For the AI-curious operator

You went down the Claude rabbit hole. Now you need someone who can actually build it.

You've watched the webinar. Sat through the cohort. Saved 88 TikToks about AI agents. You know AI is the wave, you just don't know where to start, what to build first, or what platform to bet on. I do. I've been architecting business systems and automations since 2018. The AI layer is just the new generation of tools.

TitleOperations & AI Systems Architect
DisciplineCPA · Auditor · Lean Six Sigma
Doing this since2018 (the tools changed)
ClaudeOpenAIZapierMakeAirtableHubSpotn8nNotion AIChatGPTVector DBsCustom GPTsSkills & Connectors ClaudeOpenAIZapierMakeAirtableHubSpotn8nNotion AIChatGPTVector DBsCustom GPTsSkills & Connectors
— 02 / Before it had a cool name

I've been doing this since before it had a cool or trendy name.

I started building CRM workflows, multi-platform automations, and trigger-based operating systems in 2018, back when most "automation consultants" were selling Mailchimp drips. The clients changed. The tools changed. The thinking (root-cause, holistic, financially-aware, compliance-aware) didn't.

"AI architecture" isn't a pivot for me. It's the same skill applied to a new generation of tools. Find the manual task. Find the broken handoff. Find the human doing what a system could do. Build the system. That was true with Zapier in 2019. It's true with Claude agents in 2025.

Now → which level are you on?

— 03 / Three levels of AI maturity

Most businesses are at Level 1. They just don't know it.

Before we talk about AI agents and decision layers, we have to know where you actually are. Most of the work I do is helping founders see honestly which level they're on, and which level they actually need next.

Level 01 · Foundation

Automate the manual.

The repeatable, non-emotional, non-decision work that's still being done by hand. Most of what eats your day doesn't need AI; it needs a system that actually exists.

Tools at this level

  • CRM workflows & trigger automations
  • Zapier · Make · n8n · native integrations
  • Onboarding portals & document automation
  • Notification & reporting triggers
  • Form-to-system pipes (intake, contracts, payments)

You're here if your invoicing isn't automated, your onboarding still pings a VA, or your CRM is "we use it sometimes."

Level 02 · Operations

Use AI to enhance the work.

Once the manual stuff is wired, AI starts earning its keep: drafting copy, analyzing calls, generating creative, building dashboards, producing the artifacts a small team used to outsource.

Tools at this level

  • Claude Projects · custom skills · domain-trained agents
  • Marketing & content generation pipelines
  • Sales call analysis & objection mining
  • Auto-generated proposals, briefs, recaps
  • Internal "business brain" with every doc indexed

You're here if Level 1 is solid and you're paying humans to do work AI can do faster, cheaper, and at higher volume.

Level 03 · Decision Layer

Stop flying blind. Make data-driven decisions faster.

By Levels 1 & 2, you have a team of humans, agents, and automations generating data every single day. The question is whether you're actually using it. Level 3 is your AI data analyst plus your AI board of advisors, working in tandem: dashboards and agents continuously reading the signal, then thought partners (CFO, CMO, COO personas) helping you decide what to do about it. Better decisions, sooner, more strategically, in partnership with advisors you otherwise couldn't afford.

Tools at this level

  • AI data analyst agents · always-on dashboards with commentary
  • Trend, churn & anomaly detection · early-warning systems
  • AI advisor boards (CFO, CMO, COO personas) for strategic calls
  • Pricing, pitch & investment decision support
  • Risk & compliance monitoring agents
  • Proactive insights pushed to you on a regular cadence

You're here if Levels 1 & 2 are working and you're tired of making major decisions without consulting the data first.

"No."The honest answer

If your invoicing isn't automated, you don't need AI. You need a CRM. I'll tell you that on the call and save you $40,000.

Most "AI architects" will sell every client a Level-3 buildout, whether they need it or not. I won't. The biggest unlock for most businesses I talk to isn't AI; it's the Level-1 automation they should have done five years ago. We start where you actually are.

— 04 / Three ways to work together

Start with a call. Build when you're ready.

Three engagement levels. The first is a 30-minute paid assessment to diagnose where you actually are. The second is a fixed-scope build sprint with three flavors, scoped on the call. The third is an ongoing architect retainer for whatever comes next.

Option 01 · Diagnose

AI Assessment Call

30 min · 1:1 · paid intake form

Fill out a short intake. We get on a call. In 30 minutes I tell you whether you're Level 1, 2, or 3, what you actually need to build first, and whether AI is the right answer at all. Honest, financially-aware, root-cause diagnosis.

$197
Credited toward any engagement booked within 30 days
  • Pre-call intake form (tools, team, processes, pain)
  • 30-minute working session, diagnostic, not pitch
  • Level 1/2/3 assessment + which to build first
  • Recommended next-step (or honest "you don't need me")
  • Recording & written summary delivered after
Option 03 · Continuous

Continuous Architect

Monthly or quarterly · no commitment

Once the build is shipped, the tools keep changing. Claude Design dropped two weeks ago, MCP connectors are weekly, agents get smarter every month. Stay on retainer so the new capabilities show up in your business as they ship, not three years later.

Monthly RetainerOngoing access · cancel anytime
$1,500/mo
Quarterly Deep-DiveOne intensive · saves vs. monthly
$3,500/qtr
  • New-tool review & implementation
  • Quarterly deep-dive on your full stack
  • Slack/email access between calls
  • First call on every new release I find useful
  • Cancel month-to-month, no minimums, no fine print

Every engagement starts with the $197 assessment call. Don't buy a build until we've talked.

— 05 / Don't be the next one

Every fifteen years, the way work gets done changes. The businesses that adapt survive.

You don't need to be reminded what happened to Blockbuster. You watched it happen. You watched it happen to Borders, to Kodak, to BlackBerry, to every newspaper you grew up reading. The story isn't new. The technology layer is.

The cost of waiting isn't the cost of an architect.

It's the ground you lose to a competitor who didn't wait. It's the buildout that gets twice as expensive next year because the rebuild costs more than the build. It's the operator who books the assessment call this week — and the one who books it after their next slow quarter.

Cost of one call
$197
1995–2005

The Internet Era

Casualties: Blockbuster, Borders, Yellow Pages, Tower Records.

Adapters: every business that built a website before they thought they needed one.

2005–2015

Mobile & Cloud

Casualties: Kodak, BlackBerry, hotel chains that ignored Booking.com.

Adapters: every operator who moved to SaaS, mobile-first, and the cloud.

2015–2024

Digital & Social Marketing

Casualties: print-only retailers, agencies that wouldn't touch Meta ads.

Adapters: brands that learned the new acquisition channels in time.

2024–?

The AI Layer

Casualties: TBD. The category is being written right now.

Adapters: operators who treat this like the internet in 1998 — not a fad.

If you got into business after 2010, you've never lived through one of these shifts as an owner. You're living through one now.

— 06 / The differentiator

A technical AI architect builds you a system. A business architect builds you a system that won't backfire.

The market is filling up with prompt engineers who can wire a tool but can't read a P&L. They'll deploy agents with database-delete permissions. They'll send 50,000 SMS messages without thinking about TCPA. They'll automate your sales follow-up and accidentally violate FINRA. As a CPA, ex-KPMG and Delta auditor, and Lean Six Sigma certified operator, I've spent my entire career being trained to find that exact kind of risk.

"Anyone can build an AI agent. Building one that doesn't get you fined, sued, or hacked, and that actually moves revenue, is a different job." — The job description, in one sentence
01 Compliance & Risk Lens

What an architect without finance training will miss.

Agent permissions. Database-delete authority. SMS & email regulation. Data-residency rules. Privacy compliance. Audit trails. The controls that decide whether your AI is an asset or a liability.

  • TCPA · CAN-SPAM · GDPR · CCPA
  • Agent-permission boundaries & guardrails
  • Data-loss prevention & rollback design
  • Audit trails the regulator will accept
02 Financial Lens

Build what makes money. Skip what doesn't.

What I bring as a CPA: I price every potential build by revenue impact, profit margin, and time-to-payback before we touch a tool. Half the buildouts I see in the wild are technically impressive and financially worthless.

  • ROI scoring on every proposed automation
  • Cost-of-process baseline (what is this hour costing you?)
  • Payback-period model before scope
  • Phase plans ordered by revenue impact, not novelty
03 Audit Lens

Find the actual root cause. Not the obvious symptom.

Auditors are trained to find what's wrong before someone tells them. KPMG and Delta paid me for a decade to do exactly this. Half of "we need AI" calls become "we need to fix Step 3 of your sales process and the AI you're proposing won't matter."

  • Process map → gap analysis → root-cause
  • What's the real problem under the stated one
  • Eliminate · consolidate · automate (in that order)
  • Build only what survives the diagnosis

I don't bandage. I find the root.

Anyone can map a process. Anyone can stand up an agent. I'll tell you which step of your business is actually broken, and then we'll automate, consolidate, eliminate, or rebuild it. The deliverable isn't a tool. It's a business that runs differently than it did before we started.

16yrsfinding root cause for a living (auditing since 2010)
— 07 / Who this is for

This is for you if…

If even three of these read your mind, the call's worth $197. We'll know in 30 minutes whether you're a Level 1, 2, or 3 — and whether AI is the answer or you've been sold a problem you don't have.

You went down a Claude (or ChatGPT) rabbit hole and came out more confused than when you started.

You took a cohort, watched the webinar, sat through the all-day workshop — and now you can't tell what to build first.

You know AI matters. You also know your invoicing still pings a VA. You'd like the truth on which to fix first.

You're tired of "AI consultants" who can't explain why this build pays for itself in nine months instead of nineteen.

You're worried about compliance, security, agent permissions, data — and your last quote came from someone who wasn't.

You're a solopreneur, a 5-person team, or a 200-person company. The question isn't size, it's whether you're being honest about Level 1.

You don't want a prompt engineer. You want someone who can read your P&L, your process map, and your contracts in the same hour.

You're not afraid of AI. You're afraid of choosing the wrong thing to build first and burning $40K finding out.

— 08 / Honest answers

Questions you should be asking before you hire anyone.

I just took an AI cohort. Why do I still need you?+

Cohorts teach you what's possible. They don't have time to teach you what to prioritize for your business, what compliance to worry about, what your stack actually needs, or how to keep an agent from doing something catastrophic. That's the gap I fill — between knowing AI is real and having it shipped, safely, in your business.

Why is the assessment call paid?+

Because in 30 minutes I'll save you tens of thousands of dollars in misdirected build cost. $197 keeps the call serious — both directions. It's also fully credited toward any engagement you book within 30 days, so if we work together it's effectively free.

What's the difference between you and a "regular" AI consultant?+

Most AI consultants are technical. They can stand up an agent. Few of them have ever read a P&L, owned a business, signed an enterprise contract, or sat through a compliance audit. I have. I'm a CPA, ex-KPMG and Delta auditor, Lean Six Sigma certified, and an operator who co-founded a company in 2022 and 11×'d revenue using exactly this work.

To be clear: that doesn't mean every engagement starts with a full P&L review or a deep financial diagnostic — it means the lens I apply to every question, every recommendation, and every build is shaped by finance, compliance, and operations. The questions I ask on the call come from that lens. The risks I flag come from that lens. I'm not just looking at the tool you want to build; I'm looking at the business you're building it inside.

What if I don't actually need AI?+

I'll tell you on the call. Most of the businesses I talk to don't need AI yet — they need Level 1 automation (CRM workflows, Zapier, basic system hygiene) that's been possible since 2018. If that's you, I'll say so, and you'll save your AI budget for when you're actually ready.

Will you decide my AI platform & tool stack on the call?+

I may make recommendations on the assessment call — I've worked with most of the major platforms and I'll have an instinct from the intake form. But I won't commit to a stack until I'm in the build and have actually seen how your business is wired: your data, your team's skill level, your compliance footprint, and what's already in place.

Picking a stack before doing the diagnostic is exactly how clients end up with $40K of tooling they can't use. The call gets you a directional recommendation. The build gets you the final stack — selected after I understand the system, not before.

How fast does this typically move?+

It depends on more than just which Level you're on. Three things actually drive timeline:

1. Your level. Foundation builds (Level 1) are usually 2–4 weeks. Operations builds (Level 2) run 4–8 weeks. Decision-layer builds (Level 3) are 8–16 weeks.

2. Scope & scale. Team size, volume of data, number of platforms in play, and how clearly the business already understands itself. If I have to dig to figure out what your packages are, who your customers are, or what your mission actually is, the diagnostic takes longer — because I'm not skipping foundational steps to get to a build that won't hold.

3. Stakeholder availability. If a Level-3 build needs me to interview ten stakeholders, and we can only book one meeting per week, the project takes ten weeks before I write a line of build logic. The faster you can get me information and decisions, the faster the build moves. That's the part of timeline you control.

Can you work with my existing tools, or do you push your own stack?+

I'm tool-agnostic. I've built on Zapier, Make, n8n, HubSpot, Airtable, Notion, ClickUp, Claude, ChatGPT, custom GPTs, custom skills, MCP servers — the list goes on. The right stack is the one that fits your team, your budget, and your compliance requirements — and that recommendation comes after I'm in the build, not on the first call.

Do you sign NDAs and handle sensitive data properly?+

Yes. NDAs are non-negotiable and on file before we start any engagement, before any information moves in either direction, and before I'm granted access to a single tool. Data-handling protocols are part of every build — encryption, access boundaries, agent permissions, audit logs, the works.

And — wink — don't forget my first career. Before any of this was AI architecture or operations consulting, I was an auditor at KPMG and Delta. Risk management is at the core of how I see every system. The audit lens isn't a marketing line on this page; it's the lens I look through first. Your risk is my problem before it's yours.

One call. Thirty minutes. Where you actually are.

Fill out the intake. We get on a call. By the end of it, you'll know which level you're on, what to build first, and whether I'm the right person to help you build it. If I'm not, I'll tell you that too.

$197 · 30 minutes · Credited toward any engagement