Global Institute of Data Science
If you can't, the model won't save it. The A.I.R. Framework finds the real problem before your board does.
Request a MeetingA.I.R. Framework™ — U.S. Copyright Registered · Cross-Industry Advisory: Financial Services, Insurance, Marketing, Manufacturing, Retail, Logistics & Legal · Founder's background: Top-Tier Hedge Funds & Bulge-Bracket Banks · Keynote Speaker & Judge, Top-Ranked University Master's Programs · Top-Ranked University AI Strategy Instructor
The Upstream Problem
RAND
0%
of enterprise AI projects fail to deliver their promised value
MIT
0%
of generative AI pilots produce zero measurable P&L impact
S&P Global
0%
of companies scrapped most of their AI initiatives in 2025 alone
These are not data quality problems. They are not talent problems. They are not vendor problems. They are upstream problems.
In every failed initiative, the same thing was skipped: someone defining precisely which business problem the AI must solve, for whom, and how you will know when it worked. That is problem formulation. It is the least glamorous skill in the building. It is also the one that determines everything.
I have seen this across 20+ years in quantitative finance. A model on the wrong problem does not produce a bad slide deck. It produces a loss. The cost of a wrong AI initiative in a Fortune 500 company is the same kind of event. It just moves slower.
At Future Alpha 2026 — a room of quants, portfolio managers, CIOs, and heads of risk — the most-requested talk was not about architecture. It was about how to state a problem clearly.
The Situations We Walk Into
The pilot that never left the room.
Your team ran a proof of concept. It worked in the demo. The data scientists were proud. Then someone asked what system it would actually run on. Who would maintain it. Whether the compliance team had signed off. The project went quiet. Three months later, someone filed it under "learnings."
The board asked a question you couldn't answer.
They didn't ask about the model. They asked: what does this produce for us, by when, and how will we know? You had three answers. None of them were the same. The meeting ended. A follow-up was scheduled. The follow-up was rescheduled.
The initiative that had five sponsors and no owner.
Legal said maybe. Finance said not yet. IT said they needed six months. The business unit said they needed it now. Everyone was right. No one was in charge. It is still happening, technically. No one talks about it at the weekly.
The vendor who sold you a solution before hearing the problem.
They came in with a demo. The demo was impressive. You signed the contract. Six months later, the tool was doing exactly what it was designed to do — just not for the problem your organization actually had.
"The hype on LinkedIn says everything has changed, but in our operations, nothing fundamental has shifted."
— COO, mid-market manufacturing firm. MIT Project NANDA research.
The Decision Intelligence Spectrum
When an organization gets the problem formulation right, everything downstream becomes easier. The data team builds the right thing. The board gets answers they can act on.
Brain-Driven
Pure intuition, no data. Fast and confident — frequently wrong in complex systems.
Data-Informed ← The goal
Human judgment shaped by relevant data, not replaced by it. Use what is useful. Discard what is not.
Data-Driven
Outputs without judgment. When the model is wrong — and eventually it is — no one knows how to course-correct.
The diagram maps four positions. At one end: Brain-Driven — pure intuition, no data. At the other: Data-Driven — pure automation, no judgment. In between is Data-Informed — where human judgment is shaped by relevant data, not replaced by it. Above all three sits Augmented Intelligence.
What lifts an organization there is a state GIDS calls Cognitive Clarity: knowing exactly what your AI should do, why, and how you will measure whether it worked — the state the Assess phase exists to create.
Proof
Insurance · Enterprise
Wrong model class selected — the prior failure was a model-selection problem, not a data or talent failure.
C-suite plan to roll out sales forecasting across all subsidiaries. An earlier effort had already failed in production. GIDS identified the failure point and re-grounded the initiative on deep learning time-series modeling.
Investment Banking · Global
Structured data only — an oversimplification that left a direct P&L insight source untouched.
Organization resisted applying AI beyond structured data. GIDS identified that data transformations on trader user files could surface systematic P&L insight tied directly to the department P&L.
Digital Marketing · Mid-Market
"We told a client to spend less on AI." A right-sizing problem — a leaner solution was the better strategy.
About to invest in an expensive LLM solution. GIDS identified a right-sizing problem and pivoted to a leaner agentic recommendation engine — opening a new revenue channel.
The A.I.R. Framework™
Most AI initiatives start with the solution. A timeline is set. A vendor is selected. Then someone asks what problem this solves. The answer is not specific enough. But the timeline is locked. So everyone moves forward.
This is not incompetence. It is the default setting of any organization under pressure to "do AI."
Hover each phase to explore →
A.I.R. Framework™ · U.S. Copyright Registered
Assess
Organizational Readiness
Where everything begins — before any tool is selected, before any budget is approved. It asks one question:
"Can you describe your AI initiative in one sentence that a smart, non-technical colleague would push back on?"
Implement
Evidence-Based Deployment
What happens when Assess is done. Not before. When you know exactly what you are building, for whom, and how success is measured, implementation is disciplined execution — not an expensive experiment.
Revolve
Continuous Optimization
The discipline most enterprises skip. AI that does not continuously improve becomes a liability. Revolve closes the loop: systematic review, adaptation, and documented learning.
The organization that wins is not the one with the best model. It is the one that asked the right question first.
Supported by the major studies on enterprise AI failure published in the last three years.
The Engagement
What it is.
A Feasibility Assessment begins with one question: can you describe your AI initiative in a single sentence a colleague would push back on?
The 5-Step Assessment covers data readiness, stakeholder alignment, use case clarity, ROI definition, and deployment risk.
The output is a board-ready feasibility report — exactly what is ready, what is not, and what to do before anyone writes a line of code.
Most organizations use this report to stop a project that would have failed — or to accelerate one held back by the wrong concerns.
What it is not.
GIDS does not build AI systems.
GIDS does not manage your data infrastructure.
GIDS does not replace your data team.
GIDS does not validate decisions already made.
GIDS is not a fit if you need a vendor to confirm a direction already chosen.
John Thomas Foxworthy
Founder & CEO · GIDS
Most AI consultants started in software. John started in econometrics — where the cost of a wrong model is not a missed deadline. It's a number.
That discipline became a decade of applied machine-learning consulting, then three years teaching computer vision and deep learning, and ultimately an AI-governance and model-evaluation practice built for the organizations where it matters most.
Creator of the A.I.R. Framework™️, U.S. copyright registered. Author of The Augmented Intelligence Revolution (Fall 2026).
M.S. Data Science, Northwestern · B.A. Economics, UCLA
Luis Varela
President · GIDS
Most AI governance advisors know the framework. Luis knows what happens when the framework meets a legacy system that was never designed to support it.
Nearly a decade at Cisco, then cybersecurity leadership at JPMorgan Chase and Accenture, then AI governance and transformation — with a Carnegie Mellon master's and three certificates as the academic spine running through all of it. His work covers the operational layer: how AI initiatives are governed, secured, and built to last past the first quarter.
M.S. Information Technology, Carnegie Mellon
Certificate in Artificial Intelligence, MIT Professional Education
CDAIO Certificate, Carnegie Mellon Heinz
The first conversation is a diagnosis, not a pitch. John reviews your situation before the call. If GIDS is not the right fit, he will tell you what is.
Institutional Partnerships
GIDS works with organizations that need AI competency built across a workforce, a curriculum, or a country. This is not individual professional development. It is institutional capacity-building.
The A.I.R. Framework™ is the foundation. Scalable, evidence-based, and built to complement existing academic credentials, not compete with them.
FAQ
"Can you state your AI initiative in a single sentence that a smart, non-technical colleague would push back on?"
— John Thomas Foxworthy · Future Alpha 2026 · Top 3 keynote from 120+ speakers
