• Pre-read assembly, 48 hours before the meeting
• Live decision capture, with owner and date
• Exception detection against thresholds you set
• Assumption ageing: it tells you which comparator is six months old
• Acts on what it finds, day or night: drafts the change, updates the pack, tells the owner. You approve in the morning.
• Follow-up tracking across cycles
• A searchable memory of every cycle it has attended

• No admin approval. It's an Office add-in, installed from the ribbon.
• No data leaves your tenancy unless you send it.
• Free to try on one cycle.
• Reads the consensus forecast, the exceptions tab and the notes column. Writes the executive summary, the exception commentary and the slides.
• Every sentence it writes cites the cell range it came from.

• Draft-and-approve by default. You see everything before it leaves.
• Per-channel and per-stakeholder permissions. Email to the plant, yes. Calls to finance, not yet.
• A full log of every message and call, with transcripts.
• Every message says it is from Pal, on your behalf. It never pretends to be you.

Inside your business

• Consensus forecast
• Open orders
• Inventory positions and stock-outs
• Production capacity and downtime
• Master data changes
• Promotional calendar
• Campaign plans
• Phasing and bias history
• The last six months of S&OP decisions

Outside your business

• Competitor pricing and promotions
• Retailer listings and range changes
• News and events by geography
• Weather against your seasonal SKUs
• Commodity and freight movement
• Category demand indicators

• Read-only by default on every connected system. Writes require a named human approval.
• BYO cloud tenancy, BYO model, or fully hosted. Three deployment options, priced the same.
• Data residency options: UK, EU, US.
• Meeting participation is visible: Pal appears in the participant list, always. It never records a call it is not visibly in.
• Retention and deletion controls per business unit. Export everything on exit.
• Pal never trains on your data.

We would rather tell you what is in progress than imply it is complete.

Drafts. Publication dates are targets, not history. Bylines marked pending have not yet been approved by their authors.

Named with permission.

• AI first. If a model can do the work, the model does the work and a person checks it.
• Small teams. Fewer people, more scope each.
• Revenue before anything. Design partners pay on outcome; we build what they will pay for.
• We run the company on Teams, because our customers do.

Demand arrives as a rolling release from the OEM: firm for two weeks, planned for twelve, forecast for a year, and revised every week. The supply plan is built on the release, the release is built on the OEM's own forecast, and the difference between the two is where the stock goes. Semiconductor and sub-tier constraints surface in a call from the supplier, not in the plan. Engineering changes arrive by email with an effectivity date that somebody has to translate into a run-out.

• Reads the OEM release and the call-offs from EDI, the supplier constraint from the email, the engineering change from the notice, and shows the three against each other.
• Flags the part numbers where the call-off is running ahead of the release, or the release has dropped and the supply plan has not.
• Captures what was agreed in the daily and weekly supply calls: who owns the recovery, when the line rate returns, which customer takes the shortage.
• Chases the supplier for the confirmed delivery date and the plant for the changeover, in your name, and logs the answer against the part.
• Writes the weekly supply review and the monthly S&OP pack in your template, with the source for every movement.

Reads from: SAP, Kinaxis, o9, EDI (DELFOR, DELJIT), supplier portals, the plant's Teams channel, engineering change notices.

Tuesday 06:50. A tier-two supplier emails that a connector will be short for three weeks. By 07:30 Pal has listed the eleven finished-goods part numbers that use it, the OEM call-offs they feed, the days of cover on each, and a draft note to the OEM planner for the two that will miss. The supply planner approves the note at 08:05.

The monthly cycle starts on the first working day. The demand planner updates last month's sales; the sales manager has two days to revise with the regions; on day three the country manager signs the forecast and owns it. From day five the replenishment plan goes to manufacturing with a six-to-twelve-month order horizon, and the supply side works the constraints with internal sites and CMOs. When a market is short, the fix is a transfer from another market, which needs regulatory approval, batch documentation and a paper trail that will be audited. Shelf life and cold chain make over-forecasting expensive in a way that shows up as write-offs a year later.

• Prepares the day-three demand review pre-read: the SKU-level forecast against last year, the run rate and the campaign calendar, with the sales justification captured from the thread rather than retyped.
• Records the country manager's sign-off and the assumptions it rested on, so the September revision can be explained in March.
• Ages every assumption: the competitor stock-out that justified the uplift in June is flagged when it is still in the plan in October.
• Tracks tender and hospital contract dates against the plan, and flags the SKUs where a tender loss has not yet come out of the forecast.
• Drafts the transfer request between markets with the batch, expiry and regulatory fields filled from the source documents, for a named person to submit.

Reads from: SAP, Kinaxis, o9, Anaplan, the tender calendar, the campaign calendar, the QA release system, CMO portals, Teams and Outlook.

Day three, 09:10. The Portugal sales manager has revised the forecast up 14 per cent on one SKU, citing a competitor stock-out. Pal posts to the review: the last two times this justification was used the uplift delivered 6 and 4 per cent; the competitor's stock-out was reported resolved on 22 August; current cover is 11 weeks against a 9-month shelf life. The country manager holds the forecast at plus 5 and signs. The reasoning is in the log.

A third of the portfolio is innovation or renovation. Each launch carries a phasing assumption, a distribution build and a cannibalisation guess, all agreed in a meeting and rarely revisited once the product is in market. Retailer listings change monthly. A single creator post can double a SKU's sell-through in a week, and the sell-in that follows is what gets counted in the plan. Sizes, shades and formats multiply SKUs faster than planners can review them.

• Tracks every launch against its own assumptions: phasing, distribution, cannibalisation, and the sell-through the assumption implied.
• Reads the retailer range files, listings and delistings, and reconciles them with the distribution build in the plan.
• Watches the category signal: search, creator activity and competitor launches, and reports which SKUs moved with them.
• Separates sell-in from sell-through where the retailer data allows, and flags the SKUs where the plan is chasing sell-in that has not sold through.
• Writes the demand review pack by brand and by launch, with the source for every uplift.

Reads from: SAP, Kinaxis, o9, Anaplan, retailer EPOS and forecast files, the launch calendar, the promo calendar, PIM and master data, Teams and Outlook.

Monday 10:14. A brand manager asks why the spring launch forecast has not moved after a viral week. Pal answers: sell-through at the two retailers sharing EPOS is up 210 per cent over seven days; the plan's phasing assumption from January had week 12 as the peak; the launch has 3.4 weeks of cover in the retailer DCs. It proposes a revised phasing for review and drafts the note to the demand planner.

The consensus forecast carries a promotion calendar that the retailer changes monthly, an uplift assumption per promotion that nobody re-measures, and a seasonal curve that was built on last year's weather. Retailers send their own forecasts, on their own cadence, in their own formats; EPOS arrives weekly for some customers and never for others. Peak-season dispatch rates fall to 85 per cent and the lost volume is booked as forecast error. Forecast accuracy in the low 60s and a positive bias of four to five per cent are common at large businesses, and the bias goes straight into write-offs on short-life products.

• Assembles the pre-read from the plan, the promo calendar, the retailer forecasts and last month's decisions, exceptions first.
• Measures every promotion against its assumption: this uplift said 18 per cent, the last three on this range delivered 11.
• Reads weather against the seasonal SKUs and reports the comparator being used and whether it is still reasonable.
• Reconciles retailer forecasts and EPOS with the consensus, by customer, and flags the gaps above threshold.
• Captures the demand review and the supply review as they are spoken, and chases the commercial inputs that are owed.

Reads from: SAP, Kinaxis, o9, Anaplan, Blue Yonder, retailer forecast and EPOS files, the promo calendar, the campaign plan, weather feeds, Teams and Outlook.

Tuesday 09:14, 48 hours before the S&OP. Pal posts the pre-read: consensus up 3.1 per cent; four exceptions above threshold including a bias drift on one range; two open assumptions carried from last month, one six months old; three decisions from last month not yet in the plan. The commercial view on the promotion is the only field left blank, and the account manager who owes it has been asked.

Demand comes from projects that are won months ahead and start late, from merchants who stock up before a price increase and go quiet after it, and from a season that a wet April can shorten by three weeks. Capacity is regional and heavy to move: a kiln, a line, a quarry. Planning runs on the order book plus judgment, with the judgment held by a handful of regional sales managers whose forecasts are optimistic by design. Freight and energy costs move the economics of where to make and where to ship faster than the plan is revised.

• Reads the project pipeline from the CRM, the order book from the ERP and the merchant stock reports, and reports the gap between what is won and what is called off.
• Flags pre-buying ahead of a price change and separates it from underlying demand in the consensus.
• Watches weather and seasonal indicators by region against the seasonal SKUs, and tells you which region is running behind its curve.
• Tracks capacity constraints and maintenance windows raised in plant calls against the supply plan, and chases the plant for confirmed dates.
• Writes the regional S&OP pack and the executive summary, with the source for every movement.

Reads from: SAP, Kinaxis, o9, Anaplan, the CRM pipeline, the order book, merchant stock reports, energy and freight indices, weather feeds, Teams and Outlook.

Thursday 14:00. A regional sales manager forecasts a strong Q2 on the strength of four large projects. Pal posts: two of the four have slipped start dates in the CRM since the forecast was made; merchant stock in the region is 30 per cent above last year after the March price letter; the ten-day forecast is wet. It proposes holding the regional plan flat and moving the review to the next cycle, and logs the four projects with owners.