I'm Shawn Kimble, the founder of AI·IA. Over twenty-five years I've administered, automated, migrated, consolidated, and audited the systems that run companies' money — from Financial Director & CIO of a family construction group on QuickBooks, to global multi-subsidiary ERP environments spanning dozens of entities, hundreds of users, and five countries. At nearly every stop, I found something someone hoped nobody would look at.
AI·IA exists because AI agents are now inside those same systems — executing under trusted identities, at machine speed, against controls built for humans reviewing at human speed. Almost nobody is testing what those agents will actually do. That's the job I've been training for my whole career, whether I knew it or not.
Ceramic engineering is materials science under heat. You can design a perfect body and glaze on paper, but the only truth is what comes out of the kiln. The training is process control, designed experiments, tolerance analysis, and failure forensics — systems thinking where the failure always lives at an interface.
That degree installed the reflex I've carried through every job since: don't trust what a system says it does. Instrument it, stress it, and measure what it actually did.
For twelve years I was Financial Director & CIO of Bay 10 Ventures — a vertically integrated, multi-subsidiary family business spanning custom residential construction, commercial and residential real-estate holdings, home-maintenance services, and real-estate investment, all running on QuickBooks. I owned finance, tax, risk, and HR — and got my real education in what ledgers hide. Working inside it, I uncovered intercompany fraud, payroll fraud, credit card abuse, construction-site theft, and identity fraud. Not case studies — real losses, in a real business, found by someone willing to reconcile what everyone else waved through.
It's also where I built my first real control: an Excel macro system that automated the timesheet process end to end, replacing the manual workflow where the payroll games had lived. Two lessons walked out of that business with me: fraud is ordinary, and the fix that survives is the one built into the system, not the one written in a memo.
At a management-consulting firm I analyzed vertically integrated conglomerate spend for arbitrage opportunities and scripted Automation Anywhere bots optimizing backend procedures for one of the country's largest telecoms — plus Excel macros that replaced whole repetitive business functions. That was my first look at software robots executing business processes under human credentials, at volumes no reviewer could follow: enterprise RPA, the direct ancestor of today's AI agents.
Everything AI·IA now tests for — automation identity, machine-speed execution, review controls designed for human pace — I first watched take shape there, years before anyone called it agentic.
At a global insurance-technology company I wore two hats. As Senior Operations Analyst I supported the CFO through the turnaround of six business units into a pre-IPO global technology company — managing, at peak, a team of 27 across finance and HR for roughly $100M of business, building the three-year plan ahead of acquisition by one of the world's largest private-equity firms, and reconciling consolidated financials and FX across two systems that used different consolidation methods. Consolidation is where every upstream sin surfaces: intercompany that won't eliminate, currency that won't tie, mappings nobody documented.
As ERP administrator I ran the migration off the legacy stack onto NetSuite OneWorld and OpenAir for 400 users across the United States, Canada, the United Kingdom, Australia, and Pakistan. That taught the other half: every control you have is only as strong as its survival through the next cutover. Migrations are where segregation of duties quietly dies — someone has to hold all the keys for a weekend, and the weekend has a way of lasting.
After a consulting run doing transaction-level QuickBooks-to-NetSuite migrations for global, multi-currency, multi-subsidiary companies, I spent the most recent chapter of my career — nearly a decade — running finance systems through the full private-equity lifecycle at a global, PE-backed company: NetSuite OneWorld administered through aggressive M&A, at one point covering 95+ finance users across 45+ subsidiaries; QuickBooks data migrations and NetSuite-to-NetSuite subsidiary mergers; ASC 606 revenue recognition implemented across 12 subsidiaries concurrently; segregation-of-duties and change-management policies I wrote and enforced myself; a timesheet platform serving 1,600+ users; recurring external audits and IPO readiness.
From there I moved up the stack — managing the full financial-systems portfolio (planning, close automation, expense, tax, integration middleware), then into FP&A data analytics: wiring BI directly into the ERP and putting AI to work on finance reporting. Which means I've watched the seam between IT and Finance widen from inside every layer of it — while AI tools moved into the gap faster than any control framework followed. The systems checklist on our free triage is, not coincidentally, the stack I've personally administered.
Along the way I've also been the investor. I backed a heritage snowboard brand that had everything a term sheet dreams of — nostalgia, real sales, real margins, customers who loved it — and watched a quarter-million dollars burn anyway. Not for lack of demand: growth ran faster than cash could follow, inventory ate the runway, and when I offered to take over the financials as a partner in the business, the offer wasn't accepted. The company went bankrupt with demand still on the table.
The lesson runs this whole firm: good companies don't die from lack of opportunity. They die from what the founder won't let anyone look at. It's why AI·IA sells the outside look — and why the business itself carries no inventory, grows at the speed of cash, and takes help when help is offered.
Nights and weekends, I'm a potter — a ceramic engineer come home. Glaze chemistry is designed experimentation: the math on paper, then cone 6 in the kiln to find out what's true. Every firing is a small audit. You don't grade your intentions; you grade what the heat actually did.
That's the same discipline AI·IA sells: not what your AI policy intends — what your agents actually do.
The name is a palindrome on purpose. AI delegates human judgment to systems; IA — internal audit — verifies the delegation stayed in bounds. Same discipline, read in both directions.
The risk line that drives this firm: AI can execute control-defeating actions at machine speed and volume, under a trusted user's identity, in systems built for humans to review at human speed. And companies can't fully police this themselves — conflicts of interest and org charts get in the way. The testing has to be independent, from outside, and designed for how AI actually operates.
Every finding on the free triage, every test in our packs, comes from somewhere in the story above — a fraud found, a migration survived, a control watched failing in production. This isn't a framework we read about. It's a career, productized.
Start with the free three-minute triage — it runs entirely in your browser and nothing you select leaves the page. Or go straight to a conversation.
Schedule a consultation Run the free risk triage