Our customer wanted to reduce the load on their IT support team and proposed creating a chatbot capable of identifying hardware models and software issues from smartphone or webcam photos. Once the hardware is recognized, the user can describe their issue, and the bot takes over from there – guiding them through troubleshooting and providing solutions automatically.
Challenge
As a large corporation, our client had thousands of users experiencing frequent hardware and software issues. With high manpower costs and an overworked IT helpdesk, they needed a chatbot capable of resolving many problems autonomously.
Solution
We developed a chatbot web interface functioning as AI assistant for IT Support that asks users to upload a picture of the affected hardware or specify the software in use. Once identified, the workflow guides the user to describe the issue. The bot then searches the internal knowledge base – aggregated from dozens of internal and online sources—to pinpoint the problem and provide a solution.
If no solution is found, the user is offered the option to automatically create a Zendesk ticket. The ticket includes an AI‑generated summary of the conversation, the problem description, and any uploaded images. We also built an internal Power BI dashboard to track chatbot usage and success rates.
Results
The bot identifies hardware with 95% accuracy, and new hardware can be trained at any time. It delivers precise solutions whenever the issue exists in any of the indexed sources. If not, the helpdesk resolves the ticket, and the new solution becomes part of the bot’s knowledge base – making it self‑learning. Overall, the helpdesk workload has been reduced by 50%.