Customer Service Automation
SLTN’s five-person HR team was overwhelmed by a steady volume of basic and sensitive employee questions, and employees sometimes waited two to three days for HR answers that should have been available instantly.
Reported impact
SLTN gave employees instant, privacy-aware HR answers and actions, reduced pressure on its HR team, and created a governed AI showcase it can also use to demonstrate value to clients.
SLTN built Sophie, an AI-powered HR chatbot, on IBM watsonx so employees could receive personalized HR support without exposing sensitive employee data. The team started with watsonx.ai and watsonx Assistant to create a proof of concept that answered questions about employment conditions, then connected the bot to the core HR system so it could return employee-specific answers such as remaining vacation days and also submit transactions like vacation requests. Sophie was rolled out company-wide through Microsoft Teams and Active Directory, which gave the assistant context about the user while IBM watsonx.governance supplied the privacy and governance controls SLTN considered essential for responsible use. Beyond self-service support, the assistant’s analytics help HR see which policies or workflows need improvement, creating a feedback loop between employee questions and better people management. The result is a 24x7 HR channel that handles routine work in the background so human HR staff can spend more time on the personal, high-empathy issues where they matter most.
Employees ask HR questions in Teams, Sophie uses watsonx Assistant and watsonx.ai to interpret them, queries the HR system and identity context through Active Directory, returns a personalized answer or transaction result, and logs query patterns that HR can use to improve training and policies.
“By integrating with Microsoft Teams and Active Directory, the chatbot delivers personalized, context-aware responses”
“IBM watsonx.governance gives us the tools to maintain data privacy”
“the chatbot’s large language model (LLM) was trained to answer questions based on SLTN’s employment conditions.”
“the productivity and satisfaction of employees are sometimes impacted due to the two-to-three-day response times”
“using IBM watsonx.ai AI studio and IBM watsonx Assistant to develop Sophie”
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
IBM Sample May 19, 2025 Inspire Inspire needed to reduce the manual effort in routine HR tasks so employees could complete leave, expense, and insurance requests quickly and the company could offer a more modern employee experience to attract and… 15% boost in productivity, driving business growth and efficiency IBM watsonx Assistant · IBM watsonx.ai · Meta Llama 3 IBM Sample Feb 11, 2025 DI Square DI Square needed to improve call-center operator productivity and reduce dependence on individual know-how as labor shortages and higher staffing costs made it harder to hire, train, and retain operators. IBM Granite · IBM Watson Discovery · IBM watsonx Assistant IBM Sample Nov 6, 2024 Humach Humach wanted to evolve its contact-center offering so digital assistants and human agents could work together more effectively, because rising customer expectations required faster, more contextual resolutions without… 18% increase in overall customer satisfaction Amazon Web Services · Deepgram · IBM watsonx Assistant 