Chat with your Data searches your Recurrency data — inventory, demand, purchasing, and sales — and answers in plain English. A good prompt tells it what to look for, where to look, and what a useful answer looks like. A vague prompt still returns something; it just may not be the cut you needed — and very broad questions can time out.
This guide covers what belongs in a prompt, what not to ask, and a library of prompts you can copy.
💡 Closed Beta Feature. Pair this with Chat with your Data for where to find the feature.
Why a good prompt matters
Chat with your Data has to interpret your question, search your Recurrency instance, and explain the result. If the prompt leaves gaps, it fills them in. Those guesses are where answers get generic or off-target — for example treating OP/OQ items as min/max, or explaining a min change as if you asked to convert replenishment methods.
A strong prompt usually names the pieces that would change the answer:
What — overstock, dead stock, fill rate, recommended min, EOQ, a specific item
Where — location, supplier, buyer, product group. “Company-wide” is fine; leaving it blank is a guess
When — last 90 days, MTD, next 7 days
Which items — A items, stockable, min/max vs OP/OQ, included in planning
How to show it — top 10, by branch, a downloadable table, and why
You do not need every piece every time. You need the ones that would change the result.
Copy-paste template
Use this when you’re not sure how to start:
For [location / all locations], show [what] for [item scope] over [time window]. Rank by [impact]. Limit to [top N]. Explain why, and give me a downloadable table.
💡 Tip: If the prompt would be too vague for your best buyer, add a location, time window, or “and explain why.”
Vague vs specific
Vague | Specific |
overstock | Which locations appear overstocked? Which items have the largest excess inventory? |
dead stock | How much dead stock do we currently own, and which vendors create the most of it? |
stockouts | Which items stocked out the most this year, and which of those look preventable? |
min max | Which items have mins that appear too low, and which recommendations have the biggest financial impact? |
how are we doing | Give me an executive summary of our inventory performance over the last 90 days. What changed the most since last month? |
what do our rules do | Which validation rules triggered the most items in the last 30 days? Show the top 5 by item count and inventory $ impacted. |
analyze everything | Give me an executive summary for [location] over the last 90 days. Top 3 problems and top 3 opportunities only. |
The specific versions take a few extra words. They save a follow-up, and the answer is easier to trust.
What not to ask
These patterns usually return weak answers — or time out — because they ask Chat with your Data to scan too much at once.
Too broad / no scope
“How’s inventory looking?”
“Tell me everything about our inventory.”
“Analyze our whole business.”
“What should we do?”
“Give me a full report on all items.”
Ask instead: Give me an executive summary of inventory performance for the last 90 days at [branch]. Top 3 problems and top 3 opportunities only.
Vague “explain this rule / setting” with no ID
“What do our rules do?”
“Explain all validation rules.”
“Why is Recurrency recommending changes?”
“How does forecasting work?”
Ask instead:
Which validation rules triggered the most items in the last 30 days? Show the top 5 rules by item count and total inventory $ impacted.
For rule [rule name], how many items did it hit last month, and what did it change?
I’m looking at item [item] at [location]. Why is Recurrency recommending this min/max change? Show the data and assumptions — min/max only, not OP/OQ.
Mixing topics in one prompt
Don’t ask for stockouts, dead stock, EOQs, and transfers for all suppliers in one message.
Ask instead: One topic per chat. Finish stockouts → New chat → start dead stock.
Missing the replenishment method
“Which mins are wrong?” when you mostly run OP/OQ will send the search down the wrong path.
Ask instead: Only include items currently on min/max (not OP/OQ). Which mins look too low at [location] based on the last 90 days of demand?
Asking Chat with your Data to take action
“Update my mins.” / “Create the POs.” / “Turn on this rule.”
Chat with your Data is read-only. Use Planning, Daily Workflow, Purchasing, or Rule Builder to act. Ask Chat with your Data to find and explain, then confirm in the product.
Be as specific as you’d be with a teammate
A useful mental model: you sat down with an inventory analyst who can see Recurrency but does not know what you already have in your head.
You would not walk up and say “inventory.” You would say:
I’m reviewing Dallas A items this morning. Which ones are going to stock out before the next PO lands, and what’s driving it — usage, lead time, or an open order that slipped?
That sentence does five jobs:
Intent — reviewing stockout risk, not browsing.
Scope — Dallas, A items.
Time — before the next PO.
Output — which items, and why.
Constraints — usage vs lead time vs late PO.
You do not need a perfect sentence. You do need enough of those five so the search is pointed at the right slice of data.
Follow up in the same conversation
The first prompt finds the area. Follow-ups are how you get to a decision.
A productive thread usually looks like this:
Ask what is true today (current mins, EOQs, fill rate, dead stock).
Ask where it doesn’t match history or recommendations.
Ask what to change first, and why.
Ask for the impact or a what-if.
Click into Planning and confirm before you act.
Stay on one topic per chat. If you switch from EOQ to a single item’s lead time, start a new conversation.
Worked example: Economic Order Quantity (EOQ)
Copy the first line, then the next, in the same thread:
What EOQ values are we currently using?
Which EOQs don’t align with our purchasing history?
Which suppliers appear to have EOQs that are too low? Too high?
What EOQ changes would you recommend, and which five suppliers should we review first?
What is the expected impact of changing EOQ — inventory, service level, and buyer workload?
Which suppliers appear constrained by freight minimums instead of EOQ?
A one-word prompt of EOQ will give you a starting cut. This sequence gets you something you can take to a supplier meeting.
Worked example: A recommendation you don’t trust
Vague: why min max changing
Strong, first turn: I’m reviewing items where Recurrency is recommending a lower min and a higher max. Show the three min/max items (not OP/OQ) with the biggest min decrease and the largest max increase, and explain why for each.
If the answer is still wrong, say what was wrong — don’t just rephrase:
Those three are OP/OQ. I only want items currently on min/max. Redo the ranking and the explanation. Don’t tell me what would happen if I converted replenishment methods.
Follow with the “explain this” prompts in the library below.
Quick follow-up script (same chat):
Filter to [location / supplier / A items only].
Rank the top 10 by $ impact.
For the #1 item, explain why and show the calculation.
What should I do first this week — accept recommendation, change min, transfer, or talk to supplier?
Give me a downloadable table of the final list.
Use the page that’s already open
If you are staring at an item, a filtered Planning view, or Daily Workflow, open the page assistant (chat icon, upper right). The green bar — Reading the page you have open — means it can interpret “this item,” “these filters,” and “why is this flagged?”
Then ask:
Why are you recommending this change? What data did you use, what assumptions are you making, and why is it different from our current settings?
Start a new chat (or hit the eraser) when you change topics.
Tell it your house rules in the first message if they change the answer: we plan on min/max not OP/OQ; Dallas is the hub; lead time is a company-average override on purpose; ignore locations we don’t plan.
When the answer is wrong (it will be sometimes)
This is a beta AI feature. It can mix up min/max with OP/OQ, blame the wrong driver, or sound confident either way.
Don’t accept a narrative that doesn’t match Planning.
Correct it in the thread with the fact it missed.
Ask it to show its work: “Show me the calculations behind this recommendation.”
Click the item link and confirm before you change anything.
Start a New chat if the thread has gone sideways.
Chat with your Data cannot update Recurrency or your ERP. A wrong answer cannot silently change a min. A wrong answer can send you to the wrong item if you don’t verify.
It also cannot create purchase orders, accept recommendations, or deploy Rules. Use Planning, Daily Workflow, Purchasing, or Rule Builder to act.
Prompt library
Copy one prompt at a time. Swap bracketed fields like [location] or [supplier] with your values. After the answer, pick the next prompt in the same group so the search stays on that topic.
Ask for a downloadable table when you need to take the list into a meeting.
Open a topic below to copy prompts.
Morning / daily buyer triage
Morning / daily buyer triage
For [location], what should I work in Daily Workflow today? Top 15 by risk of stockout in the next 14 days, stockable A/B items only.
Which of my assigned items at [location] will stock out before the next expected receipt? List item, days of supply, open PO due date, and recommended action.
Show items at [location] where on-hand + on-order is below min right now. Top 20 by extended cost.
Which recommendations for [buyer name / supplier] have the biggest $ impact if I accept them this week? Top 10 only.
What changed since yesterday for [location] — new stockout risks, new excess flags, and large min/max recommendation swings?
Use Executive summary when you need a starting picture, then immediately follow up on one problem.
Executive summary
Executive summary
Give me an executive summary of our inventory performance over the last 90 days at [location / all locations].
What are the three biggest inventory problems our business is facing today?
What are the three biggest opportunities to improve inventory performance?
What should I focus on this week at [location]?
What changed the most since last month?
What should my purchasing team be paying attention to today?
If you were my inventory consultant, what would be your top recommendations for [location]?
What metrics concern you the most?
Which KPIs are improving? Which are trending in the wrong direction?
Inventory value
Inventory value
Where is most of our inventory investment tied up at [location / company-wide]?
Which product categories have the highest inventory value?
Which vendors account for the most inventory dollars?
Which branches carry the highest inventory investment?
Which locations appear overstocked? Which appear understocked?
How much inventory hasn’t sold in the past 12 months at [location]?
How much dead stock do we currently own at [location]? Break out by supplier, top 20 by inventory $.
Which items have the largest excess inventory?
If we reduced inventory by 10% at [location], where should we start?
The last Forecasting prompt is best from the page assistant while you have the item open.
Forecasting
Forecasting
Which forecasts are least accurate at [location] over the last 90 days?
Which items experienced unexpected demand spikes in the last 60 days?
Which items have highly seasonal demand?
Which forecasts should we manually review?
What demand patterns changed this month at [location]?
Which items recently shifted from stable demand to erratic demand?
Which forecasts appear inflated? Which appear too conservative?
Which items would benefit from exceptional usage adjustments?
Explain why you’re forecasting demand for this item.
EOQ
EOQ
Current state
What EOQ values are we currently using for supplier [supplier]?
Which suppliers have the highest EOQs? The lowest?
Which EOQs appear unreasonable?
Which EOQs don’t align with our purchasing history?
Recommendations
What EOQ changes would you recommend?
Which five suppliers should we review first? Explain why.
What is the expected impact of changing EOQ?
Which EOQ changes would reduce buyer workload?
Which EOQ changes would reduce inventory investment?
Which EOQ changes would improve service level?
Which suppliers appear to have EOQs that are too high? Too low?
Which suppliers are placing more purchase orders than expected?
What if
What happens if we double ordering cost?
What happens if carrying cost increases?
What happens if we reduce EOQ by 20%? Increase it by 30%?
Which suppliers benefit most from increasing EOQ?
What-if answers are reasoned from your Recurrency data, not a full simulation. Confirm the items in Planning before you change a setting.
Order cycles
Order cycles
Current state
What order cycles are we currently using?
Which suppliers have unusually short order cycles? Unusually long?
Which order cycles appear inconsistent with purchasing history?
Recommendations
Which suppliers should increase order cycle? Decrease it? Why?
What impact would changing order cycles have?
Which order cycles generate unnecessary purchase orders?
Which suppliers should review supplier purchase targets?
Safety stock
Safety stock
Current state
Which items have the highest safety stock at [location]?
Which items have no safety stock?
Which A items have very little safety stock but stocked out in the last 12 months?
Which safety stock values concern you?
Recommendations
Which items should increase safety stock? Decrease it?
Which safety stock changes would reduce stockouts?
Which changes would reduce inventory investment?
How should we treat A items differently?
Min / max
Min / max
Current state
Only items on min/max (exclude OP/OQ) at [location]: which mins look too low vs last 90 days of demand? Top 20 by stockout risk.
Which items have mins that appear too high?
Which max values seem unreasonable?
Which items repeatedly hit max?
Which items repeatedly fall below min?
Recommendations
Which mins should we review first?
Which max values should change?
Which recommendations have the biggest financial impact?
Which recommendations reduce stockouts the most?
I’m reviewing item [SKU] at [location]. Show current min/max (or OP/OQ), recommended values, demand used, lead time used, and why they differ.
If you replenish on OP/OQ, say so in the first message. If you replenish on min/max, say not OP/OQ. Mixing those is the most common way a thread goes wrong.
Transfers
Transfers
Current state
Which branches perform the most transfers?
Which suppliers generate the most transfers?
Which items transfer most often?
How many transfers and what $ value moved from [hub] to other branches last month?
Which transfers move the fewest units? The least inventory value?
Operational improvement
Which transfers could be eliminated?
Which branches could purchase directly instead?
Which suppliers are good candidates to remove hub-and-spoke?
Where is [spoke] short while [hub] is long on the same item right now? Top 20 transfer candidates by service risk.
How much transfer activity could be eliminated?
Which suppliers would reduce operational workload the most?
Suppliers
Suppliers
Which suppliers are performing best at [location]?
Which suppliers are creating the most inventory problems?
Which suppliers have unreliable lead times?
Which suppliers frequently cause stockouts?
Which suppliers require the most manual intervention?
Which suppliers should we review first?
For supplier [supplier], compare expected vs actual lead time over the last 12 months. Which items are worst?
Stockouts
Stockouts
Which items stocked out the most this year at [location]?
Which vendors generate the most stockouts?
Which branches experience the most stockouts?
Which stockouts cost us the most sales?
Which stockouts appear preventable?
Why did [item] stock out at [location]?
Which stockouts were caused by low mins? Supplier lead times? Forecasting? Purchasing delays?
Show me items that have stocked out multiple times.
Which A items at [location] stocked out in the last 90 days? Rank by lost sales $. For the top 10, was the driver low min, lead time, forecast, or late PO?
Chat with your Data can analyze how Rules are behaving. It cannot deploy or edit a Rule — use Rule Builder for that.
Rules
Rules
Existing rules
Which validation rules trigger the most items the most often?
Which rules create the most inventory changes?
Which rules should we review keeping or changing?
Which rules appear unnecessary?
Which rules have the largest financial impact?
For rule [exact rule name]: how many items did it affect last month, at which locations, and what’s the net inventory $ impact?
Don’t explain every rule. Only explain rules that hit more than 50 items last month, in plain English, one paragraph each.
Recommendations (ideas only — you still request the Rule in Rule Builder)
What new validation rules would you recommend?
Which repetitive decisions should be automated?
Which manual changes happen frequently enough to become rules for [supplier / product group]?
Deadstock
Deadstock
Analysis
Which deadstock is growing fastest at [location]?
Which vendors create the most deadstock?
Which branches hold the most deadstock?
Which deadstock has the highest inventory value?
Which items are unlikely to sell?
Show items that became dead stock in the last 6 months (had sales before, none recently) at [location].
Recommendations
What should we liquidate?
Which items should become non-stock?
Which purchasing behaviors are creating deadstock?
If we had to free $500k of inventory at [location] without hurting A-item service, where should we start? Top 25 SKUs with reason codes.
Fill rate / service levels
Fill rate / service levels
Why isn’t our fill rate improving at [location]?
Which items have the lowest fill rates?
Which suppliers impact fill rate most?
What changes would improve fill rate?
What’s preventing [location] from reaching a 98% fill rate? Name the top 10 items dragging fill rate and the likely cause for each.
Use Root cause after you already have a symptom (too much inventory, too many stockouts, too many transfers). They work better once you’ve named the problem.
Root cause
Root cause
Why are we carrying so much inventory at [location]?
Why are we still experiencing stockouts?
Why are we transferring so much inventory?
Why are buyers reviewing so many items?
Why are forecasts changing so often?
Why do we repeatedly override recommendations?
Same caution as EOQ: What if is analysis on current data, not a sandbox that rewrites your plan. Keep what-ifs narrow.
What if
What if
What happens if we increase review cycles? Decrease them?
What happens if we reduce transfers by 25%?
What happens if we increase safety stock on A items at [location]?
What happens if we remove hub-and-spoke for supplier [supplier]?
What happens if we reduce lead times?
What happens if supplier [supplier] lead time increases by 7 days for [location]? Which A items are at risk in the next 60 days?
What happens if demand increases by 15%? Decreases by 20%?
Explain this recommendation is best from the page assistant on the item you’re looking at.
Explain this recommendation
Explain this recommendation
Why are you recommending this change?
What data did you use?
How confident are you?
What assumptions are you making?
What information is missing?
Show me the calculations behind this recommendation.
Why is your recommendation different from our current settings?
What evidence supports your recommendation?
Those results included OP/OQ items. Redo using min/max only and explain again.
Prioritization
Prioritization
If we only had time to make five changes this month at [location], what would you recommend?
Which recommendations deliver the biggest ROI?
Which recommendations reduce buyer workload the most?
Which recommendations reduce inventory investment the most?
Which recommendations reduce stockouts the most?
Which changes should we implement first?
Which recommendations carry the highest risk?
What can wait until next quarter?
Inventory consultant prompts work after you’ve already pulled a current-state cut. A consultant frame with no scope still leaves the search too wide.
Inventory consultant
Inventory consultant
Based on our last 12 months of data at [location], what are the three biggest mistakes we’re making?
If you joined our company tomorrow, what would you investigate first?
What operational habits are costing us the most money?
What purchasing behaviors concern you?
Where do you think we’re over-managing inventory?
Where are we trusting tribal knowledge instead of data?
What assumptions do you think our buyers are making that the data doesn’t support?
If you could automate three repetitive decisions for our buyers, what would they be?
What recommendations would have the biggest impact with the least effort?
Based on our data, what story is our inventory trying to tell us?
What question should I be asking that I haven’t asked yet?
FAQ
Q: Should I write “you are an expert warehouse manager” at the start?
A: Don’t spend the prompt on a persona. Recurrency already knows this is your inventory data. A short consultant frame is fine after you’ve asked for the numbers (“based on the last 12 months, what would you change first?”). An empty role-play with no scope is not.
Q: Is a longer prompt always better?
A: Longer only helps if you added constraints (location, min/max vs OP/OQ, time window, table vs summary). Two precise sentences beat a paragraph of background.
Q: Can I paste a whole category of prompts at once?
A: One question, read the answer, then the next. The assistant keeps context in the thread, so the follow-up can be short.
Q: Can I ask it to “just look at everything”?
A: The executive-summary prompts are the right first move. Then pick one problem (dead stock, EOQ, stockouts) and stay there. “Analyze everything” often times out or returns a generic answer.
Q: Why did my question time out?
A: Usually the prompt was too broad (all locations, all items, all rules, multiple topics). Narrow to a location, time window, and one topic — or ask for a top 10 — then follow up.
Q: Why did it answer a different question than I asked?
A: It guessed a missing constraint. Tell it what it got wrong and re-ask. If the thread is mixed up, New chat and paste a tighter version of the original question.
Q: Should I still use Planning and reports?
A: Yes. Chat with your Data is the fastest way to find and explain. Planning, Daily Workflow, Mission Control, and exports are still how you verify and act.