Strategic Work Sample

AI Operating-Model Hypothesis
for BJ's Wholesale Club

An operator's read on the public signals, and the operating system I'd build under the CIDO to turn AI activity into governed, measurable enterprise capability.

Scott Maiocchi Chief Creative Technology Officer May 2026
The brief is not "introduce AI." The brief is "turn the AI you already have into a governed enterprise capability — without breaking your margin discipline."

This is a work sample, not a pitch. It is built from what I can see publicly about BJ's today: the Head of AI Strategy & Governance posting, the two adjacent roles (AI Opportunity Development and AI Adoption & Enablement), the Simbe Tally rollout across clubs, the digital and IT modernization agenda visible in the broader transformation strategy, and the club's investments in retail media, drop-ship, and digital commerce. I have made hypotheses where I had to. I have flagged them as hypotheses.

I do not work for BJ's, and nothing here is insider knowledge. The point is to show how I read the situation, what I would build, and how I would make the tradeoffs.

My read

BJ's is not hiring an AI evangelist. It is formalizing the operating layer between AI demand, business value, governance, adoption, and delivery.

The reason I frame the work this way is because I have already had to build it: an AI council, tool registry, policy stack, training program, internal hub, and executive reporting model inside a services organization supporting regulated healthcare, insurance, and retail clients.


Read on BJ's today

The pattern in the postings

Three roles posted at once is not a coincidence. Head of AI Strategy & Governance, Lead AI Opportunity Development, and AI Adoption & Enablement Lead are the structure of an enterprise AI function, not a single hire. Read together they describe a small team being formalized under the CIDO to do four things at once: shape strategy, govern usage, find the use cases that matter, and get the workforce to adopt them.

The pattern suggests BJ's is not hiring one "AI person." It is formalizing a centralized AI Strategy & Enablement function — one lane for strategy and governance, one for opportunity development, one for adoption and enablement. The senior mandate is to make those lanes operate as one system.

Lane 1 · Strategy & Governance

The role I'm applying to. Owns the AI operating system — strategy, council, standards, roadmap, measurement, executive and board reporting.

Lane 2 · Opportunity Development

Feeds the system with disciplined use cases. Intake, scoring, business cases, hand-off to delivery teams, value realization.

Lane 3 · Adoption & Enablement

Drives the workforce side. Training, AI Champion network, productivity tools, communities of practice, adoption depth.

Partner · Build & Deploy

Product, Data, Engineering. They actually build and run the AI-enabled solutions; the function does not.

Partner · Demand & Adoption

Merchandising, Supply Chain, Club Operations, Digital, Corporate functions. They bring the problems and adopt the outcomes.

Above · CIDO & ELT

Set the bar, fund the work, and hold the function accountable for AI maturity and business impact.

What public signals suggest is already in motion

What the new AI function likely needs to formalize

A hypothesis based on public signals, not an audit. These are the components any enterprise needs to coordinate AI activity into a system — whether each is already in flight at BJ's is a question for the first 30 days.

The mandate as I read it

A hypothesis, not a verdict: the role exists to make the three lanes — strategy, opportunity development, adoption — operate as one governed enterprise system, in a way that is consistent with how BJ's already runs: disciplined, member-focused, and unforgiving on margin. Not "be the person who knows AI." The job is to build and run the operating layer underneath.


What I'd propose to build

The operating model

Five components. Each one has a published artifact, an owner, and a measurable. The shape below is a proposal — refined and re-prioritized with the existing team in the first 30 days, not imposed.

1. AI Council with five workstreams

The Council is the decision body, not a discussion forum. Each workstream has a named owner from the existing organization, a documented charter, and a quarterly outcome target. The Opportunity Development and Adoption & Enablement leads work across these workstreams; I run the Council operationally and report up to the CIDO.

Policy & Ethics

Standards, release gates, synthetic media, IP, member-data rules.

Technology & Innovation

Tool registry, vendor reviews, integrations, infrastructure readiness.

Operations & Merchandising

Inventory, pricing, demand, club ops, supply chain.

Member & Digital

Personalization, search, retail media, e-commerce, member service.

Workforce & Productivity

Adoption, AI Champion network, role-based training, productivity tools.

(Council)

Intake, prioritization, approvals, decision documentation, board reporting.

2. Intake-to-approval flow

One front door. One process. Same path whether the request is from a business team, a vendor, or a leader.

1IntakeIdea, owner, business problem, data class, proposed value.
2ReviewSecurity, Legal, Commercial. Standard turnaround.
3Council decisionFund, defer, or decline. Documented.
4Scorecard + handoffSuccess and kill criteria, review cadence, named delivery owner in Product / Data / Engineering.

3. The AI Hub — the front door, not a wiki

One destination. Approved tools by discipline, current policy stack, council decisions, training, and the status of in-flight pilots. Usage analytics feed back into what the Hub surfaces next. Treated as a product, not a SharePoint.

4. The policy stack

Five short documents, each owned by a named lane, refreshed quarterly. None of them lives in a vacuum — every staff role is taught the parts that affect them.

5. Measurement, not theater

Every initiative gets translated into the language a CFO will recognize: hours recovered, decisions accelerated, lift delivered, risk avoided. We do not report "engagement with AI." We report what changed in the business.


Priority use cases

Where I'd push first

Ten candidate use cases scored against three dimensions: value (how much it moves the P&L), feasibility (how realistic in a 6–12 month window), and tight-margin fit (does it earn its keep in a club retailer's cost discipline). 1–5 scale. Top three flagged for first funding.

This is illustrative, not prescriptive — built from the public picture of BJ's. The real list would be re-scored with you in the first 30 days.

Use case Value Feasibility Tight-margin fit Note
Retail media targeting & measurement Fund first 5 4 5 Revenue-positive, brand-funded, measurable. Lowest-risk early win.
Inventory & shelf intelligence (Tally signal expansion) Fund first 5 3 5 Data already produced by Tally. Governance/data-pipeline problem first, model problem second.
Pricing & promotion optimization Fund first 5 3 5 Direct margin lever. Needs guardrails around member-trust and promo-rule integrity.
Demand forecasting (DC and club level) 4 3 4 Likely already in motion in pockets. Council role is to coordinate, not duplicate.
Member personalization (digital, app, email) 4 3 4 Retail-media-adjacent. Needs strong member-data rules and a clear opt-out posture.
Private-brand content automation (Berkley & Jensen, Wellsley Farms) 3 4 4 Direct adjacency to work I've done at CVS Health and Staples. Fast ROI on cost-per-asset.
Member service triage & deflection 3 4 3 Useful, but member-trust risk is real. Phase carefully.
Workforce productivity (governed ChatGPT / Copilot / Claude) 3 5 3 Useful, easy. Risk: overspend on tools without measurable lift. Cap and measure tightly.
Finance & procurement automation 3 3 3 Slow but compounding. Fold under existing Finance modernization, not a standalone AI bet.
Labor planning / club-staffing optimization 4 2 4 High value, high sensitivity. Needs HR, legal, and culture alignment before pilot.

First 90 days

What I'd do

Three phases. Each ends with a published artifact and a measurable. Illustrative — the actual sequencing would be re-shaped against what is already in flight at BJ's once I have the inside read.

Days 1–30
Listen
Map what's already happening.
  • One-on-ones with the CIDO, the two AI lead hires, Finance, Legal, Risk, HR, Merchandising, Member Marketing, Digital, Data, Product, and Club Operations.
  • Inventory current AI work, current tool usage, current policy state, current risk exposure.
  • First-pass scoring of the use-case list above against your actual P&L, not my guess.

Deliverable: AI Maturity Baseline memo for the CIDO and the AI Council.

Days 31–60
Stand up
Get the operating model running.
  • Council chartered with named workstream owners. First Council meeting on the calendar.
  • Intake-to-approval flow live. AI Hub v1 published.
  • Internal AI Policy v1 and Staff Quick Guide approved by Legal and IT, rolled out with role-based briefings.

Deliverable: working Council, working front door, published policy stack v1.

Days 61–90
Prove
First three pilots scored, funded, and starting.
  • Pilot scorecards published for the three "fund first" use cases (or whatever you and I decide actually wins).
  • First quarterly report to the CIDO and the executive committee: AI maturity, adoption depth, and projected business value.
  • AI Champion network seeded across the corporate functions.

Deliverable: three pilots in flight with documented kill criteria. First measurable already in motion.


Workforce architecture

Redesign, not reduction

Most enterprise AI conversation treats adoption as a software rollout. The harder, more valuable work is workforce architecture — deciding how roles get reshaped, which ones compound with AI, which ones expand, and which new ones appear. The BJ's JD asks for "a talent and organizational impact plan for agentic AI adoption, including identifying skills gaps, workforce changes, and potential new roles." That is the workforce-architecture question. Here is the lens I use.

Frame 1

Company maturity

  • AI emergentExperimenting with tools, pilots, and pockets of productivity. Most large enterprises today.
  • AI nativeRebuilding workflows, roles, governance, and operating models around AI. A small but growing group.
  • AI laggingTreats AI as optional or incremental while competitors compound advantage. The exposure a club retailer in a tight-margin category cannot afford to drift into.
Frame 2

Workforce archetypes

  • BuildersDesign, configure, and improve AI-enabled systems. Tech, data, product, engineering, parts of digital and marketing operations.
  • Agent managersOrchestrate and supervise AI-enabled work. Merchandising, marketing, HR, finance, legal, club operations — the corporate and operational middle layer.
  • Front-linersOwn the human edge with members, employees, and partners. Club teams, member service, field leadership.
  • StewardsGovern, audit, and decide whether to scale, pause, or kill. The AI Council, this role, and the partners across legal, risk, IT, and finance.

The synthesis: AI-native companies are not just adopting tools — they are redesigning work around new role types. The executive question for BJ's is not "who gets ChatGPT?" It is "which functions should become builders, where do we grow agent managers, which roles stay front-liners, and who stewards the whole thing?" That mapping decides where AI compounds value and where it just generates risk.

The agent-manager category is the value-and-adoption center

In a retailer, the agent-manager category — the corporate and operational middle layer running merchandising, marketing ops, finance, HR, legal, and club operations — is where much of the near-term value sits. It is also where adoption will need the most care, because these roles are closest to the work AI will reshape. Underestimating either side stalls the program.

How I'd communicate it across 30K employees

Sixteen years of cross-functional leadership through ambiguous, high-stakes change in U.S. Army Civil Affairs is where I learned how to communicate disciplined change without false comfort and without fear. The same posture applies here — and it matters more in a 30,000-person club retailer than in any single creative-services organization.


Adoption proof

Access is not readiness

AI adoption is emotional before it is technical. In a recent five-hour AI enablement session I led for cross-functional agency teams, the strongest feedback was not simply that people liked the session. It was that people who were already using AI every day still needed the bigger picture made clear.

One client-facing leader wrote that the session made AI feel "exciting instead of daunting" and helped the team move from "dancing around" client AI questions to feeling ready to lead the conversation. Another senior leader noted that the room stayed "engaged and energized" for the full five hours.

That is the work of enterprise AI adoption: not just giving people access to tools, but giving them the confidence, governance, examples, and language to use AI responsibly in the flow of business.


Sample artifact

What a governed pilot looks like

One page. Standardized. Forced to declare success criteria and kill criteria up front. This is the document that turns an exciting idea into a fundable initiative, or correctly kills it.

Pilot Scorecard — Sample

PilotRetail media targeting model v1
SponsorMember & Digital workstream lead
Business problemBrand partners need stronger targeting and measurement to renew at higher CPMs.
Data classMember behavioral + transactional. Confidential. No PII to third-party tools.
Success criteria≥ 12% lift in measured campaign conversion vs. control over 8 weeks; ≥ 2 brand partners agree to expand.
Kill criteria< 5% lift at week 6; OR any member-data exception flagged by Privacy; OR cost-per-incremental-conversion exceeds existing baseline.
Review cadenceWeekly with sponsor; biweekly Council update; full readout at week 8.
Decision rightsCouncil recommends; CIDO and CMO co-approve scale, pivot, or shutdown.

The pilot scorecard is one of seven artifacts in the policy stack. Full set available on request: intake form, council decision template, tool-approval workflow, data-classification matrix, release gate, synthetic-media review, and the staff quick guide.


About the author

Scott Maiocchi

Chief Creative Technology Officer at (add)ventures and Door G. Author of AI Ops, the policy stack and operating model the agency uses to govern AI for healthcare, insurance, and retail clients — including 15+ years on CVS Health across Private Brand, retail pharmacy, in-store, digital, and corporate communications.

30+ years leading creative, content, and emerging-technology programs at enterprise scale. 7× Emmy. 16 years U.S. Army Civil Affairs.

scottmaiocchi@gmail.com · 401-301-3778 · linkedin.com/in/scottmaiocchi