.png)
This is a serious, well-placed bet. The Big 4 are investing where it matters, and the second tier will follow. The audits, tax engagements, and advisory work of the next five years will be faster, broader in coverage, and more accurate because of it.
But there is one layer of the engagement the money has not reached yet. And it is the layer clients experience the most.
What the money is actually buying
Each of the Big 4 is pointing its AI investment at the same categories of work.
AI that reviews audit documentation and suggests improvements. AI that navigates draft financial statements and answers questions about their content. AI that extracts and reconciles data across thousands of files. AI that drafts communications, flags anomalies, and performs initial controls testing.
Deloitte has been explicit about it. Its Omnia platform now includes GenAI features that perform initial reviews of audit documentation, navigate uploaded financial statements, and support tie-out procedures. The firm is integrating agentic AI capabilities that remember context, coordinate with other agents, and handle multi-step audit tasks.
KPMG has embedded Azure OpenAI into Clara, its smart audit platform, so that 85,000 audit professionals can point directly at client data instead of ingesting it manually. The firm projects over $12 billion in incremental growth from the alliance.
EY built EY.ai as a unifying platform across strategy, transactions, tax, assurance, and risk, backed by a $1.4 billion investment and its proprietary LLM, EYQ.
PwC has gone the broadest on workforce deployment. Signing 100,000 employees onto ChatGPT Enterprise made it OpenAI's single largest enterprise customer, with custom GPTs for tax return review, proposal drafting, and reporting.
The tooling is live. The deployment is in motion. The billions are already spent.
A quieter conversation is starting around fees
Audit, tax, and advisory fees have historically been built on hours. Partner rate times partner time. Manager rate times manager time. Associate rate times associate time. That model made sense when humans did every step.
As AI handles more of the first-pass work, the hours on routine tasks shrink. The question of what that means for pricing has started showing up in boardroom conversations.
Early in 2026, the Financial Times reported that KPMG renegotiated a 14% fee reduction from its own auditor, Grant Thornton, by pointing to AI efficiency gains in the audit process. Companies House filings confirmed the fee dropped from $416,000 to $357,000.
Verdantix called it an early signal of a pricing inflection point for audit. A 2024 CPA Australia study found that firms implementing AI-assisted workflows reported efficiency gains of 40 to 60 per cent on routine compliance work, with direct implications for how engagements get priced.
A rethink is underway across the profession. Subscription models, fixed fees, and outcome-based pricing are all on the table. The firms that get ahead of this conversation, rather than having it forced on them, will be in a stronger position with their clients.
The layer AI still has to reach
Here is the part that matters most for the client experience.
AI on the firm's side only works on data the firm actually has. The testing, the drafting, the anomaly detection, the reconciliation, all of it assumes the source documents have already arrived, in the right format, from the right person, at the right time.
That assumption is where the story gets interesting.
The front end of every engagement, across the whole profession, still largely runs on email attachments, Excel-based PBC lists, shared folders, and follow-up messages that pile on follow-up messages. A recent Caseware analysis of modernizing audit data collection framed it cleanly: when the first exchange between the audit team and the client is a cluttered request list and a set of formatting instructions, the engagement starts in friction.
That friction absorbs professional capacity in the background. Clarifications. Missing attachments. Wrong versions. Status checks. Someone on the client side who went on leave and never handed over.
This is the layer where the firms that are investing in AI today have the most room to add value next. A GenAI model that drafts an audit communication in 30 seconds is brilliant. It lands even harder when the client did not spend three weeks chasing bank confirmations through an inbox with 400 other threads to get the data there in the first place.
Where the coordination layer fits in
Professional services has never struggled because the expertise was missing. The expertise has always been there. What slows firms and clients down together is everything around it. The coordination. The follow-ups. The document chasing. The invisible machinery that moves work from one party to the next.
For audit, tax, M&A, legal, and compliance work, that machinery still runs largely through email. Responsibilities live in people's heads. Progress depends on asking. Critical documents move as attachments. Follow-ups depend on memory. Status depends on chasing.
Email became the default operating system of professional work, yet it was never built for that role. It is a chronological list of unstructured text. It does not understand priority, responsibility, deadlines, dependencies, or business impact.
Structured work has been forced through an unstructured pipe for too long. And the firms that solve the coordination layer will capture a second wave of efficiency on top of the AI investments already underway.
What this looks like in practice
Alkmist is built for this layer specifically.
Every engagement starts with structured taskflows instead of a PBC list pasted into an email. Document requests, clarifications, approvals, and sign-offs live in one shared environment where both sides see the same thing. Clients see what is required and what is pending. Professionals see who owes what by when. Every interaction is traceable, accountable, and auditable by default.
Intelligence is embedded directly into the flow of collaboration. Unstructured communication converts into structured execution. Tasks, responsibilities, priorities, and follow-ups are extracted automatically, so professionals can focus on judgement calls and client conversations instead of status chasing.
This is where the real compounding effect lives. AI on the firm's platform plus structured collaboration with the client equals an engagement that is faster on both sides, with a clear trail of what happened and why.
A complementary story, not a competing one
The Big 4 are investing intelligently in their side of the equation. The next logical step for the whole profession, Big 4 and mid-market alike, is closing the coordination gap on the client side.
When both layers move forward together, the productivity gains stop being a one-sided story that clients have to take on faith. They become something both parties can see, measure, and talk about openly.
The technology shift is real. The client experience shift is the natural second act. And both of them depend on something quieter than a billion-dollar press release: making the way professional firms and their clients actually work together, request by request, document by document, signature by signature, finally match the quality of the work itself.
That is the part that changes everything downstream.
See what a modern engagement actually looks like
Alkmist replaces the email-and-Excel front end of audit, tax, M&A, and advisory work with structured taskflows built for professional services. White-labeled. Used by more than 8,000 professionals across 62 countries.
If AI is about to reshape how firms deliver engagements, the coordination layer is the foundation that has to move with it. Book a demo and see how Alkmist fits into your engagement flow.
