Operations
Questions in plain language.
Aurélien
The retailer
A 9-store streetwear retailer, two buyers, three thousand five hundred SKUs.
Our retailer operates 9 stores across France in the streetwear contemporary unisex segment. The team is small and focused: a 2-person buying team running roughly 3,500 SKUs in season.
The stack is mainstream for the sector — Polaris as core retail management, Shopify for e-commerce, and Excel as the connective tissue between everything else. It works. Until it doesn't.
"Every question I had took two emails and three days to get answered. By the time the answer came, the question had moved on."
— Chief Merchant
Context
A chief merchant separated from her data by a wall of analysts.
2024. The chief merchant of the streetwear retailer ran the buying strategy, the brand calendar, and the supplier relationships. Every operational decision required data — sell-through patterns, stock cover, store performance — and every data request went through her two-person analyst team.
The analysts were good. They were also one analyst short, and overwhelmed. So the chief merchant's questions queued. By the time an answer came back, three more questions had emerged.
She had the dashboards. She could even read them. But the moment she had a specific question — "why are weekend sales dropping in store 4?" or "which capsule items should we re-run for SS?" — she needed an analyst.
The Problematic
The data was right there. Just not for her.
The chief merchant could see the dashboards. She could even read them. But the moment she had a specific question — "why are weekend sales dropping in store 4?" or "which capsule items should we re-run for SS?" — she needed an analyst.
The analyst team had thirty other requests. The chief merchant's questions waited 1-3 days for answers, then often required a follow-up. The decision cycle stretched. Mondays were spent emailing analysts, not making decisions.
Three friction points compounded:
Specific questions required analyst capacity that didn't exist.
Answer turnaround of 1-3 days made tactical decisions slow.
Follow-up questions added another cycle, multiplying the delay.
The chief merchant had the data. She just didn't have the access.
The Solution
Plain-language questions, structured answers, sources shown.
Solya pulled the canonical retail model, the team's metric definitions, and the encoded business rules. The chief merchant's questions ran straight against the model — no analyst in the loop, no SQL.
She typed her questions naturally: "Why are weekend sales dropping in store 4?" or "Which capsule items should we re-run for SS?" The agent pulled the relevant data, ran the analysis, and returned a structured answer — with the calculation visible, the sources shown, and the confidence level attached.
80% of questions answered in seconds, with full transparency.
Edge cases still routed to the analyst team — but with the partial work already done.
Every answer traced back to the data, the metric definition, and the rule applied.
The analyst queue dropped from thirty requests to six. Everyone wins.
How we did it
Inside the loop.
The analytics agent runs on the Application Layer, with the metrics layer providing definitions and the data layer providing the substrate. Here's how the system works, end to end.
01 — Plug into the canonical model.
The agent had direct access to Solya's canonical retail model — every SKU, every store, every transaction — with native retail entities and the team's metric definitions already encoded.
02 — Parse plain-language questions.
The chief merchant typed questions naturally. The agent translated them into queries against the canonical model — no SQL, no analyst middleman.
03 — Run the analysis.
The agent computed the answer using the team's encoded definitions and business rules. No ad-hoc math, no shortcut SQL.
04 — Show sources and reasoning.
Every answer came with the data sources, the calculation, and a confidence level. The chief merchant could verify, drill down, or push back.
05 — Escalate edge cases.
When the agent's confidence was low, it routed the question to the analyst team — with the partial work already done. Analysts saved time on the easy cases.
The chief merchant got direct access. The analyst team got their queue back.
The Impacts
Decisions that don't wait three days for an analyst.
After the analytics agent went live, the chief merchant operated on her own questions — and the analyst team focused on the deep work that actually needed them.
80% — Of questions answered in seconds.
30 → 6 — Analyst queue at any time.
Sources shown — Behind every answer.
Days → seconds — From question to action.
"I stopped scheduling questions. I started asking them."
— Chief Merchant
Explore more use cases
One platform. Every retail decision.
Inventory, allocation, pricing, planning, execution — connected in a single operational layer.
