The Challenge
A major international pharmaceutical company's sales expansion team faces a unique analytical task: understanding product performance not in isolation, but against the backdrop of each product's lifecycle, go-to-market strategy, and competitive situation — across indications, markets, and time.
That means holistically connecting approval dates, reimbursement timelines, competitor activities, and multi-year sales forecasts and actuals — and surfacing the correlations that explain performance: Is an underperformance driven by a launch delay? A competitor movement? A reimbursement gap? How are key metrics trending, and how will they extrapolate?
Existing tools like Microsoft Copilot, Snowflake, or Anaplan struggle to combine business understanding and data processing into the analytical rigor required to carry out multi-step reasoning across such diverse data sources.
This means business leaders are constrained by prebuilt reports or need to reach out to analyst teams, which are expensive bottlenecks, introduce hours of waiting time, and don't allow for fast follow-up questions. Embedded insights teams are burdened with ad hoc requests above 60% of their time — time better used for strategic analysis. Fragmented data and reports cannot be fully leveraged by the commercial organization.
The Solution
In April 2026, after a successful multi-month proof of value, explai launched its agentic analytics platform with the international sales expansion team, embedded directly into their existing Microsoft-based Business Intelligence.
Senior leaders on the team are now ramping up using explai to analyze global product performance forecasts and actuals. The platform spans simple natural language queries to multi-step reasoning. Key capabilities include:
- Multi-step correlation analysis — explai surfaces performance deviations and their drivers, such as launch delays, competitor movements, or market access changes.
- Competitive intelligence integration — Daily reports from the company's global market intelligence team are ingested and scanned for clinical trial schedules, earnings call signals, and other events that could impact performance vs. plan.
- Tailored business understanding — explai learns company processes, metric definitions, forecasting logic, and analysis templates through its long-term memory system.
- Analysis standards — Agents follow established analysis templates and produce graphical reports aligned with company standards, spanning custom chart guidelines and a network of process templates about how to reason and identify root causes.
- Clarification and guardrails — The system proactively clarifies ambiguous questions and rejects those it cannot answer faithfully, preserving analytical integrity across high-stakes decisions.
What Was Built
explai's agentic analytics platform adapts to the data science needs of any industry and function. For this global commercial analytics team, it was configured for global product performance analysis in a structured project approach:
- In the first weeks: Sandbox — explai stood up a working sandbox quickly, giving the business team a hands-on environment to shape scope and priorities early and continuously provide feedback throughout the process.
- Business knowledge ingestion — The team encoded company processes, metric definitions, forecasting logic, and analysis templates directly into explai's long-term memory.
- Golden-answer database — explai built a growing reference answer set from day one and evaluated against it continuously to track and improve agent quality.
- Agent configuration — explai's data science agent fleet was configured for the use case, with custom agents added on demand — including a market research fact extractor and a competitive intelligence newsletter agent.
- Private cloud installation — explai deployed entirely within the customer's AWS accounts, serverless, with centralized observability and secrets management. No data leaves their environment.
- Live data integration — Live Snowflake syncs give the agents real-time access to millions of records across multi-year approval, launch, and performance data, embedded within the customer's existing Microsoft BI cockpit.
- Rollout — The platform targets 100 users in 2026 and 1,000 in 2027, phased via multiple waves of acceptance testing.
Why It Works
Pharmaceutical commercial analytics cannot be separated from business context. Metric definitions, forecasting processes, product lifecycle stages, and market-specific launch dynamics all shape how data must be interpreted — generic AI tools can't operationalize this reliable end-to-end analytical workflow.
explai combines configurable agents, long-term business memory, and rigorous evaluation to deliver the nuanced, multi-step analysis pharma leaders need, at the speed and flexibility business demands.
Path to 1000 Users
The explai standard product is iterating fast to safeguard the ramp-up throughout 2026 and get ready for large-scale adoption in 2027:
- Plan-based reasoning: Adapt latest reasoning models for more faithful knowledge instruction following.
- Data quality checks: Profile incoming data for high accuracy.
- Proactive knowledge curation: Agents continuously curate long-term memory based on usage to speed up and standardize subsequent answering.
- Generative UI: Interactive answer widgets for diverse answer templates and exploration with fewer turns.
- Consulting Efficiency: Increase automation for business consultants and forward-deployed engineers for optimization and extensions.
- Question guardrails: Senior or casual users don't have the patience to follow question guidelines. Clarify rigorously.