Articles grouped by business problem, industry, and system type.
An AI model suggests catalogue parts from a customer's description as a list of candidates to check. When two variants differ by a detail the description lacks, for example the connector, the system points out the gap and leaves the choice to the salesperson. Check search with synonyms first.
A custom spend classification model makes sense when supplier-based rules and your current purchasing system confuse different kinds of purchases from one supplier. Before training, agree the category split and check whether line item descriptions let you tell the categories apart, and settle the choice with a test on new line items.
A custom model for defects in photos makes sense when the image shows what the inspector judges, and a ready-made vision product misses defects or flags too many good parts. First check the lighting and the number of physical parts in the photos. At 20,000 photos a month, a ready-made vision product cuts errors from 3% to 2% and fine-tuning to 1%, which together saves 20 hours of corrections, or €698 a month.
Fine-tuning for transport documents makes sense when the image is legible and the model keeps mixing up fields, for example the loading date with the delivery date. First separate image, reading and assignment errors, because a poor photo is fixed by a legibility gate and a request for a new shot. At 3,000 documents a month, a fall in wrong fields from 8% to 3% saves 10 hours of corrections, or €349.
Fine-tuning helps classify tickets when a general model regularly confuses the company's specific categories. First check whether people understand the categories the same way and whether past assignments show the correct decision. If tickets went to whichever team had time, the history teaches the model who was available.
Private models
We select the hardware and model, deploy it and measure quality on your examples. You can start with a computer you already own.
Times are shown in your own time zone. We work with clients across time zones.
See private modelsOpenAI makes ChatGPT and Anthropic makes Claude. We belong to the partnership programs of both companies, so we work with their tools every day and learn from the creators themselves.
Claude Partner NetworkParticipation in programs does not mean that OpenAI or Anthropic endorses our services.