Customizing VOC Analysis Criteria

Deskroom now supports customizing VOC analysis criteria. You can train Deskroom AI directly so that VOC can be analyzed by your company's own criteria.

Why we made this

The same customer inquiry can be interpreted by different criteria from company to company, and it is often hard to capture that simply through predefined type names. This is also why, when using general-purpose AI like ChatGPT or Claude, the AI misses the real intent or the analysis results fail to connect with organizational decision-making.

Deskroom released this feature to help AI interpret customer inquiries according to each company's criteria, so that the results can be put to use in decision-making right away.

Use cases

When you want to break down a shared VOC type that is interpreted differently by each company

• You can split a "refund request" into "product dissatisfaction," "delivery issue," "promotion error," and so on, instead of treating it as a simple refund, so it can be analyzed to fit your criteria.

When you want the AI to understand expressions or terms used only by your service

• You can define "subscription cancellation" as "temporary pause," "benefit termination," or "at-risk churning customer," or register terms that exist only in your domain such as "non-member purchase" or "trial group refund," to improve analysis accuracy.

How to use

In the [Analysis Criteria] tab, select the VOC type you want, and you can directly enter analysis criteria that fit your company.

When you write which type values to use for each criterion, along with their definitions and examples, Deskroom's AI learns from them and analyzes customer inquiries from your organization's perspective.