AI fluency instruction · Vendor-neutral AI evaluation
I'm an independent AI practitioner and advisor. Since 2023 I've been building AI solutions, deploying them into production, and teaching the people who use them. I work directly with multidisciplinary teams across procurement, legal ops, sales enablement, finance, and operations.
AI tools I built at Accenture were named as strengths by Everest Group
300+ practitioners using AI daily within 6 months of my first deployment
The Problem
Here's what I see happening: companies invest in generic AI training, hire consultants to redesign the work in slide decks, then bring in managed services to run the tools. Each step looks sensible. Together they deepen vendor dependence.
AI training that sits outside a team's actual work never sticks. Tools bought or built without the people who use them don't get trusted, and nobody in-house can keep them current once the vendor moves on. AI capabilities shift every few weeks. The only way to keep up is to teach your people to think and build with AI themselves.
Invest in your people first. That investment compounds as AI advances, and it protects your institutional intelligence.
The Approach
Hands-on, in-context practice is what builds AI instincts.
My cohorts start using AI on real work by Day 2, including people who've never touched it. As they get better, their output quality rises, and a good number of them start building their own agents and solutions.
Applying AI to today's work is a start, but AI transformation comes when the work itself gets redesigned around what AI now makes possible. That redesign only happens from inside the work, done by the people who own it.
AI hype gets people burned. AI doom keeps them on the sidelines. Neither builds skill. I show teams where AI earns trust and where it doesn't. I work as a player-coach, in the trenches, building real tools alongside the team while they learn to build their own. The team keeps the tools and the know-how that made them.
How It Works
Anyone can develop AI fluency. I teach the method I built by teaching myself.
Engagements are scoped to where your team is; they can start at any of the three stages below, in remote, in-person, or hybrid form. Remotely I work with up to seven people at a time; in person, up to fifteen, so every participant gets hands-on instruction. Larger teams run as multiple cohorts.
An initial phase runs two weeks or longer, on a syllabus built from what your team is actually delivering, using tools you already own. Office hours continue after formal sessions end, for the moment your people apply the teaching on their own and hit the first snag.
The value of any AI tool is capped by the fluency of the people using it, and the value of any fluency program is capped by whether the tools fit the work. Selection and instruction are two halves of one investment, so evaluating AI products and services is its own engagement: an AI purchase on the table, a consulting proposal to vet, or tools already owned that the team has outgrown. Enterprises are sold to constantly and rarely equipped to cut through the noise. I am, and I take no side while doing it.
What changes: Your team learns AI on the work already in front of them, not generic exercises.
How it happens: For instance, a contract lead doesn't study "how to write prompts." They use AI to draft a clause for a negotiation they're actually running, refine it, turn it into an agent, and rebuild their process around it. By the end of this stage, the team has working tools they built themselves.
What changes: AI instincts develop under real work with an experienced builder alongside. I work with your team on live deliverables, catching weak spots and raising output quality as the work happens.
How it happens: Together we find brittle solutions as they happen and push toward harder use cases as fluency grows. When AI disappoints, I show teams why it failed and how to fix it, instead of concluding AI doesn't work and walking away.
What changes: A few people will show the aptitude to go past foundational fluency. They evolve to build real applications with agentic coding tools.
How it happens: Anyone can vibe code their way to a demo. The critical skill is building solutions that hold up: properly scoped, structured, debugged, and maintainable. I learned to engineer by building with AI, and your internal builders will learn the same way, becoming the people your organization turns to first.
Your team ends with tools that fit the work, the understanding to keep improving them, and people who own the result.
Proof
My method was proven inside one of the world's largest consultancies.
I earned my AI credentials at Accenture, where I started building with AI in 2023 and designed and built two production AI products single-handedly: S2C IQ augments sourcing and contracting practitioners across the source-to-contract cycle, including negotiation simulation. TechTariff IQ models real-time geopolitical risk for adaptive sourcing. I taught the people who would use them: over 300 practitioners using AI daily within six months, serving 120+ enterprise clients. When Everest Group assessed Accenture's procurement capability in 2025, both tools were named among its strengths. Read the Everest Group report
Solutions Showcase
A sample of what I've built and taught, from sourcing to sales to LLM evaluation.
Everything below comes from the same practice: sitting with a specific team, learning how they work, and building what they need.
Fact-checks suppliers against sourcing criteria for faster, defensible diligence. It cuts research time and catches unsupported supplier claims before they become a due-diligence problem.
Guides legal teams drafting contracts by checking alignment with legal and contract playbooks. It reduces missed approvals and escalation steps, and keeps risky redlines from slipping through untriaged.
Drafts RFI/RFP answers for sales teams using only approved company content. It speeds up proposal responses, tailors them to each client, and keeps unsupported claims out of client-facing proposals.
Helps nonprofit founders turn funder priorities and program details into grant-ready material, from letters of inquiry to full proposals and budget narratives. It finds the strongest alignment between a funder's goals and the organization's work, follows each application's required structure exactly, and replaces scattered notes with a consistent, credible case for support.
Checks whether a project update actually holds up against the documents, facts, and figures behind it, and returns specific fixes. It cuts the back-and-forth between reviewers and project owners and stops weak updates from moving forward.
Turns demand forecasts into required-staffing views for client delivery and operations. It replaces ad hoc spreadsheet guesswork with a repeatable staffing model, surfacing capacity gaps before they become delivery risk.
A thinking partner for team leads coaching frontline agents whose day-to-day performance drives call center KPIs. It helps leads diagnose performance issues instead of settling for surface-level explanations, and cuts down on unnecessary escalations.
Blind-scores model outputs to evaluate different AI models. It replaces gut-feel model comparisons with a consistent scoring method, so rollout decisions are defensible.
Structured evaluation of AI products and services against your actual use cases, not vendor demos. My two decades in technology sourcing meet a market full of hype: it stops teams from over-buying, catches inflated claims before they reach a signature, and builds the evaluation discipline in-house so the next selection doesn't need me. I take no fees, commissions, or partnerships from any AI vendor.
How I Think
AI must extend your thinking, not replace it. The work of making sure it does is what keeps your thinking your own.
FAQ
A: You bring me in to make your team independent. If you bring me back later, it should be for new work or new teams.
A: Two things: your people need to be relentlessly curious and willing to change how they work. And leadership needs to give them psychological safety to experiment and fail without punishing it.
A: When a team wants a finished tool handed over without learning how to own it. Or when there's no real appetite to question how things are done today. Neither of those is a fit, and I'd rather tell you that upfront than take the engagement anyway.
A: Yes, and this is where my two backgrounds meet. Before I built AI solutions, for two decades I advised clients on IT investments, so I evaluate AI products the way a procurement leader does and the way an AI builder does, at the same time: what the vendor claims, what the technology actually does, and what it will deliver for what it costs to own. I have helped ensure clients don't over-buy or, worse, invest in vendor hype. And my advice has no side: I take no fees, commissions, or partnerships from any AI vendor.
A: We do an honest fit check on your use cases, your tool environment, and what your organization can actually commit to. If the conditions aren't there, I'll tell you. A bad-fit engagement helps no one, including me.
A: The method transfers because it starts from your team's work, whether that's procurement, legal ops, sales enablement, finance, or operations. What it depends on is the conditions in the fit check: curious people and leadership cover to experiment. The function has never been the constraint, but readiness sometimes is.
A: Everything I do is executed alongside your people. The skill spreads through the team via hands-on practice.
A: Whichever fits the work, starting with the tools you already have. I teach how to think with AI, not how to drive one vendor's product. When the tools change, your people will keep up. I build and work daily with tools from the leading AI labs, OpenAI, Anthropic, and Google, alongside open-source models.
A: We scope an initial phase against your priorities, with defined outcomes agreed upfront. It can be as short as two weeks. At the end of that phase we jointly decide whether a further phase adds value or whether the team is ready to run without me. The engagement is designed to end.
A: No. If I build them for you, you're back to renting. We build together so your people can keep improving the tools through ownership. And when a solution needs to scale across the enterprise, your IT or engineering team takes it there; I build for handover, not dependency.
Start Here
I work with leaders who own a function's output: heads of procurement, legal ops, sales enablement, finance, and operations.
If your team has real work to build on and you're ready to change how it operates, let's talk.
And if you're facing an AI purchase decision, or own tools that aren't paying off, an unconflicted analysis is worth having before you spend more.
First step: a short chat, and an honest read on fit.
Start a conversationI also write about what's working in AI, what isn't, and what I'm still figuring out.