COALA mapped on Significant innovations and lessons learned, training aspects

• Pioneering application of an open, privacy-focused digital assistant in manufacturing. The COALA Digital Intelligent Assistant (DIA) core software will base on the privacy-focused open assistant Mycroft, addressing industry-grade security, privacy, and ethics needs.
• Introduction of digital-assistant-mediated prescriptive quality analytics using manufacturing resource data. The PREVENTION (PREscriptiVE aNalyTIcs for quality optimizatiON) service of COALA will address Prescriptive Quality Analytics.
• Proof-of-concept for the effective augmentation of prescriptive quality analytics via a voice-first interaction.The DIA will use voice as a primary medium (voice-first) to receive user input and provide responses – this speeds up the interaction, is hands-free, and in general an advantage over, oftentimes, overloaded graphical interfaces.
• Pioneering use of an explanation engine with a focus on manufacturing analytics. Fundamental research and the prototype development of the so-called “WHY engine” will enable the assistant to explain its responses (e.g. predictions and advices).
• Introduction of a novel performance measurement procedure (“X”-WHY test) for the explicability of AI-systems, such as digital assistants in manufacturing. It identifies how many WHY questions the user expects to ask in a scenario in a row. Then, the test measures how many questions in a row the assistant is capable of answering.
• Joint application of a Machine Avatar (Digital Twin). This feature is able to track and manage the information gathered by each IoT-connected machine with a Product Avatar, able to track the lifecycle of each product.
• Availability of machine and product information through avatars in sharable environment, independent of the data format and totally product-agnostic.
• A cognitive advisor with dedicated human-assisted AI methods for enabling transfer of tacit knowledge of experts to novice workers.
• Introduction of a new didactic concept to teach workers AI competencies.
• Pioneering use of digital assistant for on-the-job training with a voice-first interface.
• Availability of a new concept for tacit knowledge transfer based on small training datasets.
• Change management process to prepare the adoption of digital assistants at work.
• Pioneering use of an explanation engine with a focus on on-the-job training.

• COALA highly involes in the Digital Innovation Hubs (DIHs) and other regional innovation infrastructures to address SMEs issues that may not be able to afford the education of their labor force, through COALA's didactic concept or related services.
• A cost/benefit analysis will be included in the exploitation plan to verify the affordability for SMEs.

Evaluation of the use case implementations will provide lessons learned and recommendations for further development and research of the COALA solution.

The COALA vision for AI in manufacturing is the development of human-centered digital assistant that provides a more proactive and pragmatic approach to support operative situations characterized by cognitive load, time pressure, and little or zero tolerance for quality issues. COALA will help shaping the complementarity in the collaboration between the AI-based assistant and the human so that

• the AI will take over time consuming and stressful tasks reliably and credibly, while
• the human will focus on understanding and problem solving in complex, knowledge-intensive situations.

COALA’s AI-focused education and training concept will prepare the human-side of the collaboration by offering concept for teaching professionals systematically and, in the language of the workers, about the capabilities, risks, and limitations of AI in manufacturing. The COALA solution will transform how workers perform their jobs and it allows companies to maintain or increase the quality of their production processes and their products.