Promotion Analyst: what this AI agent does and how to use it

Try the Promotion Analyst live demo →

What it does

Price ladders and promotions are easy to get wrong. These agents show where prices cluster, where gaps are, and how promotions performed. The Promotion Analyst agent is the specialist for this job: evaluates promotions and markdown candidates from sales velocity and stock.

Who it's for

Your team. It runs in the console for your staff and through the REST API. It is a team agent, so it is never exposed to shoppers.

Questions it answers

Ask it in any of our languages, for example:

• "Which promotions are running and how did their categories sell?"
• "How did the Party season dresses promotion perform?"
• "Which category should we promote next?"

The live demo shows these as sample prompts in your language, on a fictional store's data.

How it works

1. It understands the question in the language it was written in, even when your catalog is in another language.
2. It looks things up with tools, never from memory:

• Read-only sql over your retail data (sandboxed per workspace)
• Sales analytics by category, channel or period

3. It answers from your data, in your brand voice and currency, and shows which tools it used.
4. Rules stay in code. Prices, eligibility and permissions come from your data and settings; the agent explains them and can't invent them. Why this matters.

The data it needs

Products, Orders, Order lines, Promotions. As soon as these tables are loaded, the agent is ready to switch on. Files can use your own column names: they're mapped automatically. See Connect your data.

Works well with

• Price Architecture Analyst: Explains price ladders and positioning by category, brand and department.
• Markdown Advisor: Finds slow sellers with too much stock and suggests markdown depth and timing.
• Price Sensitivity Explainer: Shows how sales respond across price points and during promotions, in plain language.
• Key Value Item Identifier: Identifies the items that shape price perception, from purchase frequency and reach across customers.
• Promotion Cannibalisation Checker: Checks whether a promotion's uplift came at the expense of other products or later sales.

Your team can use them side by side in the console, or call each one through the API.

Turn it on

1. Load the data it needs.
2. Install Promotion Analyst from the Agent catalog and add your own instructions (guide).
3. Test it in the Playground, in every language you serve.
4. Invite your team or call it from your own systems.

New to ServiceBuddy? Start with how onboarding works.