DeepIP is a workflow-native AI platform for patent professionals, built to support work across the full patent lifecycle — from invention capture and drafting through office action responses, portfolio intelligence, and enforcement readiness. Its defining design choice is that the AI assistant lives inside Microsoft Word, fitting into the document-centric workflow patent attorneys already use rather than forcing them into a separate application.
Founded in 2024 by François-Xavier Leduc and Edouard d'Archimbaud — previously CEO and CTO of AI data company Kili Technology — DeepIP has raised $40M in total funding, including a $25M Series B in 2026 co-led by Korelya Capital and Serena. The company reports powering more than 10,000 patent applications and lists firms and corporate IP teams such as Greenberg Traurig, Mewburn Ellis, Dexcom, and Philips among its users.
The product combines large language models with retrieval-augmented generation over public and proprietary patent data to summarize complex inventions, define technical terms, and flag gaps in disclosures. DeepIP markets drafting and office-action-response time savings in the range of 50–70%, positioning the AI as a first-draft and review accelerator rather than a replacement for attorney judgment.
On security — which matters given the confidential invention disclosures handled in patent work — DeepIP states that it does not retain client data or use it for model training, encrypts data in transit, segregates customer and organizational data, and holds ISO 27001, SOC 2 Type II, and HIPAA certifications. Pricing is not published; the company offers a trial.
Hands-on review pending.
