Measured results, on systems customers already run

The published story

From Excel exports to decisions in production

Databricks has published a customer story about an A2go Decision Intelligence Platform (ADIP) deployment at an operation with 65 production facilities, 17 sales channels, and more than 150,000 customers. The before-state is one most planning organizations will recognize: analysts pulled data into Excel and circulated static recommendations, and by the time a recommendation reached the person who could act on it, the data behind it was stale.

What the deployment changed, as reported in the story:

• Forecasts that took 28 hours to produce now complete in under 1 hour.
• Pricing is simulated across 60,000+ SKUs in minutes rather than days.
• More than 25 market-intelligence users run scenarios daily without data-engineering support.

The full account, including how the deployment runs inside the customer's Databricks environment, is published here: read the Databricks customer story.

How we count

What these figures are, and what they are not

The figures on this page are reported as received from the engagement. We publish the measurement in the customer's own terms, against their own baseline. We do not average results across customers, extrapolate a pilot into an annual projection, or restate a number in a stronger unit than the one it was measured in.

Results vary by engagement. What a deployment changes depends on the decision you point it at, the condition of the systems that feed it, and the cadence your team already runs. A 28-hour forecast cycle and an 18-hour scheduling run were the measured starting points at those operations; yours will differ, and so will the ending point.

If you want proof at the level of mechanism rather than outcome, we have written up two decisions end to end, showing what the platform reads, what it proposes, and where a person stays in the loop: the ATP short-ship decision and the clear-to-build decision.