For all that's holy, can you just leverage the Web, please? Author: Thomas Steiner Date: September 3, 2025 Tags: Life, Technical --- Story Summary In 2005, Thomas and his wife Laura moved in together in Barcelona. They had an old washing machine that broke, leading to a temporary fix using a chair. Later that year, they bought an Electrolux washing machine that lasted nearly 20 years across several moves. Recently, the machine’s motor failed, and Thomas ordered a new Electrolux washing machine online — this model notably lacks Internet or WiFi connectivity. Warranty Registration Experience The new washing machine included a voluntary 10-year warranty if registered. The brochure featured a phone number and a QR code for registration, but: The QR code was a plain-text phone number (not a clickable tel: link). Calling the number was frustrating due to high volume, dropped calls, and eventual inbound capacity limits. An SMS offered a chat service URL — however, the link was broken and insecure (invalid SSL certificate). After manual fixes, the chat page still did not work. Thomas used Google to search "electrolux warranty register" and found a promising registration page behind a login. Registration was painless, allowing input of the washing machine’s product number via: Manual entry of the product number from the identification plate. Uploading a photo of the identification plate — an AI service extracted the product number automatically, making registration extremely easy. Thomas reflects on the poor initial user experience versus the excellent AI-enabled web solution. --- Key Observation "Why they didn't just put this URL on the brochure and the QR code is beyond me." Thomas advocates for companies to leverage the Web as the default for such straightforward tasks. He notes telephone-based support is often cumbersome and expensive. In 2025, digital-first solutions should be standard, with phone support as a fallback. --- Bonus: AI Product Number Extraction Demo Thomas works on built-in AI at Google Chrome. He built a demo using the Chrome Prompt API to identify product numbers from photos of identification plates. The demo processes a multimodal input (text + image) and extracts a 9-digit product number using regex constraints. Demo Highlights: Create a LanguageModel session, specifying expected inputs (English text and images) and outputs. Send a prompt streaming request with both text instructions and an image. Use a regex /\d{9}/ to enforce output format. Process response stream chunks in real-time. The demo showcases how client-side AI can enhance workflows like warranty registration by automating tedious tasks. Code snippet (overview): Interactive demo available: AI Product Number OCR Demo --- Related Links & Info Blog GitHub Repository for editing: Edit on GitHub Contact Author: Thomas Steiner License: Creative Commons BY-SA 4.0 ISSN: 2941-0452 --- Final Thought Companies should prioritize web-based,