United KingdomTechnology and telecom2025 – ongoingMatillion
Maia, an AI assistant that builds data pipelines
Matillion's Maia cut one customer's complex pipeline work from 40 hours to 1
Problem
Building data pipelines on Matillion's platform took skilled engineers many hours, and business analysts could not do it themselves. Matillion wanted an assistant that builds and maintains pipelines from plain-English requests.
Approach
Matillion built Maia on Claude after testing several model providers for 18 months. It used three models in its test setup: one running Maia, one acting as tester and one judging results. Claude handled the company's own YAML-based pipeline language without special training. Maia became generally available in June, after two years of work that began with auto-documentation. Matillion's own developers also use Claude Code.
Results
Takeaway
Matillion tested models side by side for 18 months before choosing one. The company reported results for customers and for its own engineers.
Categories
- Industry
- Technology and telecom
- Business function
- IT and R&D
- Innovation type
- Product and service innovation
- AI technology
- AI agentsGenerative AI
- Adoption
- Off-the-shelf tool or platform
- Evidence
- Vendor case study
Sources
Source summaries
Anthropic's customer story describes Matillion's Maia assistant, built on Claude, and the company's internal use of Claude Code. It reports one customer's transformations falling from 40 hours to 1 hour, 80% of Matillion developers using Claude Code and a 41% average cut in pull request time. The page carries no date; the Internet Archive first saved it on November 26, 2025.
Full details are in the original sources.
