The workforce consequence is already visible. Big Four entry-level graduate hiring fell 44% year-over-year in 2024. KPMG cut some cohorts by nearly 30%. EY's head of tax is explicit about why: AI agents now handle the document management, research, and compliance routing that junior staff spent years learning on.

If you run a 5-15 person accounting firm, this news probably triggers one of two reactions: either it sounds impossibly far away from your daily reality, or it sounds like exactly the kind of thing that will erode your client base in the next 24 months.

Both reactions are right. And there's a third option.

---

What EY Is Actually Automating

It's easy to hear "150 AI agents" and imagine futuristic technology your firm couldn't possibly access. That's not what this is.

The workflows EY is automating with its agent deployment are not exotic. They're the same tedious, time-consuming administrative tasks that eat up capacity at every accounting firm — regardless of size.

Document upload walkthroughs. When a client needs to upload their prior-year return, W-2s, 1099s, and bank statements, someone on your team currently manages that process. They send emails, follow up, answer questions about which documents to upload, and chase down missing items. EY built an agent for that. So has TaxDome, Financial Cents, and Karbon — tools your small firm can use today.

Research routing. When a foreign tax question comes in, finding the right staff member and getting them the relevant context takes time. EY has an agent that identifies the question type, retrieves the relevant rules, and routes it. A small firm can build something similar with a trained AI assistant in a shared workspace — it won't be as sophisticated as EY's enterprise deployment, but it can handle 80% of the same function.

Compliance case routing. When an IRS or state notice arrives, someone reads it, figures out what type of notice it is, determines who should handle it, and gets it to them. That process currently requires judgment and time from someone on your team. AI can handle the classification and routing step — the judgment call about what to do with it still belongs to a credentialed professional.

Client Q&A. When a client calls or emails to ask about extension deadlines, standard deduction thresholds, or filing requirements for a new state, that question lands in someone's inbox. EY has an agent for that. A small firm can build a training-data-free version using a well-configured AI assistant with access to IRS publications and your firm's FAQ document.

None of these are new capabilities. They're existing AI tools applied to the workflows that consume the most administrative time in a typical small accounting firm.

---

The Gap That's Opening (and What It Means for Fees)

Here's the practical implication of EY's deployment for small firm owners.

When EY can process compliance cases at scale with AI, their cost per engagement goes down. When their cost per engagement goes down, their pricing flexibility increases — or their margins do. Either way, the pricing dynamics in the accounting market shift.

Your clients don't necessarily know what EY's internal costs look like. But they know what their experience looks like. If an EY-level workflow produces faster turnaround, fewer follow-up emails, and more proactive communication — because AI handles the routine communication layer — clients begin to notice the difference when they compare notes with peers.

This is already happening in the legal market. Clio's 2026 Legal Trends data shows that mid-sized law firms using AI are handling 65% more work volume with the same teams — a productivity gap that's increasingly visible to the clients comparing proposals from different firms.

The accounting equivalent is coming. The firms that close that gap in the next 90 days are the ones that will have an answer when a client asks why their competitor's firm turned around a document request in four hours instead of four days.

---

The Small Firm 90-Day Sprint

You are not EY. You don't have an enterprise AI engineering team, a dedicated model training budget, or 80,000 practitioners to test on.

But you have something EY doesn't: you can move in a week, not a quarter.

A 5-15 person accounting firm can run the equivalent of EY's agent deployment in 90 days — not with 150 agents, but with three targeted automations that address the same categories of work. Here's how.

Weeks 1–4: Document Collection Automation

What EY has: An agent that manages the client document upload process — reminders, status tracking, follow-up.

What you can do: Configure the document collection workflow inside your practice management platform. TaxDome, Karbon, and Financial Cents all have native client portals with automated reminder sequences. If you're already using one of these platforms and haven't activated the automated reminder feature, you're leaving the most accessible productivity gain in your stack untouched.

Set it up for one engagement type first — individual returns, for example. Configure the portal to send a reminder at day 7, day 14, and day 21 after document request. Track how many follow-up emails your team sends compared to before. That delta is your time savings number.

Target outcome: Reduce team time on document follow-up by 50% for the engagements you automate. For a firm processing 200 individual returns, even 15 minutes per engagement saved is 50 hours recovered in a season.

Weeks 5–8: Tax Notice Triage

What EY has: An agent that reads incoming notices, classifies them by type, and routes them to the right staff member with relevant context.

What you can do: Create a shared inbox rule set combined with a trained AI assistant. This doesn't require a custom-built agent. You need: (1) a designated inbox for all client notices; (2) a prompt template that takes a notice's text and returns a classification (CP2000, CP14, state balance due, state audit, etc.) plus the relevant response deadline; (3) a routing protocol that assigns each classification to a specific staff member.

You can do this with a standard AI tool — Claude, ChatGPT, or a similar assistant — plus a shared spreadsheet or Notion database. It takes a day to set up, a week to calibrate, and it removes the cognitive overhead of notice triage from your senior staff permanently.

Target outcome: Reduce the time from notice receipt to staff assignment from 24-48 hours to under 2 hours. More importantly, reduce the number of notices that sit in general inboxes without being seen until a deadline is close.

Weeks 9–12: Client Q&A Triage

What EY has: An AI system that handles routine client questions about tax rules, filing requirements, and standard information — so practitioners receive only the questions that actually require their judgment.

What you can do: Build a trained Q&A assistant using your firm's most common client questions. Start with the questions your team answers most often: extension deadlines, estimated payment schedules, W-2 vs. 1099 classification for a specific situation, standard deduction thresholds, FBAR filing triggers.

You don't need to build a client-facing chatbot to do this. A well-organized internal document (your firm's FAQ plus relevant IRS citations) plus a trained AI assistant that references it gives your staff a triage layer — they paste the client question in, get a draft answer, and review before sending. The client gets faster responses. Your senior staff spends their time on the 20% of questions that actually require professional judgment.

Target outcome: Reduce average response time for routine client questions from 24-48 hours to same-day. Reduce senior staff time on client communication triage by 30%.

---

The Number That Matters

At the end of 90 days, calculate the hours recovered across these three workflows.

If your firm processes 300 engagements and saves 30 minutes per engagement on document collection, that's 150 hours. If your notice triage saves 1 hour per notice and you receive 200 notices in a quarter, that's 200 hours. If your Q&A triage saves 15 minutes per client interaction and you have 100 routine interactions per month, that's 75 hours per quarter.

That's potentially 425 hours of capacity recovered in 90 days — at zero cost beyond the time to configure tools you may already be paying for.

You don't need 150 agents. You need three workflows, 90 days, and the willingness to measure what changes.

EY has an enterprise AI engineering team. You have speed and focus. Use it.

---

Sources: Bloomberg Tax — Big Four Firms Roll Out AI That Can Handle Routine Tasks Solo | EY — Autonomous AI Is No Longer Theoretical Survey