344
Productivity & Workflow355
Automation & Workflow225
Software Development251
Marketing & Growth193
AI Infrastructure & MLOps175
Writing & Content Creation204
Data & Analytics142
Photography & Imaging156
Design & Creative170
Customer Support133
Sales & Outreach125
Voice & Speech135
Education & Learning131
Operations & Admin87
Legal operations leaders at Dentsu, IAG, ArcelorMittal and Shell describe how they use AI to save time, cut costs and manage new risks.
In short: More in-house legal teams are using AI to handle routine work faster, but they are running into new issues around oversight, training, and measuring savings.
Large companies are increasingly using AI to help their internal lawyers deal with everyday tasks like contracts, conflict checks, and invoice reviews. The Financial Times spoke with legal operations leaders at Dentsu, IAG, ArcelorMittal, and Shell about what is working and what is not.
At Dentsu, lawyers use a legal work platform to track requests and reuse contract information. They also build many small internal tools, including one that uses AI to spot possible conflicts of interest across a large set of contracts. Dentsu estimates about half of its lawyers save 4 to 8 hours a week, but the company also sees problems when people build similar tools twice or do not follow consistent methods.
At insurer IAG, AI is used to reduce admin work, including reviewing legal bills. The team uses software to flag invoices that look unusually high, which is important when you receive about 50,000 invoices a year. IAG says the spending on this software has paid back by more than 800 percent in a little over a year.
ArcelorMittal is using contract software that lets staff insert pre-approved contract wording, which can speed up purchasing. Shell is using AI from Microsoft and Thomson Reuters for tasks like drafting and summarising, and estimates time savings of 15 to 20 percent as part of a wider plan to cut legal spending.
Two challenges keep coming up. One is control, because lawyer-built AI tools can multiply quickly without enough oversight (like everyone making their own spreadsheet and no one knowing which one is correct). The other is budgeting, since some vendors may charge based on usage, which can be harder to predict than a yearly license.
Source: Financial Times