Writing from inside the workshop.
Plain-language thinking on custom AI for professional-services firms. What to build first. Why platforms keep missing. How a real two-week sprint actually runs.
Everything we’ve written, in reading order.
What Florida Tax Ballot Battles and New AI Compliance Warnings Mean for Firm Automation
Recent tax-vote disputes show how fast policy changes can reshape workloads, client questions, and review risk. For law and accounting firms, the practical response is custom AI and automation built to track change, route exceptions, and support stronger supervision.
ReadWhat Recent CPA Practice Advisor News Says About Custom AI Workflows for Accounting Firms
Recent news on compliance enforcement, advisory growth, and firm acquisitions points to the same operational need: accounting firms need custom AI and automation that can handle more than basic keyword matching or one-off tasks.
ReadCustom AI Workflows for Law Firms and Accounting Firms Are Moving from Chat to Connected Systems
Recent news from legal and accounting technology points to the same direction: firms want AI that understands firm context, connects to existing systems, and helps staff do real work. For professional-services leaders, the practical opportunity is less about generic chat and more
ReadCustom AI Workflows Help Professional Services Firms Tackle Fraud Risk, Compliance, and Rising AI Spend
Recent news on stolen tax refund checks, nonprofit compensation rules, and law-firm AI platform partnerships points to the same lesson: firms need custom AI and automation built around real workflow risk, not generic chatbots. For law and accounting leaders, the opportunity is in
ReadWhat Ramp Stack and Wordsmith Say About Custom AI Workflows for Law and Accounting Firms
The latest AI vendor news points in the same direction: professional-services firms want systems that do the work, keep decisions reviewable, and fit real workflows. For law and accounting leaders, that means focusing less on generic chatbots and more on custom AI, automation, an
ReadWhat Tax and Regulatory News Means for Custom AI Workflows in Law and Accounting Firms
Recent state tax fights, SEC settlement scrutiny, and legal-tech dealmaking all point to the same need: firms should design AI workflows that can handle changing rules, document review, and workflow routing with more consistency than generic chat tools.
ReadWhat Hanson Bridgett and Kirkland Signal About Custom AI Workflows for Law Firms
Two major law firm moves point to the same trend: firms are shifting from general AI curiosity to specific, workflow-based adoption. For law firm leaders, the practical question is no longer whether to use AI, but which tasks to standardize, govern, and automate first.
ReadWhy Big Firms Are Building Custom AI Platforms and What It Means for Law and Accounting Firms
Recent news shows professional-services firms and regulators moving in the same direction: more custom AI, more automation, and more attention to how work is governed. For law and accounting firm leaders, the lesson is clear-generic tools can help, but durable advantage comes fro
ReadWhat the Latest Tax and Audit News Means for Custom AI Workflows in Professional Services Firms
Recent tax, benefits, and audit headlines all point to the same operational need: firms need tighter systems for updating client guidance, routing exceptions, and documenting judgment. Custom AI and automation can help teams turn changing rules into repeatable workflows without a
ReadThe Two Walls Every Enterprise AI Deployment Hits: Reliability and Runaway Cost
Harvey is worth $11B, Legora $5.6B, and Kirkland is spending half a billion to build its own. Underneath the legal-AI gold rush, two hard walls are showing up: agents that fail the math of reliability, and token bills that scale in ways no SaaS budget was built for. Here is what that means for buy-vs-build — and the way through.
ReadKirkland & Ellis Is Spending $500M to Build Its Own AI. The Lesson Isn't Just for Big Law
The world's highest-grossing law firm just decided to build its own AI instead of leaning on Harvey or Legora. The headline is the half-billion-dollar budget. The real signal is why a firm with unlimited options chose bespoke over a one-size-fits-all platform.
ReadAfraid Your Firm Is Falling Behind on AI? Here's Where to Actually Start
If you run a law or accounting firm and the AI conversation feels like noise you can't act on, you're not behind because you lack a strategy. You're stuck because nobody has turned the worry into one concrete workflow. Here's how to start without betting the firm.
ReadWhat CFO-Style AI Governance Means for Law and Accounting Firm Automation
Recent finance-tech announcements point to the same shift: firms are moving from simple AI assistance to governed automation, anomaly detection, and agentic workflows. For law and accounting leaders, the opportunity is not just speed, but better controls around the work that shou
ReadLegal AI Workflows Are Moving From Chatbots to Measured, Agentic Systems
Recent legal AI news points to a practical shift for professional-services firms: usage benchmarking, specialist agents, and stronger review controls matter more than generic chatbot access.
ReadLegal AI Workflows Are Moving Past Chatbots Into Contract and Case Operations
Recent legal AI news points to a practical divide: many teams still lack clean contract systems, while AI-driven case platforms are demonstrating end-to-end workflow automation. For law, accounting, and advisory firms, the near-term opportunity is to automate specific operating w
ReadLegal AI Tools Are Getting More Vertical, but Firm Metrics Still Decide the Workflow
Recent coverage points in two directions at once: legal AI tools are becoming more specialized, while firm operators still need better visibility into performance. For law and accounting firms, the practical move is to connect AI workflow decisions to the metrics that explain pro
ReadCustom AI Workflows Are Moving Into the Daily Work of Law and Accounting Firms
Recent coverage of Claude for Legal, group audits, and accounting-firm lead generation points to the same operational lesson: AI is becoming more useful when it is embedded into the actual workflow, not treated as a separate chatbot.
ReadLegal AI and audit standards point to custom AI workflows, not generic chatbots
Recent legal AI news, Clio's retention data, and the AICPA's updated confirmation standards all point in the same direction: firms need AI that fits real work. For law and accounting leaders, the opportunity is to build targeted workflows around high-friction tasks, evidence hand
ReadClaude for Legal Launches: What Anthropic's Plugin-and-Connector Bet Means for Your Law Firm
Anthropic launched Claude for Legal on May 12, 2026 with 12 practice-area plugins, more than 20 MCP connectors, and Claude embedded inside Microsoft 365. Here is what actually shipped, why the connector list matters more than the plugins, and where a platform stops and your firm's own build has to begin.
ReadMike OSS: The Open Source Legal AI That Clones Harvey and Legora — And What It Means for U.S. Law Firms
A former Latham attorney spent two weeks rebuilding the core of Harvey and Legora, released it on GitHub under AGPL v3, and watched it cross 2,200 stars in 72 hours. Here is what Mike OSS actually does, where it falls short for U.S. firms, and what its existence proves about every legal-AI invoice your firm signed this year.
ReadCross-Border Service Desks and AI Advisory Tools Point to the Same Workflow Problem
Recent moves from Forvis Mazars, Wolters Kluwer, Intuit, and legal AI educators all point to a practical theme: firms need better coordination around client work, not just more software. Custom AI and automation should be designed around that operating model.
ReadAgentic AI for Dental Practices: The Workflow Layer Between Patients, PMS, and Production
Dental practices do not need another chatbot. They need applied agentic AI development that can sit between the patient, the practice management system, insurance, imaging, and recall workflows, then move bounded work forward with citations, audit trails, and human review.
ReadAccounting Firm Automation Lessons from Embezzlement, Caregiving, and Advisory News
Recent accounting news points to three practical workflow priorities for firm leaders: stronger financial controls, less manual compliance work, and more repeatable advisory delivery.
ReadWhy AI Chatbots Fail Law Firms and Accounting Firms: Workflow Beats Search
Bolting a chatbot onto a database sounds like custom AI for law firms and accounting firms, until the attorney or CPA stares at the prompt box and goes back to Outlook. Real adoption starts when the app meets the professional inside the workflow, removes the next annoying step, and gives them less to think about, not more.
ReadTax Season 2026 Postmortem: Where AI Actually Helped Small Accounting Firms (and Where It Quietly Added Work)
Tax season 2026 is the first one most small CPA firms ran with AI in the building. Three weeks after the April 15 deadline, the partner conversations sound very different from last year. Here is where AI saved real hours, where it quietly absorbed new work nobody put on the dashboard, and what to fix before tax season 2027.
ReadKilling the Billable Hour: How AI Is Forcing Mid-Size Law Firms to Reprice Their Work in 2026
When a partner can deliver in two hours what used to take twenty, the billable hour stops being a unit of value and starts being a tax on the firm’s own efficiency. Here is why AI is finally collapsing the model that built modern legal practice, the pricing structures replacing it, and how mid-size firms should plan the transition before clients force the question in their next RFP.
ReadThe Tokenizer Tax: Why AI Costs More Than the Pricing Page Says
Tokenizer changes, reasoning tokens, and subsidized subscriptions are turning AI pricing into a real operating discipline. The winners will know when to use frontier models, cheaper models, and local open-source AI.
ReadDeep Models Are the Future of Enterprise AI
Enterprise automation does not need one generalist model that is decent at everything. It needs models that are unusually deep at the specific work the company actually runs.
ReadThe $100M Solo Law Firm: Why a Single Attorney With the Right AI Stack Will Out-Earn an AmLaw 200 Office by 2031
An AmLaw 200 firm generates roughly $1–2M of revenue per attorney. By 2031 a single licensed lawyer running a custom-AI stack — intake, conflicts, drafting, e-discovery, scheduling, billing, K-1s, client comms, all owned in one repo — will clear $100M in fees. This is the architecture that gets the first one there, the niches that produce the first ten, and what it means for the partnership model that has run U.S. legal practice for 150 years.
ReadMCP for Law Firms: A Practical 2026 Guide to the Model Context Protocol
Model Context Protocol is becoming the USB-C of legal AI — a standard way for chat clients, agents, and internal tools to call your firm's verbs safely. Here is what MCP is, why it matters more for law firms than the platforms admit, and how to ship a useful first server in two weeks.
ReadAI Client Intake for Accounting Firms: From PDF Hell to Structured Onboarding
Most accounting firms quietly burn 20 hours per new client on intake — chasing documents, retyping fields into the engagement system, and re-validating the same numbers four times. Here is the AI-shaped intake pipeline that small CPA firms are starting to build, what it actually replaces, and the one rule that keeps it audit-defensible.
ReadEvals Are the Real Moat: How Professional Services Firms Should Test Their AI Agents
The model you pick this quarter will be obsolete by next. The eval suite you build, if you build a good one, compounds for years. Here is what an eval system actually looks like for a custom AI agent in a law firm or accounting firm — and the four failure modes that wreck most attempts.
ReadBuilding Workflow Blueprint: A Task Workspace Ready for Headless Agents
Workflow Blueprint is a Next.js task planning workspace with boards, notes, invite-only accounts, and a private read-only API built for headless agents. This is the build story: why the product exists, how the architecture is shaped, and why the agent contract matters as much as the UI.
ReadHeadless Agents Are the Real Software 3.0 Shift
AI agents do not need a GUI. Salesforce Headless 360 makes the enterprise signal hard to ignore: the next platform war is not prettier dashboards, it is programmable surfaces agents can safely operate.
ReadBuilding JobFinder.guru: An AI-Ranked Job Search Command Center, Shipped Solo
JobFinder.guru is a single-operator AI job search platform that replaces spreadsheets, tab sprawl, and scattered notes with one disciplined weekly workflow — AI-ranked job matches, a personal job search CRM, and a visible pipeline. Here is how it was built, who it is for, and what it teaches about AI-native product design.
ReadWhat Actually Happens on a 30-Minute Bottleneck Audit
Every engagement I take starts with a 30-minute call. Most partners expect a pitch. Here is what the call actually is: three questions, a real answer, and an honest take on whether custom AI is the right tool for the firm at all.
ReadWhat "AI-native" Actually Means for a Small Accounting Firm in 2026
Every accounting-tech vendor is calling itself AI-native this quarter. For a 2 to 20 CPA firm, most of those products are still a chatbot stapled to a 2019 workpaper tool. Here is what the phrase actually means, and why the difference is worth billing hours.
ReadThe Three Custom-AI Mistakes That Kill Professional-Services Builds in Year One
The firms that get the most out of bespoke AI in year one do not pick the most exciting workflow or the fanciest model. They avoid three specific mistakes the rest of the market keeps making — none of them technical.
ReadHow to Measure AI ROI at a Professional-Services Firm: A Six-Metric Framework
Every AI vendor has a hockey-stick deck and a case study with a suspiciously round number. Partners keep asking a different question: what is this thing actually doing for the firm? Here is the six-metric framework we use to answer that — the one built for professional-services economics, not SaaS-dashboard theater.
ReadThe Partner's Guide to Writing Your Firm's AI Use Policy in 2026: A Ten-Section Framework
Your malpractice carrier is about to ask whether your firm has one. Your largest client already has. A state bar inquiry could require one tomorrow. Here is the ten-section framework partners can use to draft, circulate, and adopt a working AI use policy in a month — without turning it into a compliance-theater document nobody reads.
ReadHow Much Does Custom AI Actually Cost a 20-Attorney Firm? A Three-Year Line-Item Breakdown
Legal-AI pricing is deliberately opaque in 2026. This is the spreadsheet I walk partners through before they sign anything — three columns, three years, and the line items the vendor slides leave out.
ReadAttorney-Client Privilege and AI: A Partner's Guide to Vendors, Data Paths, and the Waiver Problem
When your firm's data leaves your infrastructure and lands on a vendor's model, what exactly happens to privilege? Here is the framework I walk partners through before signing any AI contract — and why the deployment model matters more than the feature list.
ReadHow to Pick Your Firm's First AI Workflow: A Partner's Checklist
Most firms that fail at AI fail at the first decision: which workflow to automate first. Here is the five-question test I walk partners through before we agree to build anything.
ReadAnatomy of a Two-Week AI Sprint in a Professional-Services Firm
A complete walkthrough of a typical ten-day engagement: from the first discovery call to a production-ready custom module pushed to the client's GitHub.
ReadAccounting Firms in 2026: The Case for Building Instead of Buying
Your workpaper tool, your workflow platform, your tax prep suite, and your document manager are all raising prices and adding AI features you didn't ask for. Here is the alternative.
ReadWhy Law Firms Should Stop Buying Legal AI Platforms
Harvey, Legora, and every other legal-AI vendor is selling you the same product as the firm down the street. Here is why a custom AI module, scoped to one workflow, will beat any platform you can license.
ReadBuild vs. Buy: Why Professional-Services Firms Are Done with SaaS
The SaaS subscription model promised simplicity but delivered dependency. Here is why forward-thinking law and accounting firms are walking away, and building instead.
ReadIn Software 3.0, the Rules Change
Software is no longer built for humans with a sprinkle of AI. It is AI-first, adaptable, and vibe-coded. Welcome to the agentic future, where most software will never be sold.
ReadThe Software 3.0 Manifesto: Build, Hand Off, Own
A deep dive into the philosophy behind Software 3.0: precision over platforms, AI as the engine, software ownership, and why the build-vs-buy debate has been permanently rewritten.
ReadEvery post here started with a real conversation with a real partner. If there's a problem you see in your practice, we'd like to hear it.