Case Studies

Field inspections require adjusters to carefully observe and document conditions while working on-site. Traditionally, adjusters either stop their inspection to manually enter information into digital forms or make notes and complete the form later, relying on memory and handwritten or voice notes.

Speech2Form introduces a more natural and efficient approach. The AI-powered solution allows adjusters to simply speak about what they observe during a claim inspection, while the application intelligently converts those spoken observations into structured inspection form data.

Using AI, Speech2Form understands natural-language descriptions, identifies relevant information, maps observations to the appropriate inspection fields and supported values, and presents the suggested updates to the adjuster for review before applying them to the form.

The result is a more efficient inspection experience that allows adjusters to focus on the inspection rather than the form.

The Challenge

Claim inspections require adjusters to capture a significant amount of information about the property and observed conditions.

Traditional digital inspection workflows often require adjusters to:

• Stop their inspection to enter information into a form
• Navigate between multiple fields
• Select values from dropdowns and checkboxes
• Remember observations until they can enter them later
• Make notes and manually transfer them into the inspection form
• Repeat this process throughout the inspection

This creates friction in the field. When an adjuster is focused on documenting roof conditions, measurements, damage, materials, and hazards, repeatedly interacting with a mobile device can interrupt the inspection process.

There is also a risk that information may be forgotten, entered incorrectly, or interpreted differently when the form is completed later from memory.

The goal was to create a more natural interaction model that allows adjusters to describe what they see and let AI handle the form-filling process.

The Solution

Rheal developed Speech2Form, an AI-powered solution that converts an adjuster's spoken inspection observations into structured form data.

Instead of requiring the adjuster to interact with individual form fields, the adjuster simply taps the microphone and speaks naturally, just as they would describe their findings to another colleague.

For example, an adjuster could say:

"Cause of loss is wind, roof condition's fair, architectural shingle, medium pitch, drip edge's present but not painted, gutters are damaged on the front slope."

Speech2Form analyzes the spoken observation and identifies the information relevant to the inspection form.

Key Features & Capabilities

1. Natural Voice-Based Data Entry

Speech2Form allows adjusters to interact with the inspection form using natural speech rather than manually entering every field.

The adjuster can speak continuously about their observations without needing to follow the exact order of the form.

This provides a more natural field experience and allows the adjuster to maintain their focus on the physical inspection.

2. AI-Powered Inspection Understanding

The spoken transcript is passed to Claude AI, which analyzes the content and identifies information relevant to the inspection form.

Rather than simply converting speech into text, the AI determines:

• What the adjuster is observing
• Which inspection field the observation relates to
• What value should be assigned
• Whether the observation represents a positive or negative condition
• Whether the information is sufficiently clear to be applied

This moves the solution beyond traditional speech recognition into intelligent form interpretation.

3. Intelligent Field Mapping

One of the key capabilities of Speech2Form is its ability to map natural-language observations to the exact fields and supported values available in the inspection form. This allows adjusters to speak naturally without having to remember the structure or terminology of the underlying form.

4. Review Before Applying Changes

A critical design principle of Speech2Form is that AI suggestions are never silently written to the form.

After analyzing the transcript, the application presents a Review Sheet showing the fields that AI believes should be updated.

The adjuster can review the suggestions before anything is committed.

• Human oversight
• Transparency
• Error prevention
• Greater confidence in AI-generated updates
• Control over the final inspection data

The adjuster can accept the suggested changes, reject them, or continue narrating to provide additional information.

Business Benefits

• Reduced Form Entry – Adjusters spend less time manually navigating inspection forms and entering individual observations.
• Improved Field Productivity – Adjusters can focus on inspecting the property instead of repeatedly interacting with their mobile device.
• Natural User Experience – The system allows adjusters to describe their observations naturally rather than forcing them to communicate using rigid form terminology.
• Reduced Data Loss – Important observations can be captured closer to the time they are made, reducing reliance on memory and later manual entry.
• Improved Data Accuracy – AI maps observations to predefined form fields and values, while validation and review mechanisms help prevent incorrect updates.
• Human-in-the-Loop AI – AI does not make silent changes to inspection records. Adjusters review and approve suggested updates before they are applied.
• Scalable AI Foundation – The Speech2Form architecture provides a foundation for extending voice-driven workflows to other inspection forms and claims-related processes.

Conclusion

Speech2Form represents a new interaction model for field inspections—moving from form-driven data entry to conversation-driven data capture.

By allowing adjusters to simply describe what they see and using AI to translate those observations into structured inspection data, the solution reduces interruptions, minimizes repetitive form entry, and helps adjusters maintain their focus on the inspection itself.

The combination of natural-language understanding, intelligent field mapping, discrepancy detection, negative-observation handling, and human review makes Speech2Form more than a speech-to-text solution. It is an AI-powered assistant that helps transform an adjuster's natural observations into actionable inspection data.

As part of RhealCMS's broader AI strategy, Speech2Form demonstrates how AI can be embedded directly into the day-to-day workflows of insurance professionals to improve productivity, data quality, and user experience.