(A) Tool for Learning Organisations

A tool that leverages employee feedback to drive sustainable, innovative growth and transform traditional organizations into learning organisations where employees learning and knowledge potential are maximized.

A

F

More

A

F

More

Bachelor Project

Innovation Strategy, 2023 - 2024

Introduction

In a rapidly evolving world, organisations must continuously adapt and learn. This tool helps them better utilize, share, and apply knowledge, fostering innovation and improving overall performance. Despite advanced technologies, individual insights are often overlooked, limiting valuable exchange and potential improvements.

Our project addresses this by developing a tool that collects, analyzes, and, most importantly, facilitates the sharing of feedback. By enabling organizations to respond effectively to changing environments, it supports their transformation into learning organisations—collaborative structures where individuals work continuously toward shared goals.

Problem Definition

Despite the availability of modern technologies, individual knowledge often remains unused, which inhibits valuable exchanges and improvements. Feedback is often collected selectively and not integrated into broader learning and improvement processes, leading to lost knowledge and missed opportunities for growth. The central question is: "How can individual knowledge be effectively used to address issues in process flows?" This highlights that feedback is not just a tool for continuous learning, but also a potential driver for innovation. In today’s fast-paced business environment, the ability to collect and implement constructive feedback is crucial for organizational success.

Hierarchical structures, common in large companies, often feature top-down decision-making, which can enhance efficiency but also limit employee involvement and hinder innovation. Effective feedback methods, such as personal feedback in meetings, surveys, and digital tools, are essential to improving innovation and competitiveness. Companies must establish strong feedback channels to avoid weak feedback cultures, which can harm performance. Research shows a clear connection between continuous, constructive feedback and higher employee engagement, ultimately driving greater profitability and success.

The diagram illustrates how feedback collection and analysis flows within an organization. As participation increases, it reveals critical problem areas with greater precision, enabling targeted interventions.

Objective

The objective of the tool is to break down the communication barriers prevalent in traditional organisation hierarchies and establish a feedback culture that enables bidirectional communication across all levels of the organization. This will allow problems to be identified and addressed at their source, and then communicated upward for further improvement. The ability to tackle difficult problems often depends on recognizing key leverage points. The tool helps companies identify and leverage these points by encouraging employees to share their individual knowledge and opinions, ultimately improving the overall performance of the organization.

Benchmarking

Benchmarking is a key part of our design process. It helps us understand the market for feedback tools, identify opportunities, and stand out. We analyzed various tools with different focuses to gain insights into their strengths and weaknesses.

Some platforms focus on performance management (Employee Engagement), while others emphasize recognition and appreciation (Employee Experience). Customer Experience tools also provided valuable parallels for feedback collection.

  • CultureAmp: A platform for employee experience, offering surveys, analytics, and action plans. It supports personal goal-setting and leadership development.

  • Leapsome: An all-in-one tool for employee engagement and development, including 360-degree feedback, performance reviews, and OKR management. However, its strong focus on evaluation may feel overwhelming.

  • OfficeVibe: Helps managers build better relationships with employees through weekly pulse surveys and anonymous feedback. It fosters a positive workplace culture but lacks strong development features.

  • 15Five: A platform for leadership and employee development with pulse surveys, continuous feedback, and "High-Fives" for recognition. However, it may add pressure on leaders due to performance tracking.

Positioning

Our tool acts like a predictive maintenance system for workplace culture, detecting patterns and signaling where action is needed. It not only collects feedback but also identifies underlying trends, enabling proactive decision-making.

Traditional surveys provide structured data but can overlook individual perspectives. Open-ended feedback, on the other hand, allows the organization to gather insights directly from employees' emotions and capture a wider range of feedback. Employees can share feedback anytime, creating a constant flow that reveals both immediate opinions and deeper, unspoken sentiments. Deep analysis processes these complex inputs efficiently, enabling a deeper understanding of workplace dynamics and fostering a culture of continuous improvement.

Proof of concept

To demonstrate the functionality of our concept, we developed an AI-powered system that structures data and extracts key themes. Instead of relying solely on prompting, we considered three advanced AI integration methods:

  • Fine-tuning an existing large language model (LLM) with our dataset.

  • Retrieval-augmented generation (RAG) using vector databases to improve contextual accuracy.

  • Topic modeling to identify recurring themes in large datasets.

Training a custom LLM was impractical due to high costs and expertise requirements. Fine-tuning and RAG provided better control over categorization. To refine classification further, we implemented BERTopic, a state-of-the-art topic modeling tool. Its process is showed in the diagrams.

  1. Embedding text into high-dimensional vectors using Cohere’s API.

  2. Reducing dimensions with UMAP for better processing efficiency.

  3. Clustering similar texts with HDBSCAN, which handles noise and outliers effectively.

  4. Extracting keywords to define clusters, generating topic descriptions.

  5. Refining topics with KeyBERT for clearer insights.

The final setup integrates a PostgreSQL database with Jupyter Notebook, enabling direct data import and processing. BERTopic structures the data, and the results are visualized on a web platform, allowing interactive exploration of categorized insights.

Concept

Gaining insights through feedback is a key driver of innovation, especially in diverse organizations. Studies by McKinsey (2019) show that companies with diverse teams can achieve up to 30% higher profitability. Feedback tools help capture different perspectives, fostering a broader range of innovations.

Individual Contributors (ICs) play a crucial role in this process. Regardless of hierarchy, they contribute directly to a company’s success by providing unique insights and innovative thinking. Feedback tools empower employees to share ideas instantly, fostering a more transparent and inclusive culture. This turns feedback into an iterative process of reflection and adaptation, actively driven by ICs.

The tool incorporates various features designed to promote knowledge exchange and organisational learning: discussion space, diverse committee, AI-guided questioning, content reporting, complaint handling, search functionality, profiles, identity protection, a knowledge repository, analytics, and sharing with only positive voting.

The application consists of three interconnected components.

System Architecture: The Three Components

The tool comprises three key components: "Feedback and Discussion", Search, and Analytics. While all users have access to "Feedback and Discussion" and Search, Analytics is reserved for a selected diverse committee that consists of employees from different hierarchical levels and age groups.

The diagram illustrates the connections between the application's components. There is an exchange between "Feedback and Discussion" and "Search" because the AI can reference discussions in search results. Additionally, both components are evaluated through analytics.

  • Search: Employees can find the information they need by asking the AI or searching within feedback and discussions. It stores all organized information from "Feedback and Discussion", ensuring easy access for future use.

  • Feedback and Discussions: This feature is crucial for the platform. Employees can provide feedback, including constructive criticism and emotion-driven feedback. On the discussion board, they can react to other feedback and contribute to collective problem-solving.
    Complaints are a part of the feedback system, allowing employees to anonymously report serious issues when they prefer not to disclose their identity.

  • Analytics: This component processes and analyzes collected data to identify patterns, trends, and insights. It supports strategic decision-making by interpreting data and enables the predictive maintenance system to function effectively.

Together, these components create a continuous cycle of knowledge sharing, feedback-driven improvement, and data-informed decision-making.

Design Process

The development of the tool followed a comprehensive design process:

  1. Initial Concept Development: Exploring various approaches to address the identified challenges.

  2. Morphological Analysis: Systematically analyzing potential solution components.

  3. Persona Creation: Developing user profiles to guide design decisions.

  4. Scenario Planning: Envisioning potential use cases and contexts.

  5. Feature Prioritization: Determining key functionalities based on user needs and project objectives.

  6. Information Architecture: Structuring the tool's content and features for optimal usability.

  7. Wireframing: Creating low-fidelity prototypes to visualize the tool's layout and functionality.

  8. User Testing: Gathering feedback on prototypes to refine the design.

  9. Visual Design: Developing a cohesive visual language, including icons, color scheme, and typography.

  10. UI Element Design: Creating consistent and intuitive interface components.

Final Design/ Branding

The final design concept of the application emphasizes user-friendliness, clarity, and versatility. The UI combines a modern, professional look with subtle playfulness. Clean lines and a simple, structured layout help users navigate with ease, while the use of feedback cards and expressive icons adds a human touch. These icons, representing concepts like communication, growth, creativity, and collaboration, visually embody the core values of the organization. Colors help distinguish the functionality of the different features. The "Search" feature is highlighted with warm tones of orange and red, creating an inviting and energetic feel. "Giving Feedback" is designed with calming greenish-white hues, evoking a sense of openness and positivity. For "Complaints/Reporting," a trustworthy purple color is used to signify anonymity and a safe space for confidential communication.

Typography plays a vital role in maintaining a neutral and accessible design. "Untitled Sans" and "Untitled Serif" were chosen for their simplicity and readability. These fonts, designed with accessibility in mind, ensure that information is communicated clearly without distraction. They reflect the application’s objective of providing functional and user-centered design, supporting a clean and professional visual language.

The app icon features the logo on a white background.

Four key screens showcase the final design of the application: the start screen and the three main ways employees can interact with the system—asking the AI, giving feedback, and reporting issues.

The analytics dashboard: Only members of a diverse committee have access to in-depth insights within the system.

Outlook and Next Steps

The next step is a pilot implementation in a company to assess its impact on change projects, starting with a team that has an established feedback culture. Expanding its capabilities with external data sources can provide deeper insights.

However, risks like algorithmic bias, misinterpretation, and over-reliance on AI must be mitigated through transparency, user training, and strong data security. These safeguards are already integrated into the tool’s foundation to ensure responsible and effective use.

Home

(A) Tool for Learning Organisations