Merchant Analytics: Intelo.ai Agent Team

Merchant Analytics is the Intelo Agent Team that gives every other team the truth. Demand forecasts. Size curves. Assortment performance. Newness trajectory. Lost sales. All continuous, all reconciled, all operational.

What is agentic merchant analytics?

Agentic merchant analytics is a continuously-running AI system that produces true-demand forecasts, size curves, assortment diagnostics, newness predictions, and lost-sales recovery at SKU-store-week granularity, and feeds those signals to every other merchandising decision in real time. Intelo's Merchant Analytics replaces weekly reporting cycles with always-on analysis that the rest of the stack learns from.

Agents in this team

• Demand Forecasting: Produces SKU-store-week demand forecasts continuously, not quarterly. Blends historicals, seasonality, pricing, weather, and macro signals. Reconciles automatically when the plan changes upstream or execution drifts downstream.
• Assortment Performance: Diagnoses why an assortment is winning or losing: by style, class, cluster, and channel. Identifies heroes, cannibalizers, and deadweight in near real time so buyers can reallocate attention, not just inventory.
• Size Optimization: Builds size curves that respect cluster-level fit behavior, not network averages. Catches oversold sizes before they become lost sales and undersold sizes before they become markdowns.
• Newness Forecasting: Predicts performance for new products with limited or no history using similar-item modeling, attribute embeddings, and early-read signals. Gives buyers conviction faster on what to reorder, what to kill, and what to scale.
• Lost Sales: Quantifies demand that never showed up in your POS: stockouts, size breaks, off-planogram events, and substitution losses. Converts invisible revenue into visible, actionable recovery.

Frequently asked questions

What is 'true demand' and how is it different from sales?

Sales records only what you had the inventory to sell. True demand is what customers wanted: it reconstructs lost sales from stockouts, size breaks, substitution, and off-planogram events. Intelo's demand model estimates true demand at SKU-store-week granularity so downstream decisions see what customers actually wanted, not what you happened to sell.

How does the demand model handle stockouts, cannibalization, and halo effects?

Stockouts are backfilled using substitution patterns and historical sell-through rates. Cannibalization is modeled across comparable SKUs: when one item sells, how much does it subtract from similar items? Halo is the inverse: how much does a hero item lift adjacent SKUs? Both effects are reconciled per cluster so the forecast reflects actual customer behavior, not accounting assumptions.

What data sources does Merchant Analytics ingest?

POS, inventory snapshots, PLM, ERP, planning platforms, PIM, weather, and macro signals. Intelo coexists with Oracle Retail, SAP, Microsoft, Databricks, Snowflake, and legacy spreadsheets. The onboarding team maps your sources during the first-quarter deployment.

Can we use Merchant Analytics with our existing forecasting tool?

Yes. Many customers run Intelo's forecasts in parallel to an existing tool for a season before cutting over. The agents export forecasts in standard formats so your existing downstream systems (allocation, MFP, pricing) can consume them without rework.

What granularity does it forecast at?

SKU-store-week by default. The agents also produce roll-ups at cluster, department, or hierarchy level on demand. The native grain stays at SKU-store-week so recommendations are never averaged away before they reach a decision.

Does Merchant Analytics replace our data team or augment it?

Augment. The agents replace the mechanical work: forecasting, variance analysis, lost-sales calculation, size curve maintenance. Data scientists shift to higher-leverage work: validating edge cases, tuning the models to the brand's assortment logic, and building the next generation of signals.