DenmarkLogistics and transport2019 – ongoingSatair
Lilly, an AI assistant that answers quote requests by email
Satair's AI assistant Lilly handled 58% of quote requests that staff would otherwise have answered by hand
Problem
Satair, an aviation spare-parts supplier, received about six million emails a year and had around 230 customer service staff worldwide to answer them. About 450,000 of the emails asked for prices of products and services. Creating quotes by hand was slow and gave customers long lead times.
Approach
Satair worked with its partner Atos, which designed and built the assistant on Google Cloud after a six-week pilot that started in June 2019. Lilly reads incoming unstructured emails with the Natural Language API, pulls out product names and quantities, and sends quotes back by email. The customer service team follows Lilly's performance on a live dashboard.
Results
Takeaway
Satair put the automation behind the existing email channel, so customers did not have to change how they ask for quotes. Staff time went to customers with other needs.
Categories
- Size
- Large enterprise
- Industry
- Logistics and transport
- Business function
- Customer service
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Partner co-development
- Evidence
- Vendor case study
Sources
Source summaries
Google Cloud's customer story describes how Satair and Atos built Lilly, an assistant that reads quote-request emails and replies with quotes automatically. It gives the volume of emails and staff behind the problem, says Lilly automated 58% of quotes otherwise sent manually, and says its quote predictions were correct 98% of the time so far. The page carries no date; the Internet Archive first saved it on November 16, 2021.
Full details are in the original sources.
