Comp Wise – Enabling Line Management | Syngenta User Case

In this session, Anukriti Kumar, Global Compensation & Benefits at Syngenta Group, shares how Syngenta built CompWise, an internally developed salary prediction tool that uses country-specific regression models to support objective, consistent and auditable pay decisions across 90 countries.

Currently piloted within Syngenta’s Compensation & Benefits team, CompWise is designed to support data-driven decisions for new hires, promotions and off-cycle adjustments replacing gut feel with a transparent, documented methodology.

Anukriti will walk through the build vs buy rationale, how the tool works in practice, and what other organisations can learn from taking an internally built approach to pay decision support.

How can organizations prevent avoidable pay gaps before they arise? This user case explored how a predictive compensation model can give managers an objective, explainable reference for decisions involving new hires, promotions, lateral moves, and salary adjustments. The session emphasized that analytics should support professional judgment, not replace it.

• Put compensation philosophy before technology. A predictive tool is credible only when the organization has clearly defined which factors may legitimately influence pay, why those factors are objective, and how they should affect outcomes. Technology should implement an approved compensation philosophy rather than compensate for an unclear one.
• Move from correcting pay gaps to preventing them. Traditional pay-equity analysis identifies unexplained differences after decisions have been made. Predictive decision support can move the intervention point forward by providing an objective reference range before a hiring, promotion, or adjustment decision is finalized.
• Do not treat historical data as an instruction. Existing pay data may contain legacy arrangements, delayed promotions, inconsistent records, and unusual premiums. Model coefficients and outputs therefore need human review, logical correction, and alignment with the organization’s documented compensation principles.
• Make recommendations explainable and exceptions useful. A transparent range, supported by clear factors, is more valuable than a single unexplained number. When a decision falls outside the range, documenting the reason can preserve appropriate discretion while revealing recurring skills, qualifications, or market conditions that the model may not yet capture.
• Pilot for trust, governance, and adoption. A phased pilot can compare model outputs with the established compensation process while assessing usability, data quality, exceptions, role clarity, access, and integration. Workforce planning or budgeting may provide a lower-sensitivity starting point before extending the approach to individual pay decisions or manager self-service.

AI generated by Copilot from designated meeting content only | Rumbold