How can AI help in the monthly planning meeting?
Whether you call it S&OP or a merchandise financial planning (MFP) review, AI can build the meeting pack before it starts: what changed since the last plan, who hasn't submitted, and the biggest risks. Afterward, it reads the transcript and turns each decision into a proposed plan change with the excerpt that supports it. The planning manager approves them, and every decision is kept with its reason.
By 2030, only 5% of organizations implementing some form of supply chain planning automation will make at least 10% of their planning decisions autonomously.
Every planning organization has a monthly meeting where the plan gets agreed: demand, supply, and the financial picture, in one room. It's where the most important planning decisions get made, and where many of them get lost. Someone takes notes, someone else updates the plan a day or two later from memory, and the reasoning behind each call survives only as long as the people in the room remember it. Next month, the same questions come back.
This final article in our series on using AI in the planning process walks through one monthly cycle, before the meeting, during it, and after it, and brings together everything in the series: answering questions, showing the work, proposing changes for people to approve, and keeping the reasoning.
One meeting, different names
The monthly planning meeting goes by different names depending on the industry, but the job is the same: agree one plan and the decisions that move it.
| Industry | What it's called | What gets decided |
|---|---|---|
| Consumer goods, food and beverage, manufacturing | S&OP (sales and operations planning), or IBP (integrated business planning) in its more mature form | Forecast by product and customer, production and capacity, inventory, and the financial plan |
| Apparel, footwear, and accessories brands | MFP (merchandise financial planning) and the open-to-buy review, usually with a consensus demand step for wholesale accounts | Sales, margin, and inventory by category and channel; what to buy, chase, or cut; receipts and markdowns |
| Retailers | MFP and open-to-buy review, plus assortment and allocation | The same, plus store allocation and promotions |
In all of them, part of the meeting is agreeing the consensus demand plan: the one demand number that sales, marketing, and planning commit to. Our example is a footwear brand that sells through department stores. Its monthly meeting combines the consensus demand review with its MFP and open-to-buy review, and it runs on Microsoft Teams.
Before the meeting: the pack builds itself
Preparing for the monthly meeting usually means a planner spending a day pulling reports together. With an AI assistant connected to the planning system, the planning manager asks for the pack and gets the agenda the meeting actually needs: what moved since the last published plan and why, who hasn't submitted their input, the biggest risks, and the decisions waiting to be made.
Build the pack for tomorrow's planning review.
The spring consensus demand plan is up 4% since the September publish, mostly from Bellmont. Three decisions need the room.
- Bellmont spring +15%: buyer confirmed extra doors sales input
- Carroway's March −8%, no reason given yet needs a reason
- September actuals 11% above plan; forecast model switched to trend forecast
- PO 5120: factory cut 600 pairs and moved 1,800 to May; two styles short in April receipts
- Spring sandals 18% below the financial plan after a key account cut its order MFP
- Accept the factory's response on PO 5120?
- How much of the spring sandal open-to-buy to hold back?
- One rep hasn't submitted (Territory West, due Friday) consensus input
Illustrative. Names and numbers are made up.
During the meeting: decide, don't take notes
The team reviews the plan together on Teams, screen-sharing the planning system, and talks through each item the way they always have. The difference is that nobody has to capture the decisions by hand. The meeting is transcribed, and the decisions get picked up from the conversation afterward. People can focus on the discussion instead of the notes.
It helps to make decisions explicit as you go: "So we're accepting the split on PO 5120," or "Let's hold sandals at 80% until we see February." Clear statements like these are easy for people to remember, and easy for the AI to recognize as decisions rather than discussion.
After the meeting: from transcript to proposed changes
When the meeting ends, the planning manager gives the transcript to their own AI assistant. The transcript is read there, in the manager's own AI, and never copied into the planning system. The assistant identifies each decision, matches it to the items and numbers it affects, and drafts a proposed change for each one, with the excerpt that supports it as evidence. Topics that were discussed but not decided are listed separately, so they don't silently become changes.
Here's the transcript from today's planning review. Turn the decisions into proposed changes.
I found three decisions and one topic with no decision. Each change is drafted for your approval; nothing has changed in the plan.
- Accept the factory's response on PO 5120: April receipts 6,000 → 3,600, May 0 → 1,800 purchase orderExcerpt, 14:12, Dana R.: "We take the split. April's tight, but May still makes the floor set."
- Spring sandals: release 80% of the open-to-buy now, 3,200 of 4,000 pairs, the rest after February sell-through open-to-buyExcerpt, 22:40, Sam K.: "Hold sandals at 80% until we see February."
- Carroway's March consensus forecast: confirm 3,590 with reason "store 114 closing in May" reason addedExcerpt, 31:05, J. Patel: "That cut was store 114. They confirmed it closes in May."
- Trail boot price increase for spring: finance to bring the margin impact to next month's meeting follow-up
Illustrative. Names and numbers are made up.
Notice the third proposal. The Carroway's cut was already in the plan but had no reason. The meeting supplied one, so the assistant proposes attaching it. That's how gaps in the plan's reasoning get filled without anyone chasing them.
The same meeting at a consumer goods company
At a food or beverage brand running S&OP, the flow is identical; only the decisions change. The transcript might yield "move the retailer's summer promotion from week 31 to week 32," "add a second co-packer run for the 12-pack in August," and "hold the new flavor launch until the distributor confirms shelf space." Each becomes a proposal with its excerpt, for the S&OP owner to approve.
Approval: the planning manager decides, and the reasons stick
The planning manager reviews the batch in the planning system, edits anything the AI misread, and approves. Each change is applied as the manager's edit, with the meeting excerpt attached as its evidence and the batch name tying it to the October meeting. The plan can then be published, and the people who own each item see what was decided and why.
| Decision | Owner | Evidence | Status |
|---|---|---|---|
| Accept PO 5120 split; 1,800 pairs to May | Dana R. | Meeting excerpt, 14:12 | Applied |
| Spring sandals: release 80% of open-to-buy now, rest after February | Sam K. | Meeting excerpt, 22:40 | Applied · check in March |
| Carroway's March −8%: store 114 closing | J. Patel | Meeting excerpt, 31:05 | Reason added |
| Trail boot price increase | Finance | No decision | Next meeting |
Illustrative. Names and numbers are made up.
Next month: "What did we decide, and did it work?"
The real payoff comes at the next meeting. Because each decision was kept with its reason and evidence, the assistant can answer questions planning teams usually can't: what was decided last time, why, and how it turned out.
How did last month's planning decisions turn out?
Two of the three are on track, and one needs a decision this month:
- PO 5120 split: the 1,800 pairs are confirmed for May 2; April shortfall covered by moving stock from the West DC on track
- Carroway's March −8%: store 114 closing confirmed by the retailer's announcement on track
- Spring sandals held at 80%: early reorders running 9% above the reduced plan. Decide whether to release the remaining 800 pairs decide now
Illustrative. Names and numbers are made up.
Over a year, this builds exactly the record the footwear expert in our corporate memory article never left behind: every major planning call, why it was made, and whether it worked.
Keeping the meeting private
Planning meetings include sensitive material: customer negotiations, pricing, people. Three safeguards keep it contained:
- The transcript stays in the manager's own AI. The planning system receives only the approved changes and short excerpts as evidence, not the recording or the full transcript.
- Nothing changes without approval. Every decision the AI picks up is a proposal until the planning manager approves it.
- Permissions still apply. The assistant works with the planning manager's own access, enforced by the planning system.
Starting without AI
You can get much of the value today with a simple discipline: state each decision clearly in the meeting, keep a decision log with an owner, a reason, and the items it affects, and open each meeting by reviewing last month's decisions against actuals. AI then takes over the busywork of capturing decisions and updating the plan, and the log builds itself.
That completes our series on using AI in the planning process. Start with six practical use cases, see why corporate memory is the real asset, and read how to connect AI to your planning system so it proposes and people approve.
Frequently asked questions
How can AI help with S&OP and MFP meetings?
AI can build the meeting pack from the planning system, then turn the meeting transcript into proposed plan changes, each with the excerpt that supports it. The planning manager approves the changes, and each decision is kept with its reason for the next cycle.
What is the difference between S&OP and MFP?
Both are monthly cycles for agreeing one plan. S&OP (sales and operations planning) is used mainly by consumer goods and manufacturing companies to balance demand, supply, and finance by product and customer. MFP (merchandise financial planning) is used by apparel, footwear, and retail companies to plan sales, margin, inventory, and open-to-buy by category and channel.
What is a consensus demand plan?
The single demand forecast that sales, marketing, and planning agree to commit to, usually set in the demand review step of the monthly planning meeting. It replaces separate sales, statistical, and finance versions of demand.
Does the AI change the plan during the meeting?
No. Decisions from the meeting become proposals that the planning manager reviews and approves afterward. Topics discussed without a decision are listed separately rather than turned into changes.
Where is the meeting transcript stored?
It is read in the planning manager's own AI assistant. The planning system receives only the approved changes and short excerpts as evidence, not the full transcript or recording.