CFO+AI
FAQ

Frequently asked questions

The short answers. If you don't see your question, book 15 minutes — you'll get a straight answer, not a sales process.

Is this a full-time job search?+

No. I sell engagements, not availability. One anchor part-time CFO client at a time plus one concurrent AI build. Engagement letter caps days (up to 12/month; additional days $4,500/day, pre-approved; over 16 triggers renegotiation). No hourly billing.

Can I hire the AI build without the CFO role?+

Yes. The AI architecture & implementation ($75–150K flat, 6–12 weeks) can be a standalone engagement working alongside your existing finance, accounting and IR team. The part-time CFO and the build are also commonly combined — same principal, no handoff.

What if the portfolio company is large and needs more than one person?+

Larger assignments scale through experienced contract specialists — data engineers, prompt engineers, QA — that I direct. You still manage one person. Velarion was built this way: one principal, distributed execution. The SOW defines scope regardless of headcount.

Who owns the IP?+

Custom code, prompts, and pipeline architecture developed for the engagement remain mine; you receive a perpetual, royalty-free license to use the deliverables internally. Generalized components I create are reusable. This is stated in every SOW.

How is this different from hiring a fractional CFO firm or an AI vendor?+

Fractional CFO firms staff you with a bench — I am the bench. AI vendors sell you a tool and leave you to implement it. I do both: I can run the finance, accounting and IR function and build the automation that makes it leaner, with the judgment of someone who has done both at institutional scale.

What systems do you need access to?+

Read access to finance systems (ERP, CRM, HRIS, banking, expense), last 12 months of board decks / investor reports / lender packages / close checklists, and 30-min confidential interviews with each team member. No software to install and no hardware to buy during the diagnostic.

How do you keep AI safe around financial data and SOX?+

Five control layers: AI never executes (only drafts/flags), deterministic validation gates, immutable audit trail, segregation of duties, and weekly drift monitoring. See Controls for the full matrix — AI never touches journal approval, reconciliation sign-off, or disbursement.

Where does my data go? Cloud vs local?+

Default is cloud frontier models (Claude/GPT-4) under enterprise API agreements that contractually prohibit training on your data. That covers 99% of portfolio companies. A local LLM path is available when regulation or LP agreements explicitly require data to stay on-prem for that flow. The diagnostic recommends the right architecture per process.

What do portfolio companies actually ask you to automate first?+

Board and investor deck assembly, close and variance narratives, lender covenant packages, contract key-term extraction, AP invoice coding and routing, and IR updates. The scorecard in Results ranks each process for your company.

How fast is payback?+

Illustrative math in Results shows ~8 months on a $110K build + $45K diagnostic at mid-market volumes. Your diagnostic uses your actual hours and blended rate — the payback is specific to your numbers, not a benchmark.

Where are you based and how do you work?+

Dallas, TX. Embedded on-site or remote, per portfolio company preference. I travel as needed. Availability is portfolio-based: ~100 hours/month to clients, Velarion time protected.

How do we start?+

Book 15 minutes. If there's a fit, we start with the diagnostic ($35–50K flat, 3–4 weeks). You can stop there with the roadmap, or proceed to the build — diagnostic fee credited if within 60 days.

Still have a question? Book 15 minutes or see how to engage.