Case Study Banner

CASE STUDY

Clinical resource planning: Replacing monthly reconciliation with evidence-backed decision support

industry-iconCLIENT :Confidential
industry-iconINDUSTRY :Pharmaceutical
industry-iconDURATION :3 months
CLIENT :
Confidential
INDUSTRY :
Pharmaceutical
DURATION :
3 months

Business case

Clinical resource planning at this large pharmaceutical organization ran across two structurally incompatible SaaS systems and processes. Capacity (FTE and FSP) was held by role: the designation and available supply of data managers, statisticians, clinical operations leads and comparable specialist functions.

Demand was forecast by study, derived from each program's phase plan and milestone schedule. The gap: demand versus capacity at detailed skill-role level.

The two systems did not connect, and nothing aggregated them into a common role-and-month view of demand against capacity to support workforce allocation, budget and resource trade-off planning.

Reconciliation was therefore manual and monthly. The planning horizon followed reporting cadence rather than portfolio movement: a role shortfall surfaced at reconciliation, often after the demand peak behind it had passed. Resource constraint was managed retrospectively, not in advance.

The specialist roles most exposed to concurrent demand peaks typically have the longest lead times to acquire, whether recruited or contracted. A shortfall found during or after the peak removes the lower-cost mitigation options (reallocation across studies, resequencing, phased onboarding), leaving only expensive ones, and heroics, to keep milestones on track.

Pre-emptive scenario analysis was also constrained by the system set-up and cycle: a routine portfolio enquiry, such as the role-level effect of a two-month acceleration or slip in one study, took several days. The figures carried no audit trail, because values assembled through manual reconciliation could not be substantiated to source, which limited their standing in governance review even where accurate.

Our solution

i2e designed Agentic Resource Management (ARM), a digital accelerator that holds effective role capacity against forecast demand for every role and month across the portfolio on a single planning grid, with inbuilt AI for natural language data interaction.

The engine is anchored on baseline integrity and fed by source portfolio and workforce capacity data. Each study carries one immutable, versioned baseline forecast, so variance is measured against a fixed, auditable reference, not a working position that has drifted between planning cycles.  

Five capabilities operate above that baseline:

  • Portfolio-level scenario modelling Milestone or assumption changes applied to one to three studies return their portfolio-wide consequence within minutes, removing dependence on the monthly reconciliation cycle.
  • Variance-to-baseline reporting Each scenario returns demand, capacity, uncovered demand and headroom, supporting gap identification and assessment of absorptive capacity elsewhere in the portfolio.
  • Role and month granularity Outputs sit at the level where resourcing decisions are executed, not at an aggregate that conceals role-specific constraint.
  • Full provenance Each figure carries a run identifier, scope, role, month and unit, giving an audit path from executive brief back to source calculation.
  • ARM AI Agent Answers plain-language portfolio queries and generates the executive brief, removing the interpretation step between scenario output and decision input.


ARM connects to existing planning inputs, such as study plans, milestones, role capacity demand algorithms, additional data streams for retrospective timecard actuals (to validate the accuracy of forecast demand) and resource cost rate cards, converting planned activity into realistic effort and cost analysis. The systems of record always remain unchanged; ARM makes their combined position query able and provides the intelligence layer for operational resourcing excellence at scale, in real time.

Challenges overcome

  • Resolving two incompatible planning dimensions: capacity held by role and demand held by study required mapping to a common role-and-month grid before portfolio-level comparison could be made meaningful.
  • Maintaining baseline integrity whilst performing rapid scenario modelling: scenario flexibility delivers value only against a stable reference. Immutable versioned baselines preserved comparability while permitting continuous replanning.
  • Establishing provenance in a governed environment: accelerated analysis has limited standing if its outputs cannot be substantiated. Run-level traceability was designed into the engine rather than appended to it.
  • Delivering decision-ready output: scenario results requiring manual interpretation reintroduce the latency the solution was intended to remove. ARM Agent closes that gap by producing the executive brief directly from the run

Benefits

Who: Clinical operation team, line managers, leadership teams, portfolio and finance managers.  

  • Scenario analysis resolved in minutes rather than days, allowing resourcing questions to be addressed within team to governance forums rather than after them.
  • Role-level constraint identified ahead of the demand peak, preserving access to lower-cost mitigation options and protecting budgets.
  • Demand and capacity expressed at role and month granularity, matching the level at which resourcing decisions workforce planning & portfolio trade off decisions are made.Full provenance on every figure, supporting defensible numbers in portfolio governance. Increased trust in resource data and hence leadership decisions
  • Executive briefing generated directly from the scenario run, removing the manual write-up step between analysis and decision. 

Results

Resourcing moved from monthly reconciliation to evidence-backed decision support.

Scenarios are resolved in minutes rather than days.

Demand and capacity are reported at role and month granularity.

Every figure is traceable to the run that produced it.

SPM Icon
SPM maturity calculator