The Firm

Built for the decisions that matter

Fellowship Intelligence is a senior AI strategy and governance practice, built by people who helped shape the standards for professional AI use, not just reacting to them. It exists because the people responsible for AI decisions in most organizations are being asked to commit to outcomes they don't yet have the framework to evaluate. We show leaders the decisions they need to make, then build their people to keep making them. We are not here on a vendor's behalf to install software and leave.

Why Fellowship Intelligence exists

The current AI advisory market sorts into three categories: platforms that need you to buy software, consultancies that need you to buy implementation, and individuals offering open-ended retainers. None of these are advisory in the strict sense. They are all selling something downstream of the decision.

Fellowship Intelligence sells the decision itself. We help you scope what AI should and should not do in your business, who owns each part of it, what controls verify it, and what happens when it doesn't work as expected. We hand that operating model back to you in a form you can run, audit, and defend.

That focus is deliberate. In most organizations the tools work as configured; what goes ungoverned is what they were permitted to do, who owns the result, and how that holds up as the system and the people using it change. Why the decision.

How the firm is built

Fellowship Intelligence is founder-led. Thomas is joined by certified collaborators, and every engagement is owned end to end by the people doing the work, never handed off to a rotating bench of juniors. The deliverable is an operating model your organization runs itself, documented to hold whether or not we are in the room. The AI Sustainment Program exists to keep that model current as the technology and the rules change.

Behind the practice is Fellowship Research, the firm's internal research division. It develops and stress-tests the frameworks we deliver against, and it researches and sharpens the writing published in The Wrong Default. It builds understanding, not software.

Leadership
Thomas Tornatore, founder and Managing Director of Fellowship Intelligence

AI governance analysis for organizations that haven't asked the right questions yet.

Read The Wrong Default

Thomas Tornatore

Founder and Managing Director

Thomas built Fellowship Intelligence in 2025 after more than 25 years in organizations where accountability and structured decision-making were non-negotiable. He began in regulated financial services, completing securities licensing and advising business owners on investment, retirement, insurance, and lending, then led operations across large-scale security administration, transportation and logistics, and technology. Every one of those environments ran on the same discipline: high stakes, low tolerance for improvisation, and businesses that run on repeatable systems.

From 2017 he formalized that into a systems-first methodology for operational structure, accountability, and strategic advisory. One pattern kept surfacing: leaders adopting AI with no operating model to govern it. Fellowship Intelligence is the response. Based in Las Vegas, he works directly with COOs, General Counsel, Chiefs of Staff, and Compliance Officers, and owns every engagement: strategic direction, key decisions, and final sign-off. When senior practitioners handle execution, you always know who is doing what.

He publishes The Wrong Default, a briefing on AI governance for risk-sensitive organizations, and has two books forthcoming: Decided in Silence: How Absence Becomes a Decision, and Who Pays the Cost and Answerable: Who Owns the Decision When the Machine Has Already Made It. He also founded Still in Charge, an independent student AI-safety initiative becoming its own nonprofit, helping schools, parents, and teachers give students the judgment, privacy, and accountability that AI use requires.

Book an advisory call
Our framework

Good AI governance has to answer to standards that already exist. We didn't write our own definition of what counts as sound. We built a framework to deliver against the standards your board, auditors, insurers, and regulators already recognize.

NIST AI Risk Management Framework

Its four core functions structure how we identify, score, and assign ownership of AI risk: Govern, Map, Measure, and Manage.

ISO/IEC 42001

The international management-system standard for AI. We structure ownership, controls, and continual review around it, so governance runs as a repeatable system rather than a one-time document.

EU AI Act

Where your AI use or data touches the EU, we account for its risk-tiering and accountability obligations, so your governance is built to answer to them.

AICT Practitioner Code

The code of ethics and scope of practice for AI governance practitioners, issued by AICT (Assurance, Intelligence, Certification, Trust). We adhere to it in every engagement.

The Fellowship Intelligence AI Decision Architecture

Our proprietary operating model for delivering against those standards. It structures how we assess risk, assign ownership, and build governance that holds up over time. Not a one-time document. A repeatable system.

Credentials and affiliations
CSAP certification badge

Certified Standard AI Practitioner (CSAP)

Issued by AICT (Assurance | Intelligence | Certification | Trust). License 832335737. Thomas is one of the first practitioners certified in AI governance as a formal discipline.

IAPP

Thomas is a Professional member of the International Association of Privacy Professionals, the leading global body for privacy and AI governance.

He is currently preparing for the IAPP's Artificial Intelligence Governance Professional (AIGP) certification, which centers on the legal and regulatory dimension of AI governance.

Cited in published frameworks

Thomas's writing in The Wrong Default is credited in The Disposition Protocol (v1.1), an enterprise AI-governance framework by Sougata Roy, for surfacing its observed-outcome ("authorized but wrong") trigger and the consent-versus-live-judgment distinction.

How we operate

Decisions before deployment

Everything starts with knowing where you are and deciding where you want to be.

Fully scoped engagements

Everything is agreed up front. No unknown costs. No open-ended retainers.

The AI Sustainment Program is the moat

Governance has to evolve with the technology and the organization. We make sure it does.

Refusal is part of the job

We take a limited number of clients each quarter and decline work outside our scope. Quality depends on both.

Discipline over speed

Everything in AI moves fast. The first decisions shape everything that follows.

Spreading the word

Our founder speaks to schools, parent groups, conferences, and podcasts on responsible AI use.

Book an advisory call Take our Decision Ownership Check Discuss a speaking engagement