Big Four Professional Services Firms Activate Agentic AI at Scale

In partnership with

The Latest Agentic Development

 Big Four Professional Services Firms Activate Agentic AI at Scale

The Big Four accounting and consulting firms collectively brought agentic AI platforms to production or near-production status by year-end 2025:

• EY has deployed 1,000 AI agents into production across its organization (as of 2025) with plans to scale to 100,000 by 2028; 80,000 tax staff now have access to 150 AI agents for data collection, document review, and compliance work; annual AI investment exceeds $1 billion.

• PwC deployed Agent OS across client operations with 25,000 intelligent agents executing tasks.

• KPMG launched Workbench with 50 AI agents and chatbots in production, with nearly 1,000 more in development.

• Deloitte integrated AI capabilities into its Omnia audit platform with enhanced document review and financial statement analysis tools.

Strategic Implication: The Big Four's collective move to production-scale agent deployment signals that agentic AI is now table stakes for enterprise operations. Staffing models are shifting from humans producing first-draft work to humans overseeing AI agent execution, fundamentally reshaping professional services labor economics.

BigBear.ai Finalizes $250M Ask Sage Acquisition: Mission-Critical Agent Infrastructure

BigBear.ai completed its acquisition of Ask Sage on December 31, 2025, for $250 million in cash, bringing together two agentic AI platforms for government and defense. Ask Sage currently serves 100,000+ users across 16,000 government teams and hundreds of commercial organizations, offering secure generative AI capabilities, orchestration, and agentic features within the strictest security and compliance frameworks. The combined platform positions BigBear.ai as a trusted provider of "mission-ready AI" that moves from experimentation to operational impact.

Strategic Implication: This consolidation reinforces agentic AI's shift from proof-of-concept to production-scale deployment in mission-critical government and defense environments, where security, sovereignty, and compliance are non-negotiable.

Thomson Reuters Launches CoCounsel Agentic Workflows: Legal Agent Execution Goes Live

Thomson Reuters is rolling out CoCounsel's next-generation agentic capabilities starting in early 2026, with autonomous document review workflows and customizable multi-step legal processes. The platform now supports bulk document review of up to 10,000 documents and enables legal professionals to initiate complex workflows through natural language prompts, with CoCounsel independently planning and executing steps using Westlaw and Practical Law authoritative content. CoCounsel Document Analysis specifically launches in January 2026 for auditors and accounting professionals, automating complex audit workflows at scale.

Strategic Implication: This represents the first major-market deployment of autonomous multi-step agent workflows in a regulated professional services domain, demonstrating that agentic systems can operate at enterprise scale with audit trails and accountability.

LexisNexis Unveils Next-Generation Protégé General AI: Multi-Agent Legal Infrastructure

LexisNexis rolled out the next-generation Protégé General AI in early December 2025, featuring a new agentic infrastructure with four specialized agents that collaborate on complex legal workflows: an Orchestrator Agent (coordinates multiple agents), Legal Research Agent (decomposes prompts into legal questions), Web Search Agent (integrates open web insights), and Customer Document Research Agent (autonomously analyzes firm documents). This represents the most integrated legal AI solution yet, unifying authorized legal content, customer documents, and open web data within a single secure workflow environment.

Strategic Implication: The multi-agent orchestration model demonstrates how agentic systems can move beyond single-task automation to complex, cross-functional reasoning—a blueprint for enterprise-wide agent deployment.

Agentic Access Management becomes a “real” security category

Identity security vendor Oasis says 2025 was the inflection point where AI agents + non-human identities (NHIs) blew past human-centric governance models—and positions its Agentic Access Management (AAM) as purpose-built for intent-aware, continuously enforced access controls for autonomous agents. The notable move here is the pairing of a product category (“AAM”) with a governance framework to push security teams toward standardized controls (policy, accountability, auditability) before agents scale inside enterprises.

Meta scoops “general agent” startup Manus for ~$2–3B

Meta agreed to acquire Manus to accelerate “agents that act” across its consumer + business stack (including Meta AI), paying a reported ~$2–3B and emphasizing there will be no continuing Chinese ownership post-transaction. Manus claims extremely fast traction for an agent product (reported >$100M ARR within ~8 months; some coverage cites ~$125M revenue run-rate)—a strong signal that subscriptionized, outcome-oriented agents are becoming an M&A priority for hyperscalers.

What to Watch this Week

 🎙️AI on Fire: Real Builders. Real Heat.

Stories from founders building AI Agents. Explore all episodes here

⚔️ AArena: The Battleground for AI

Stop demo-hopping. One workspace. Every agent. Real results.

👉 Enter the Arena | 👉 See the Battleboard

🏆 The Leaderboard Never Sleeps

The global ranking of AI agents is shifting every day. Who’s on top? Who just dropped?

👉 Explore the Leaderboard →

🗺️ The Map of AI Agents (Live & Growing)

We’re charting the entire AI agent ecosystem — thousands of options across categories.
Your next agent is already on the map.

👉 Explore the Landscape Map →