Delegate the work. Keep the judgment.
Agentic planning means Pinto does the work, not just the analysis. It runs your forecasting and supply planning on a schedule, checks its own work against the goals you set, explains every number, and brings you only the calls that need a person.
- Data refreshed · sales, inventory, promotions, catalog
- Gaps checked · 3 stores filled from last week’s counts
- Forecast done · 12,400 products, next six months
- Checked against your goals · 314 changes explained
- Supply plan drafted · 1,206 purchase orders, 58 transfers
- Note sent to Megan · 3 calls need a person
No one on the team was online for any of it.
Every cycle, the same five steps.
Pinto works on the schedule you choose, weekly or every two weeks. Each run follows the same loop, and each loop starts from everything the last one learned.
- 01 · Understand
It starts with context, not calculation.
Pinto pulls the latest sales, inventory, promotions, targets and product catalog, then reads your goals, your limits and what was decided last time. It checks the data for gaps, fills them where that’s safe and flags them where it isn’t.
Data refreshed · 2 years of sales · 14 promotions · 212 launches and discontinues noted · missing counts at 3 stores filled from last week’s snapshot - 02 · Work
It does the analysis a planner would.
Pinto corrects demand that was hidden by past stockouts, picks the right method for each product, forecasts new products from their attributes, and turns the final forecast into purchase orders, transfers and allocations.
12,400 products forecast · method chosen per product · 1,206 purchase orders and 58 transfers drafted - 03 · Check
It reviews its own work.
Pinto tests the result against your goals, limits and confidence rules, compares it with the last forecast, and writes down what changed and why.
Checked against all 3 goals · 314 products changed materially, each with a reason · 50 low-confidence items flagged - 04 · Ask
It brings you the calls it shouldn’t make alone.
Pinto sends a short note in plain English: what it decided, what it couldn’t do, and the handful of decisions that need you, each with options.
Rain shells selling ahead of forecast for two weeks. · Raise the next six weeks by 18%? · Raise it · Not yet · Show me why - 05 · Remember
It keeps what you teach it.
Your decisions, overrides and the reasons behind them carry into the next cycle, with a record of who changed what and why.
Saved to memory · Rain shells: treat two weeks of early sell-through as a trend, not noise · Megan
Tell Pinto what matters. It works out how.
Nobody wants a forecast that’s accurate in the abstract. You want the right products on the shelf without tying up more cash than you can afford. So instead of configuring models, you give Pinto three kinds of instruction.
What you’re aiming for
- “Protect availability on best sellers.”
- “Class A at 95%, B at 85%, C at 70%.”
- “Grow wholesale 12% this year.”
Lines Pinto never crosses
- “Lean no more than 5% high.”
- “Keep the quarter’s buying under $2.4M.”
How your business works
- “Don’t buy beyond the season; carry over eight weeks.”
- “When stock is short, protect the highest-margin channels.”
- “Never reorder discontinued items.”
Choose a model. Set the horizon. Tune safety stock. Run it. Check it. Repeat next cycle.
Set the goals once. Pinto picks the methods and settings that deliver them, and tells you when your goals pull against each other.
Pinto decides the routine. You decide the rest.
Pinto handles on its own
- Refreshing and checking data every cycle
- Filling small data gaps when it’s safe to
- Choosing the forecasting method for each product
- Correcting demand distorted by stockouts
- Drafting purchase orders and transfers within your limits
- Writing up what changed and why
Pinto always brings to you
- Anything that would cross a limit you’ve set
- Goals that pull against each other
- Forecasts it has low confidence in
- Big swings from the last forecast with no clear cause
- Data too incomplete to forecast safely
- Anything your policies mark for approval
Pinto proposes. People approve. Nothing that needs a decision changes until someone says yes.
Three ways to plan. Only one does the work for you.
| Spreadsheets | Traditional planning software | Pinto | |
|---|---|---|---|
| Who runs the forecast | A planner, every cycle | A planner configures and runs it | Pinto, on your schedule |
| How you steer it | Formulas and tabs | Model settings and parameters | Goals, limits and policies |
| Why a number changed | Only if someone wrote it down | Charts, rarely the reason | A plain-English reason for every number |
| What reaches you | Whatever you find | Lots of alerts | Only the calls that need judgment, with options |
| Meeting decisions | Re-keyed by hand | Re-keyed by hand | Read from the notes, proposed for approval |
| What it remembers | Nothing | Overrides, without the reason | Decisions and reasons, carried forward |
| Forecast to supply plan | Copy and paste | Separate modules | One agent, one shared picture |
Hand Pinto the meeting notes.
Most planning decisions are made in the weekly trade meeting, then re-keyed into a system days later from memory. Pinto reads the notes, works out what was decided, and comes back with a short list to approve. Each item shows the words it came from.
Marketing: winter campaign moves two weeks earlier.
Finance: cash is tight. Cap Q4 buying at $2.4M.
Tighten everything except the best sellers.
Lime helmets, keep or drop? Sam and Alex disagree. Parked.
Next meeting: Thursday, same time.
Move the winter build forward 14 days. 340 products.
Cap Q4 buying at $2.4M. Passed to supply planning.
You said to tighten everything except the best sellers. Which groups, and by how much?
Lime helmets. Two of you disagreed and it was parked. They’re still in the forecast, carrying about $28K of write-off risk.
Six months later, you’ll still know why.
Pinto keeps a plain-English record of what it chose and what it chose against. It builds trust today and settles arguments six months from now.
It also records what it expected to happen, so you can check later what did. Over time that becomes a simple track record: how often Pinto was right, and how much further to trust it.
- Forecast
- 1,840 units, Oct–Dec · up 22% on last year
- Method
- Seasonal model, chosen over two alternatives. Best on the last 90 days and the only one that kept last winter’s peak.
- Lean
- +4% under the best-seller rule (limit 5%)
- Why it moved
- Winter campaign moved two weeks earlier. From Thursday’s trade meeting:
winter campaign moves two weeks earlier
- Who changed it
- Megan raised the lean limit for the winter range · Mon 08:04
Teach it once. It applies it every cycle.
An override or a meeting decision shouldn’t vanish after the cycle. When you approve a change, you say why, and Pinto keeps it for next time.
Right-sized
A memory can apply to one product, a group, or become a standing rule.
Nothing hidden
Every memory shows what it says, where it applies, who added it and when.
Yours to edit
Review, edit or retire any memory. Nothing becomes a standing rule without approval.
Accepted low confidence on new trail shoes. Launch data is thin. Applies to 12 products · Megan
Standing rule: carry over eight weeks after season end. All seasonal products · approved by Dana, supply lead
Ask a question. Get the analysis, not just an answer.
Pinto answers questions about any product, supplier or number. When a question needs analysis, it writes the code, runs it on your data and shows its work, so you can check how it got there.
One shared picture of your business.
OnePint connects the systems you already run and keeps one live, trusted view of demand, inventory, supply and business context. Pinto’s five areas of expertise all work from that same picture, so a decision in one shows up in the others.
- ERP
- WMS
- OMS
- Ecommerce & marketplaces
- Spreadsheets & files
- Inventory
- Demand
- Supply
- Goals & policies
- Decision history
- Demand Forecasting
- Supply Planning
- Inventory Intelligence & ATP
- Risk & Alerts
- Simulation & Optimization
- Goals & limits
- Approvals
- Judgment calls
OnePint works alongside the systems you already run.
Autonomy you can audit.
Approvals where it matters
Decisions go to the right person, and nothing changes until they say yes.
A full record
Every decision, its reason, every change and every approver is kept.
Access control
People see and approve only what their role allows.
Five areas of expertise. One agent.
Each one has its place in the cycle.
Does Pinto replace planners?
No. Pinto takes the repetitive work: gathering data, running forecasts, drafting plans and writing up what changed. Planners set the goals, make the judgment calls and approve decisions.
Who is OnePint built for?
Mid-sized retailers and consumer brands, typically $50M to $1B in revenue, that sell finished goods. Most are planning in spreadsheets or basic tools today and have outgrown them.
Does Pinto plan manufacturing?
No. Pinto plans finished goods, from purchase order to delivery: what to buy or make, how much, when, and where it goes. It doesn’t schedule production lines or plan raw materials.
What data does Pinto need?
Sales history by channel (ideally two years or more), inventory snapshots, your product catalog with launch and discontinue dates, and your promotions calendar. Supply planning adds lead times, open purchase orders and supplier minimums.
How often does Pinto run?
On the schedule you set, typically weekly or every two weeks, and on demand when something changes.
Can we see why Pinto made a decision?
Yes. Every forecast and order comes with a plain-English reason, and the record shows what changed, who changed it and why.
What happens when Pinto isn’t sure?
It says so. Low-confidence forecasts, conflicting goals and anything outside your limits come to a person with options, instead of being decided quietly.
Bring us a real inventory problem.
A forecast nobody trusts. A buy plan that takes a week. Stockouts you only find after the fact. Tell us what’s slowing your team down, and we’ll show you how Pinto would handle it.