Start where decisions repeat. Margin per decision and decision volume pick the first lever.
Run the boundary test. If the output changes a price, a promotion, an order or a shelf, guardrails and a named owner come first.
Count what you already run. Forecasting and optimization models have run for years without the AI name; count yours before buying new.
Test retail-ready claims on your data. Ask the vendor for a specific problem already solved, run on your data.
Baseline before the pilot starts. One KPI, one process, one snapshot before day one; skip it and 12 months later the board asks what the AI yielded.
A free lesson from AI in retail and consumer brands, one of its 9 lessons. Read another free lesson
The first pay is a task
Loblaw's store-operations assistant Robin sees about a million interactions a week from store managers. It distills a flood of head-office messages into what matters today. That is the first pay: one narrow task, at volume, inside the operation. Start there. Anything that changes a price, a promotion, an order or a shelf waits for guardrails, a named owner and a measured baseline.
1. Let margin and volume pick the lever
Pick the first lever where margin per decision meets decision volume. Lauren Steinberg, Executive Vice President and Chief Digital Officer at Loblaw Companies Limited, says her company picked its early applications by traversing the business. Space planning for stores and a supply-chain booking agent came early. Each automates a single task.
2. Run the boundary test before anything goes live
Does the output change a price, a promotion, an order or a shelf? If yes, guardrails and a named owner come first; the guardrails lesson goes deeper.
Robin flags items selling at a negative margin: the price or the cost is wrong somewhere. A store manager still finds and fixes it. Andrew Laudato, Chief Operating Officer at The Vitamin Shoppe, says his distribution-center robots break down, get lost and do a little square dance. One escaped its virtual fence and was found in the back of a truck. If the output lands on a shelf, you name the person to call when it goes wrong.
3. Count what already runs
Retail has run forecasting and optimization models for years without calling them AI. The workhorse Laudato described was robotic process automation. Software robots with their own user IDs logged in daily, downloaded files and sent out mail merges. Their value is taking the most boring, standardized jobs off people. Whether RPA even counts as AI is debatable, Laudato says. Count it anyway. Andrew Lederman, VP of Global Digital Commerce at Mondelēz, says he helped build a B2B ordering app at AB InBev. Its personalized recommendations ran on machine learning when it was still called algorithmic selling. More than 70% of AB InBev's revenue was flowing through it. The lesson on counting the AI you already have goes deeper.
4. Test retail-ready claims on your data
Steinberg tells vendors to skip the platform demo, because Loblaw has to do the mental gymnastics of fitting it to its own problems. Her advice to them: arrive with a specific problem you have already solved for the retailer. Do not open by asking what keeps them up at night. Ask any vendor the same: which specific problem has it already solved for a retailer? The lesson on vendor AI, real or relabel, takes this further.
5. Measure from day one
Pick one KPI and one business process, and take a baseline snapshot before the pilot starts. Rupa Amin, Senior Director of Strategy, Retail and Consumer Goods Practice at Perficient, says what it costs to skip. Skip it, and 12 months later the board asks what the AI yielded. There is no before and after to compare. Perficient's 2026 State of the Industry study surveyed 50 director-level and above retail and consumer leaders: 20% pivoted AI into production. The financial impact of what AI did for them is missing. Perficient sells this work, so read its numbers as one firm's. The lesson on setting the baseline and the bar goes deeper.
The margin map: where the pay sits
Price and markdown: every price decision carries margin; the markdown lesson goes deep.
Stock and availability: keeping the shelf in stock; the availability lesson goes deep.
Assortment: which items earn their shelf space; the assortment lesson goes deep.
The product page: Mondelēz is rebuilding website back ends to be machine readable for the LLMs that answer shoppers; the product-content lesson goes deep.
Store labor: task lists that absorb the head-office flood; the store-operations lesson goes deep.
Media: retail media and the agent shopper; the retail-media lesson goes deep.
Shrink: loss that never reaches the till; the shrink lesson goes deep.
Exercise (20 minutes): pick and test your first workflow
Pick the AI workflow you are under most pressure to start, or the one you already think should go first, and fill in the card for it. You won't know every line yourself. Every line you can't fill gets an owner and a date before you close the card. The card download includes a filled-in example of a workflow that isn't ready, and why.
First AI workflow checklist (PDF)
A blank card and a filled-in example, one page each.
Show the three prompts: fill the card, test it, turn it into a plan
Use AI to fill in and test your card
These three prompts go into your company's approved AI assistant. Use the first to fill in the card one row at a time; it won't guess a number for you. Use the second when the card is done, to hear the case against this workflow going first. The third asks about your industry and your business, and turns the lesson into a 30-day plan, with a checklist, rubric or brief for your boss if you want one. Use placeholders for company, system and people names; the help doesn't depend on them.
Prompt 1: Fill in my card with me
Paste this, then add your workflow in one or two sentences and the card below.
Prompt
Before you start: use your company's approved AI assistant, not a personal account. Replace company, system, supplier and people names with placeholders (Company A, System X) and round any figures. The help you get does not depend on them.
You are helping me fill in a first-AI-workflow checklist for one candidate workflow in my retail or consumer-brand business. The card is pasted at the end.
Rules:
- Ask me about one row at a time, in order, and wait for my answer before moving on.
- Write down only what I tell you. Never suggest, estimate or invent a number, a name, a yes/no answer or a decision for me, even if I ask you to guess. If I ask, say that this row needs a real answer and ask who could give it.
- If I don't know an answer, write "Not known" and ask who should find out and by when. Put that in the Owner / by when column.
- After the last row, show the whole card as a table with three columns: Question, Answer, Owner / by when. Then list every row marked "Not known" with its owner, and remind me that any unanswered row means this workflow is not ready to be the first lever.
My candidate workflow in one or two sentences: [DESCRIBE THE WORKFLOW]
The card:
[PASTE THE CARD]
Prompt from The AI Decision Course by Dr. Tony Hoang, host of The Artificial Intelligence Podcast (Spotify, Apple Podcasts): tonyphoang.com/learn
Prompt 2: Stress-test my card
Paste this, then your filled card.
Prompt
Before you start: use your company's approved AI assistant, not a personal account. Replace company, system, supplier and people names with placeholders (Company A, System X) and round any figures. The review does not depend on them.
Below is a first-AI-workflow checklist I filled in for one candidate workflow. Act as a skeptical reviewer at the decision meeting and argue the case against this workflow going first.
For each row:
- Say whether the answer is specific enough to act on (a number, a named role, a date) or too vague.
- Point out where answers contradict each other, for example a yes on the boundary test against no guardrails and no named owner.
- Say what evidence you would ask for before accepting it.
Do not rewrite my answers and do not invent facts about my workflow. If a row is blank or says "Not known", treat it as not ready.
Finish with: your recommendation (FIRST LEVER, WAIT FOR GUARDRAILS or NOT READY), the three rows that most need work, and one question I should ask my team on Monday.
My card:
[PASTE YOUR FILLED CARD]
Prompt from The AI Decision Course by Dr. Tony Hoang, host of The Artificial Intelligence Podcast (Spotify, Apple Podcasts): tonyphoang.com/learn
To paste the card into prompt 1, copy its rows from the PDF.
Prompt 3: Make it yours
Paste this and answer its four questions.
Prompt
Before you start: use your company's approved AI assistant, not a personal account. Replace company, system, supplier and people names with placeholders (Company A, System X, Supplier B).
You are helping a leader apply one lesson to their own situation. The lesson's five decisions are below. Use only them and what I tell you. Never invent facts, numbers, vendors, laws or results about my organization or my industry. Where a number is needed and I have not given one, say what to measure and who should measure it, instead of guessing.
First, ask me these four questions, one at a time, and wait for each answer:
1. My industry, and roughly how many people work in my organization.
2. The AI workflow I have in mind, in one or two sentences.
3. Where it stands today: an idea, a vendor proposal, or work that already runs under another name.
4. What worries me most about it.
Then give me the following. Keep it short and practical:
A. My plan. For each of the five decisions: one action written for my workflow, the role that should own it (a role, not a name), and what should be done within 30 days. If my industry makes a decision harder, for example thin margins, regulated prices or franchise stores, say which decision and why in one sentence.
B. Three questions to ask my team on Monday, specific to my workflow.
C. Ask me which one I want: a checklist for my workflow, a rubric to score candidate workflows, or a one-page brief for my boss. Then write it, and finish with a prompt I can paste later to update it.
Mark anything that rests on an assumption with "(check)".
The five decisions:
1. Start where decisions repeat: margin per decision and decision volume pick the first lever.
2. Run the boundary test: if the output changes a price, a promotion, an order or a shelf, guardrails and a named owner come first.
3. Count what you already run: forecasting and optimization models have run for years without the AI name; count yours before buying new.
4. Test retail-ready claims on your data: ask the vendor for a specific problem already solved, run on your data.
5. Baseline before the pilot starts: one KPI, one process, one snapshot before day one; skip it and 12 months later the board asks what the AI yielded.
Prompt from The AI Decision Course by Dr. Tony Hoang, host of The Artificial Intelligence Podcast (Spotify, Apple Podcasts): tonyphoang.com/learn
Your company's assistant may answer differently, so read what it writes before you use it.
Check yourself
Two proposals wait: one rewrites product descriptions at volume, one reprices the shelf. Which goes first?
Show the answer
The descriptions; no shopper-visible price changes. The repricer waits for guardrails and a named owner.
A vendor demos a retail-ready markdown engine. What do you ask for?
Show the answer
A specific markdown problem they have already solved, run on our data.
The pilot starts Monday. What must exist today?
Show the answer
The baseline snapshot: one KPI, one process, measured before the AI starts.
Sources
Lauren Steinberg, Executive Vice President and Chief Digital Officer, Loblaw Companies Limited, on The FEED powered by Loblaw (2026): "Loblaw's AI & Digital Strategy: Lauren Steinberg, EVP & Chief Digital Officer on Grocery's Future". https://exa.ai/library/podcast/wrqn7g8wzyp/episode/fbybkbl04cn Produced by Loblaw.
Andrew Lederman, VP of Global Digital Commerce, Mondelēz, on The FMCG Guys (2026): "359. Andrew Lederman, VP of Global Digital Commerce at Mondelēz (live from Groceryshop Las Vegas): Marketing to AI agents, Why Digital Marketing is a Team Sport, Market Share as the North Star". https://exa.ai/library/podcast/9jsxfndz3dg/episode/pmr5wffk918
Rupa Amin, Senior Director of Strategy, Retail and Consumer Goods Practice, Perficient, on The Retail Razor: Data Blades Podcast (2026): "Proving AI ROI - 64% of Retail Leaders Can't". https://exa.ai/library/podcast/3djvc9p7z9b/episode/pxpfzw5tmwd Sponsored by Perficient.
Where AI pays first in retail, and the bar it must clear
Ask about this lesson, or how it applies to your team. Every answer cites the lessons it draws on.
Try asking
Free preview: 3 questions a day, answered from this lesson. Follow-ups remember your earlier questions on this visit. Don't paste confidential details.