who can see that their qualification score is incomplete at minute eight will ask the missing questions before minute twelve. Without real-time scoring, they may not realize the omission until the call is already over.

Step 5: It Generates a Post-Call Summary

After the call ends, the AI generates a summary: caller information extracted from the conversation, key facts gathered, score breakdown by criterion, and a flag list of anything missed. The coordinator reviews this summary, corrects any transcription errors, and logs it to the CRM.

This summary replaces manual note-taking during the call. The coordinator who does not have to write while listening can listen better, ask better questions, and maintain better eye contact with their own screen. The quality of attention they give the caller improves.

What the AI Does NOT Do

Clarity about limitations is as important as clarity about capabilities. Here is what eNZeTi’s AI does not do:

It does not replace the coordinator. There is no scenario in which the AI handles the call. The coordinator is always on the line. The human relationship is always present.

It does not make decisions. The AI surfaces information and prompts. The coordinator decides whether and how to use them. A coordinator who disagrees with a suggested prompt ignores it. The AI does not escalate, override, or control.

It does not guarantee compliance. The AI cannot provide legal advice and does not attempt to. It surfaces process prompts, not legal guidance. Coordinators still need training on what they can and cannot say under applicable professional rules.

It does not learn the wrong things. The system is trained on intake conversion patterns, not on improvised language from your least effective coordinators. The pattern library reflects best practices, not the average of whatever your current team does.

It does not record or store calls in ways that create confidentiality risk. The legal and compliance architecture of how call data is handled is built to protect client confidentiality in accordance with applicable professional conduct rules.

What the Coordinator’s Screen Looks Like

Picture a dashboard on a second monitor or a browser window open alongside the coordinator’s CRM. The dashboard shows:

• The live transcript scrolling as the call proceeds
• A real-time call quality score in the corner
• A checklist of key qualifier items (green when addressed, amber when partially addressed, red when missing)
• A prompt box in the center where coaching prompts appear when detected
• Timer showing call duration

The coordinator glances at the dashboard periodically during the call. It is peripheral support, not a primary focus. The primary focus remains the caller’s voice and the conversation.

What Changes When This Is in Place

The coordinator who has AI support during the call does not think less. They think more clearly, because they are freed from the cognitive load of remembering every script, every qualification criterion, every possible objection response simultaneously. The mental bandwidth that was consumed by trying to remember everything gets redirected to actually listening to the caller.

The result is a coordinator who is more present, more responsive to emotional nuance, and better at the human dimension of the call, precisely because the AI is handling the memory and tracking functions. The human gets better at being human. The AI handles the administrative load.

That is augmentation. Not replacement.

Learn More

To see exactly what the eNZeTi dashboard looks like and how real-time coaching integrates with your coordinators’ existing workflow, visit enzeti.com.