Less than you'd think, but more is always better. The simple rule: the more sales history you upload, the more reliable the forecast.

We recommend at least 2 years of sales history per SKU. That gives the AI enough signal to capture seasonality, trend shifts, and demand volatility. With 2 years, your model is reliable from day one.

If you have less than 2 years, Flowlity Lite still works. The model just starts with a wider confidence interval (more uncertainty up front) and tightens every month as new sales come in. Many of our customers started with 12 to 18 months of history and saw forecast accuracy improve quickly.

For brand-new products with no history at all, the engine uses patterns from similar products in your catalog to build a starting forecast. It is not magic: the prediction starts wider and sharpens as actual sales come in.

Practical advice: don't wait until your data is "perfect" to start. Upload what you have today. The model will tell you which SKUs it is confident on and which need more time.

No. Flowlity Lite is plug-and-play and works without any integration.

To get started, you upload your sales history as a CSV or Excel file. That is it. There is no API to configure, no connector to install, no IT team to involve. If you can export data from your ERP, your warehouse management system, or even a Google Sheet, you have everything you need.

This is intentional. Most SMBs we work with don't have the bandwidth, or the budget, to run a 6-month integration project. By keeping the data flow simple (upload, refresh weekly or monthly), we get you to value within a day instead of a quarter.

If you later want to automate the data sync with your ERP, you can connect it through a standard CSV export pushed on a schedule. And if your business grows beyond what Lite covers, you can move to the main Flowlity product, which does include native ERP integration, without losing your setup or your team's habits.

ChatGPT, Claude, and Copilot are general-purpose language models, built to read and write text. Asking them to forecast your demand is like asking a brilliant writer to design your warehouse layout: confident-sounding answers, but not what they were built for.

Flowlity is a probabilistic engine trained on real supply chain data. It learns from your sales history (your products, customers, suppliers, seasonality) and recalibrates every time you refresh your data. The output is a probability distribution per SKU: how confident the forecast is, how much safety stock you actually need, when to reorder. The same math that powers planning at enterprises like Saint-Gobain and Magotteaux, sized for SMBs.

And if you already work daily in ChatGPT or Claude, you don't have to choose. Flowlity Co-Planner is our MCP connector that lets you query your Flowlity data directly from your AI assistant. So "what should I reorder this week?" in Claude returns Flowlity's actual forecast, securely.

In short: ChatGPT for your emails. Flowlity for your decisions. Flowlity Co-Planner to bridge the two.

Faster than you'd expect. The setup is built around a single onboarding call with a member of our team.

During that call, we walk through your data file together. An Excel export from your ERP, or even a Google Sheet of your sales history, is enough. There is no IT ticket to file, no integration to schedule.

Once your data is in, our AI runs your first demand forecast within hours. You can review it the same day. Most customers start acting on the recommendations within the first 2 weeks.

Full operational use, meaning your team uses Lite in their weekly planning rhythm, typically takes one month. That includes time for your team to get comfortable with the platform and adjust their workflow.

For context: one of our SMB customers, a €17M beauty brand, was running their entire Supply Chain on Flowlity Lite within four weeks of signing. Their words: "Honestly, it was very easy."

Yes. Flowlity Lite is a monthly subscription with no long-term commitment. You can cancel anytime from your account, effective at the end of the current billing cycle.

This is intentional. Most planning software contracts run 2 to 3 years because the cost of integration is so high that vendors need to amortize it. Lite is different: there is no integration project, so there is no reason to lock you in.

You also get a 14-day free trial before any payment, so you can test the AI forecast on your own data before committing to a single euro.

If you decide Lite isn't the right fit, you export your data and walk away. No clawback, no penalty, no exit interview.

Flowlity Lite runs on the same probabilistic AI engine as our main product. The difference is in how it is delivered.

The main Flowlity product is built for mid-market and enterprise companies. It connects directly to your ERP, supports complex multi-site configurations, includes advanced features like price and promotion optimization, real-time S&OP, and supplier collaboration, and comes with a dedicated Customer Success Manager and a full integration project.

Flowlity Lite is designed for small businesses that need real AI planning but don't have the budget, or the bandwidth, for a 6-month integration project. You upload your sales history as a CSV, get an AI forecast per SKU, and start acting on it within days. No ERP integration. From €249 per month.

The forecasting accuracy is the same. The depth of features differs. Lite covers what most SMBs actually need: demand forecasting, supply planning, and order management for up to 5,000 SKUs.

If your business grows beyond what Lite covers, you can move to the main product without losing your setup or your data.