AI Signal 603 2 feeds carried it
Thomson Reuters stays with Anthropic despite building proprietary AI for legal, tax, and compliance
↗
Thomson Reuters developed its own AI model trained on proprietary professional data for legal, tax, and compliance work, but continues to use Anthropic's models instead.
Even companies with domain-specific proprietary data and the resources to train custom models may find third-party AI more practical or effective. The decision highlights that owning relevant training data does not guarantee a superior build-versus-buy outcome for enterprise AI.
Written by elseif from the cluster below · every claim links back to a sourceThe three things worth knowing
Thomson Reuters trained a custom AI model on its proprietary professional data for legal, tax, and compliance applications.
Despite having an in-house model, the company continues relying on Anthropic's AI.
The case illustrates that domain-specific data alone may not justify replacing an established third-party model.
THE CLUSTER