Sample Deliverables

Don't imagine the deliverables. Read them.

Every document below is a full sample of something my clients actually receive — the monthly report, the assessment, the backlog, the policy. Open the PDFs. If this is what your leadership team is missing, you'll know by page two.

Samples are written for Meridian & Frost, a fictional 118-person accounting and advisory firm — a composite of patterns from real engagements, not a disguised client. (The private-AI deck uses Calloway & Reed LLP, a fictional law firm whose data justifies the opposite architecture.) The names and numbers are invented. The format, the method, and the level of detail are exactly what you'd receive.

What the retainer delivers, in writing

The advisory seat is $15,000/month. These six documents are why it isn't a leap of faith — this is the paper trail one composite client generates in a single quarter.

8 pages · the flagship monthly deliverable

The Monthly AI Executive Report

Adoption metrics, shipped work, the scored backlog, blocked decisions, risk notes, vendor updates, and the next 30 days. The written record that makes the retainer concrete.

View the sample PDF →

Read: inside the monthly report →

6 pages · updated every month

The AI Opportunity Backlog

Eighteen candidate workflows scored on impact, feasibility, data tier, and time-to-value — including the killed list, kept visible on purpose.

View the sample PDF →

Read: how the scoring works →

4 pages · monthly governance artifact

Governance & Risk Review

What's approved, what's exposed, what changed: exception requests, near-misses, and a risk register with owners and dates.

View the sample PDF →

Read: governance without bureaucracy →

3 pages · quarterly money memo

AI Investment & ROI Memo

Every initiative gets a verdict in writing — fund, hold, or kill — with the P&L line it moves and the math shown.

View the sample PDF →

Read: how I decide what to kill →

3 pages · during any build

Implementation Oversight Brief

What I pushed back on, the go/no-go gates, and the acceptance criteria — supervision of the builder, documented, even when the builder is my own firm.

View the sample PDF →

Read: who watches the builders →

4 pages · the enablement program

90-Day Adoption & Training Plan

Champions, a training calendar, measurement from week one, and the playbook for when a department stalls. Unused tools have zero ROI.

View the sample PDF →

Read: why AI training fails →

What the Assessment produces

The AI Executive Assessment is the two-week, fixed-price front door ($5,000, credited toward advisory). This is the actual output. See how the two weeks run →

12 pages · the board-ready readout

The AI Executive Assessment Report

Stakeholder findings, the shadow-AI audit, a workflow and data-source map, five scored use cases, the architecture recommendation, and a 90-day roadmap.

View the sample PDF →

Read: the 90-day roadmap, in full →

4 pages · private vs. commercial vs. hybrid

Architecture Decision Memo

The decision record for a composite client — where I recommend against private AI because the arithmetic said so. That honesty is the product.

View the sample PDF →

Read: when private AI makes sense →

4 pages · scored, never a reseller's quota

Vendor & Model Evaluation Brief

Claude, ChatGPT, and Copilot scored against one firm's data tiers and workflows — gates first, weighted criteria second, price last.

View the sample PDF →

Read: how to evaluate AI vendors →

Implementation, specified

When something gets built, it gets built from a document — one an engineering team can work from and a managing partner can verify. These are those documents.

7 pages · client document intake, build-ready

Workflow Automation Specification

Swimlanes, a state machine, an exception queue, permission rules, and acceptance criteria — the spec that turns a 2,100-hour document chase into a governed workflow.

View the sample PDF →

Read: anatomy of a workflow spec →

7 pages · permission-aware retrieval, cited or silent

Governed Knowledge Layer — Design Spec

The four rules that make a knowledge assistant governed: permission-aware retrieval, version truth, citations or silence, curated corpus only.

View the sample PDF →

Read: why knowledge assistants fail →

6 pages · AI-ready data without the data lake

AI Data Foundation Blueprint

Source inventory, a governed document architecture, and the migration plan for 1.9 million legacy files — sized for 118 people, not an enterprise program.

View the sample PDF →

Read: the corpus failures this prevents →

13 slides · full system design, for a firm whose data demands it

Private AI Implementation Deck

Architecture flows, the exact component stack, five security layers, eval gates, and an eight-week deployment plan — for a fictional law firm where private AI is the right call.

View the sample PDF →

Read: when private AI makes sense →

"Nothing gets built from a conversation. Every build starts from a document an engineering team can work from and a managing partner can verify — if it isn't specified, it isn't governed."
— The standard behind every spec on this page

Enablement, measurement & the boardroom

6 pages · the 60-day rollout that made the seats stick

Commercial AI Enablement Playbook

Guardrails before users, a quick win in the first 45 days, champions, and console-measured adoption — the antidote to pilot purgatory.

View the sample PDF →

Read: the part nobody sold you →

6 pages · 150 questions, 20 traps, hard gates

Evaluation Plan: Golden Datasets & Benchmarks

How you know the AI is right before you trust it — golden datasets, superseded-version traps, and a zero-tolerance permission gate.

View the sample PDF →

Read: golden datasets, explained →

11 slides · partner-meeting ready, with presenter prep

Board Update Deck — Q2 2026

The quarter on one slide, the money slide, the decisions the board owes answers on, and the questions partners should expect — with answers.

View the sample PDF →

Read: inside the monthly report →

Governance, ready to follow

3 pages · the one-page policy, plus the desk card

AI Acceptable Use Policy

Data tiers, approved tools, and six rules people can actually follow — proof that governance at 50–500 people fits on a page.

View the sample PDF →

Read: governance without bureaucracy →

Free for any leadership team

5 pages · 15 questions, scored

The AI Readiness Checklist

Ten minutes with your leadership team and a clear read on where you stand. One scoring band reads: "you have shadow AI, not an AI strategy."

View the sample PDF →

Get it with the scoring guide →

This is what a seat looks like.

If reading these made your current AI reporting feel thin, that's the gap the advisory seat closes. Start with the Assessment — two weeks, fixed price, credited toward the retainer.

Book an AI Strategy Call Want the retainer details first? See the advisory offer →