• 151

    Projects submitted

  • 109

    Published projects

  • 47

    Comments left

  • 66

    Number of votes

  • 52

    days until voting ends

← Back to list

Quant BI: Sales and Margin Analytics for a Pharmacy Chain

  • Project manager on customer side

    Temirlan Anuarkhanov

    Sadykhan Group

    IT-Director

  • Category

  • Nomination

  • Goals

    The pharmacy group needed up-to-date figures on sales, performance against targets, stock levels and margins. Managers had been combining reports from several 1C databases by hand. The goal was to give them a common view of the business so they could spot problems earlier and investigate their causes.

    Pharmacy directors, regional managers and head office each needed access to the figures for the operations they oversaw. The wholesale division needed reporting on sales, orders and accounts receivable. The team also had to match product records across databases and calculate the cost of goods sold when a pharmacy sold only part of a pack.

  • Timeline

    november, 2025 — october, 2026
  • Project completion year

    2026

  • Project scale

    57 automated workplaces
  • Results

    Client figures as of 2-3 November 2026. In the previous 30 days, 46 retail employees used Quant BI, including 36 of 42 pharmacy directors. Over 90 days, 57 users were active.

    Since 11 February 2026, the system has flagged 4,887 exceptions. Of the 920 cases closed, eight employees closed 125 manually.

    Across two major chains, the platform calculated costs for sales representing 99.7% and 99.4% of revenue, respectively. For the largest chain, its margin figure differed from the accounting figure by 0.4 percentage points.

    For the months of 2026 with complete data, potential purchasing savings totaled KZT 195-256 million, or 3.1-3.4% of purchases. The estimate compares prices paid to different external suppliers for the same product in the same month, before supplier rebates. For October alone, potential savings across the network were KZT 34-43 million.

    Write-offs totaled KZT 190 million, or 1.5% of revenue recorded on sales receipts. Inventory counts accounted for 95% of write-offs.

  • Project uniqueness

    Quant BI calculates the margin on a sale even when a pharmacy sells only part of a pack. It uses the purchase price of the specific batch and the number of units in the pack to determine the cost of the units sold. To make this calculation consistent across pharmacies, the team matched 81,388 product records from three retail databases to 38,599 products in a single ERP catalog.

    The system flags sales below cost and pricing errors on partial packs. A pharmacy manager can open a flagged sale, review possible causes and next steps suggested by the AI assistant, assign the issue to an employee and track it through to resolution. An AI assistant also checks the initial analysis when managers review an individual pharmacy’s performance.

  • Software used

    The project uses Quant BI, Python, FastAPI, Flask, ClickHouse, PostgreSQL, React and TypeScript. The AI assistant uses Anthropic Claude and OpenAI. Data comes from 1C:Enterprise 8 and Microsoft SQL Server. A custom agent and a 1C extension extract changes from the source databases. The platform also uses a Telegram bot and Docker Compose.

  • Solution from the Global CIO catalog

    The project does not use solutions from the Global CIO catalog

  • Project implementation complexity

    The team had to match product records across several 1C databases. Matching by name alone sometimes linked a record to more than one product and distorted the consolidated figures. The team solved this by matching products by code and item number, using the product name only when it identified a single match.

    Staff could repost documents with an earlier date, and a pharmacy’s transaction history could span several databases after a migration. To keep reports accurate, the team set up the platform to capture changes from 1C and reload data for earlier periods.

    To avoid putting extra load on the live accounting databases, the platform reads from database copies and staging schemas. Each user can access only the data relevant to their role.

  • Description

    The client is a pharmaceutical group in Kazakhstan with more than 100 pharmacies and a wholesale division. Its four legal entities use several 1C databases and process around 13,000 sales receipts a day. Staff used to combine data exports by hand in Excel. Product records differed across databases, making it difficult to compare pharmacies and spot a decline in sales early. Partial-pack sales also needed to be accounted for when calculating margins. In October 2026, they represented 8.5% of revenue.

    The team introduced Quant BI across the retail division. Pharmacy directors, regional managers and head office now have hourly dashboards showing revenue, progress against sales targets, sales trends and receipt data. Each manager sees the operations they oversee. Since February 2026, the platform has flagged exceptions automatically and sent a morning summary via Telegram. Managers use the AI assistant to investigate issues, assign follow-up tasks and track their status.

    The wholesale division received dashboards for sales, margins, stock levels, orders and accounts receivable. The team also piloted an AI assistant for wholesale staff.

    In October 2026, the team built a shared data layer that receives updates from seven 1C databases and holds around 126 million rows. It reconciles documents against 1C, calculates costs by batch and refreshes purchasing, write-off and exception data every 15 minutes. Migration of retail reporting to the shared data layer and rollout to all wholesale staff are ongoing. The next phase will incorporate supplier rebates into margin calculations.

  • Project geography

    Several regions of Kazakhstan.

Comment

Only authorized users can comment.
We suggest you or register.

  • Customer

    Sadykhan Group of Companies

  • IT provider

    SOWITA

We use cookies for analytical purposes and to deliver you the best experience with our website. Continuing to the site, you agree to the Cookie Policy.