AI in Professional Services

Current, a national platform of almost 30 independent accounting firms, launched Tax AI through its partnership with Thrive Holdings and OpenAI. Its first pilot at partner firms for the 2025 tax year processed 7,000 returns, with average tax preparation time savings of 31 percent and up to 98 percent accuracy. Current is now helping the rest of its portfolio adopt the tool.

HSP GRUPPE, a network of independent tax, audit and law firms, runs ChatGPT Enterprise in a workspace shared with Kanzleipakt across 81 organizational groups, with shared agents for client communication and SKR03/SKR04 booking questions. From February 1 to July 14, 2026 it recorded 755 weekly active users and 500,000+ messages, and 98.6% of surveyed employees reported higher productivity.

Austrian trade union GPA deployed checkts.gpa.at, an AI chatbot for labour law questions, with 88% of users opening the cited paragraph and 12% clicking through to the federal legal information system.

ABC Legal Services, a legal document delivery company with 1,100 employees, runs more than 50 agents on Claude Managed Agents across service of process, eFiling, finance, marketing and compliance. Every agent is defined as code in git and deployed through pull requests. As of July 2026 about 310 employees use Claude daily, and some agents cut the cost of the tasks they cover by up to about 50%.

Wilson Sonsini, a law firm with more than 1,000 attorneys across 19 offices, compared AI contract review tools against a 90% accuracy target. According to Icertis's customer story, Vera Engage reached 95% on first-party contracts, 92% on third-party contracts and 94% on issue detection, and the firm now uses it within its Neuron Commercial service.

AArete's Doczy.ai, built on AWS with Amazon Textract and Amazon Bedrock, has processed 2.5 million contract documents (50 million pages) over 22 months, using smart chunking and dual clustering to turn contracts into structured data. AWS and AArete report 99% extraction accuracy, against about 55% for the earlier rules-based approach, 97% less manual time and about $330 million in client savings.

Redgrave LLP ran a head-to-head study of the generative AI workflow Relativity aiR for Review against active-learning managed review on about 45,000 documents, measured against an expert's blind sample. aiR for Review took about 18 hours versus 1,123, with higher recall and lower elusion but lower precision, 29 percent versus 39 percent.

Allegis Group, working with its systems integrator TEKsystems Global Services, rolled out Microsoft 365 Copilot to more than 18,000 active users, saving over 150,000 hours. It built ALTA on Azure AI Services to translate documents in minutes, saving an estimated $1.5 million so far, and a Copilot Studio PTO agent that cut case closure from 31 to 13 hours.