Employee open-ended responses are a treasure trove. Analyze them with AI and visualize instantly.

Many companies gather employee opinions on business strategy, internal systems, compensation, and other topics through employee surveys. The results collected in this way are essentially a "treasure trove." However, it seems many companies are letting this valuable data go to waste.

To address this, Dentsu Inc. Business Design Square (BDS) developed "VoiScope, " a service that uses AI to analyze open-ended responses from employee surveys and visualize issues. Yusuke Sakamoto of BDS explained the service's features and use cases.

 

Are you leaving employee survey open-ended responses as-is?

I have been involved in organizational transformation projects for various companies. What I noticed was the problem that "employee surveys are not being utilized for management reform." There are far too many cases where the candid opinions employees write in the open-ended response sections of surveys are not being fully leveraged.

Today, many companies are under pressure to pursue DX (Digital Transformation), innovation, and corporate transformation. To create new value, the need for deeper collaboration between corporate management and employees is growing ever more critical.

Furthermore, it has become clear that employee engagement correlates with labor productivity and turnover rates, and that "increasing employee engagement boosts corporate growth rates." In other words, the level of engagement between corporate management and employees is becoming increasingly crucial. Reflecting this trend, projections indicate that Japan's employee engagement market (※), valued at 6 billion yen in 2020, is expected to double by 2023.

※＝Employee Engagement Market
Employee engagement refers to employees' voluntary willingness to contribute to their company. It is a concept representing the strength of the bond between employees and the company, indicating whether they are in a mutually beneficial relationship where both contribute and grow. The employee engagement market refers to products and services that support the measurement and improvement of "engagement."

However, on the other hand, partly due to the impact of the novel coronavirus, management has expressed concerns that
"It's become harder to see what's happening on the front lines"
"We see employees' faces less often, making it harder to gauge the atmosphere."
Against this backdrop, BDS developed "VoiScope," which analyzes employee opinions using AI to boost engagement. A beta version is currently being trialed, with a fully-featured official release planned for 2021.

Find hints for solving organizational challenges from AI classification results

I have always believed that "the clues to solving organizational challenges may lie within the company itself."

Employee feedback is precisely one such source. However, simply dumping tens of thousands of raw data points onto a table with a "Here are the opinions we've received" approach makes finding those hints difficult. VoiScope visualizes the opinions and ideas submitted using AI, transforming each employee's voice into material for management decisions.

VoiScope visualizes employee opinions in three main steps.

STEP 1: Input open-ended responses from employee surveys

First, compile the open-ended responses from employees regarding their opinions and requests for the company. You can use data from past surveys as well as conduct new employee surveys.

STEP2 Analyze with AI

Simply tabulating the collected employee opinions as-is makes it difficult to effectively utilize them for management. Therefore, we use AI to classify the data, categorizing the opinions by subject matter and sentiment (positive/negative), processing them quantitatively and structurally.

STEP3 Visualize Analyzed Opinions Clearly

The analyzed data is also presented in a dashboard format, enabling an overview of the company-wide situation. Tables and graphs visualize key insights: which categories (e.g., "Organization," "Supervisors," "Systems/Processes") receive the most opinions, and the positive/negative ratio for each opinion.

Users can filter by gender, age group, position, or office location, search keywords within raw data, track data trends over time, compare data between teams, and view departments with high negative feedback ranked in a list format.

Capturing this data with VoiScope brings organizational challenges into sharp focus. Comparing teams with high positive feedback to those with high negative feedback, and hearing what initiatives the former are implementing, can provide valuable insights for solving company-wide issues.

         

Multiple companies are considering implementation. Applications for customer service are also being explored.

Kracie Group is assisting with beta version development. Generally, the larger a company grows, the greater the distance between management and employees. Moreover, executives are extremely busy and rarely have time to listen to all employees' opinions. Even so, VoiScope enables quantitative and structured viewing of all employees' opinions. We are pleased that Kracie's management recognized this point.

As an aside, when analyzing with VoiScope, I sometimes get the sense that I can grasp the "emotion" or "temperature" of the subject being analyzed. The intensity of each individual's voice seems to deepen my understanding of the organization beyond what can be captured by mere numbers.

Additionally, Dentsu Inc.'s Public Relations Bureau is considering implementing VoiScope. Previously, they stored and viewed the results of internal surveys separately. With VoiScope, they realized for the first time that "this wasn't accumulating data," identifying it as an issue.

Following the release announcement, we've received inquiries about implementation from several other companies. The next challenge is customizing AI training for each company. Every organization has its own unique phrasing and specialized terminology aligned with its corporate culture. We are currently working on ways to ensure the AI accurately classifies opinions and emotions, and we will continue refining this capability.

Furthermore, we plan to expand the service's functionality to enable more multifaceted issue discovery. This includes integrating not only employee feedback but also customer voices.

VoiScope Release:
Developed "VoiScope," a service that analyzes employee opinions with AI and visualizes challenges