A warehouse shows a surplus. Across the region another warehouse expects a shortage. Moving stock seems obvious until sales points out that the first site has promised part of its inventory to a customer, finance prices the rush shipment, and a planner notices that the forecast was calculated before a promotion changed. Aily Labs builds decision software for questions that cross those boundaries. The useful test is whether the buyer can explain and act on a better allocation with the information available that day.
- 01The software Aily joins enterprise indicators, forecasts and modelled alternatives in a decision application.
- 02The reference Sanofi publicly describes developing its cross-functional plai application with Aily Labs.
- 03The buying question Decide which recommendation needs to improve, who owns the source data and who may authorize a change in the operating system.
01 / ProductFrom a department’s question to an enterprise decision
Aily Labs describes its decision intelligence platform as a way to connect information spread across different functions. A single department might use a forecast to anticipate demand. Linking that forecast to procurement and finance changes what a user can ask. The obvious choice by sales volume might tie up scarce material or increase transport costs. The answer needs a unit of comparison that the people responsible for all three outcomes recognize.
The solution suite distinguishes several levels of scope. Aily.App is described as a departmental offering. Aily.Pro uses the company’s Correlator technology to connect indicators across modules. ProX adds scenario-based advice and potential automation. Super Agent coordinates specialist agents and is described as able to initiate actions across systems. These descriptions establish product roles. They do not establish which module a particular customer has licensed, which data has been connected or which actions a user can authorize.
There is a documented customer example. Sanofi’s account of plai names Aily as a development partner and describes bringing internal information from several functions together to support decisions and personalized scenarios. Aily’s case study includes supply and manufacturing use. That establishes that a large enterprise has taken this route. Neither company’s account provides a controlled forecast of the savings another retailer or manufacturer will realize from an inventory transfer.
02 / AudienceWho owns a decision that crosses three departments?
A sensible sponsor is a business owner who can describe a repeating choice, the deadline and what happens if nobody acts. An allocation planner may move goods between distribution centres every Tuesday after a demand update. The finance controller can describe the costs that should enter that decision. A sales manager knows which units are already promised. Having all three available for a pilot matters more than giving everyone access to a polished dashboard.
A team still arguing about the definition of available stock has a different starting point. Stock on hand might include damaged goods, goods reserved for an order, expected returns or goods already staged for dispatch. Aily may help surface those differences, but the software cannot decide which definition the business wants to use in a customer commitment. Give each field a named owner who can explain the rule and how often the source is updated.
A manager who only needs a weekly historical stock total may already have a perfectly serviceable report. Cross-functional modelling becomes more interesting when plausible actions have competing consequences. Write those actions down before the sales demonstration. If nobody can name a different action they would take after seeing the forecast, the organization may be asking for a better view of the business rather than a new way to run it.
03 / WorkflowReplay a stock decision with the facts available at the time
For a proposed evaluation, choose a completed trading week and a product with a genuine allocation choice. Preserve the forecast that existed when the choice was made, the committed customer orders, the available stock at each site and the transport lead time. Remove later sales and corrections from the input. They belong in the evaluation record. Using the final corrected forecast to recreate an older decision would give the pilot an advantage nobody had in practice.
Ask for three alternatives. The team might leave inventory in place, move a smaller quantity through normal freight or transfer enough goods to cover the whole projected gap using an expedited service. Have the planner write down the practical constraints before the recommendations appear. A truck departure, a minimum shipment quantity and an existing customer reservation can each make an option unavailable, even when demand and margin arithmetic suggest it would perform well.
Then inspect the explanation. Which stock position and demand signal drove the recommendation? What changes if the promotion reaches fewer customers, the truck misses its departure or the source warehouse receives a large order? A commercial manager should be able to challenge the assumptions in ordinary business terms. A supply planner should be able to identify the operational record that would change if the proposal were approved. Document disagreements while both people can inspect the same case.
Keep the pilot advisory until the team understands the action trail. Record what the system proposed, what the planner authorized and what the warehouse actually shipped. Those events can differ for perfectly good reasons. Someone may spot a customer promise that reached the system late. Recording that reason tells the data owner what needs improving. Treat a recommendation that was overridden for a valid reason as evidence about the process, even when a subsequent sales figure makes the original proposal look profitable.
04 / PricingWhat the enterprise offer establishes about cost
| Route | What Aily describes | What to obtain |
|---|---|---|
| Direct enterprise purchase | Package and pricing discussion with a specialist | Written scope, term, modules and implementation obligations. |
| Departmental evaluation | Single-module app described in the suite | Price and eligibility for the intended evaluation with named data. |
| Google Cloud procurement | Purchase through eligible existing cloud arrangements | Private offer, applicable commitments and included services. |
Commercial routes and questions to resolve, consulted 19 November 2026. Aily pricing enquiries and Google Cloud Marketplace announcement.
Aily’s contact page explicitly invites package and pricing enquiries. The suite FAQ identifies selected modules, user volume and deployment scope as pricing factors. The pages reviewed do not publish a universal per-seat or per-decision price. To get comparable proposals, start with one product family, the relevant locations and the historical period needed for replay. Specify whether the team wants a read-only recommendation, a planner approval screen or an actual change in a system of record. Those are meaningfully different deliverables even when the same people use them.
The Google Cloud announcement explains another way to procure the platform and describes applying eligible spend towards existing commitments. A procurement channel changes how a company can buy. It does not tell a buyer what this workload costs to run. Ask which amount counts towards the commitment, what the seller will invoice separately and who maintains the connections after a source system changes. Include the customer’s own time spent resolving disagreements about definitions.
05 / DistinctionsA planning comparison that starts with the consequence
Pigment’s planning approach gives finance and operating teams a useful comparator when the objective is to build a shared business plan. Ask prospective providers to show who can change a forecast assumption, who sees its consequences and where an approved change appears. A team that already maintains a trusted planning model may prefer to preserve it and add recommendations only at a specific point in the decision cycle.
o9 Solutions’ supply-chain planning work offers a comparison closer to physical constraints. For the same transfer case, include the truck deadline and the reserved units from the beginning. If one demonstration ignores those limitations, a more attractive projected margin is weak evidence of a better allocation. Compare the explanations, the buyer’s ability to correct bad inputs and the time a planner has to make the decision.
06 / QuestionsWhat happens when yesterday’s data was wrong?
Suppose a store posts a large return after a recommended shipment has already left. The old decision may have been reasonable with the information then available. Preserve the snapshot the system used and decide who can issue a revised instruction. Without that history a monthly review risks blaming the planner for a later correction, or crediting the recommendation for a favourable event the model could not have known.
Agree how overrides are interpreted before measuring success. An experienced planner who rejects a recommendation because there is no loading slot might be revealing a missing data connection. Another planner might reject a sound recommendation to protect their own local service target. Those cases call for different management responses. Review a small set of actual disagreements with finance, logistics and sales in the same room before making acceptance rate a performance target.
07 / DecisionWhere to start and how much authority to grant
A team with one difficult planning decision
Start with a bounded replay and the inputs that were truly available. Compare each recommendation with the practical alternatives the planner could have taken.
Departments with competing definitions
Reconcile reservations, stock availability and timing. Ask the people who own the relevant source systems to explain two difficult real examples.
An enterprise considering automated transfers
Ask for approval boundaries, change records, failure handling and correction rights on the exact receiving system. Rehearse an override on a reversible case.
A business worth understanding.
Suggest your business or one you find interesting. Tell us what you want to understand about its product, positioning, design or workflows.
Suggestions are free. Selection and publication stay with the desk.
- Aily Labs enterprise decision intelligenceConsulted
- Aily App, Pro, ProX and Super AgentConsulted
- Aily’s account of its Sanofi deploymentConsulted
- Sanofi on developing and deploying plai with Aily LabsConsulted
- Aily Labs enterprise demo and pricing enquiriesConsulted
- Aily’s Google Cloud Marketplace procurement announcementConsulted

