Voya Financial Redesigns Workflows First, Adds AI Second

Process before technology

Voya Financial is taking a disciplined approach to artificial intelligence that runs counter to the hype cycle: fix the underlying process first, then determine whether AI belongs in the solution.

Under Chief Technology and Operations Officer Santhosh Keshavan, the financial services company—which generated more than $8 billion in revenue last year—has spent years modernizing core systems, consolidating platforms, and rethinking customer workflows before selectively embedding AI capabilities. The goal is a unified digital experience where customer context persists across channels, eliminating the friction of repeated logins and redundant data entry that can derail simple transactions.

The work began well before generative AI entered the enterprise mainstream, driven by quality, cost, and customer experience imperatives rather than technology trends.

Insourcing control over end-to-end workflows

A critical early step involved bringing outsourced technology capabilities in-house. Previously, a single 20-step customer process might involve multiple external vendors handling different components, fragmenting ownership and slowing execution.

Voya moved to a single contact center platform and consolidated other systems to gain end-to-end visibility and control. Keshavan emphasized the importance of owning key workflows directly rather than delegating them entirely to partners, even as the company continues to work with external vendors where appropriate.

That consolidation created the foundation for examining customer journeys holistically and identifying where processes could be redesigned from the ground up.

Eliminating a form beats automating it

In Voya's Employee Benefits business, filing supplemental health accident claims traditionally required customers to complete a complex form. Incomplete or inaccurate submissions triggered back-and-forth exchanges with representatives, extending processing times.

The team initially considered using AI to pre-populate fields and automate the existing form-based workflow. But after consulting with customers and internal legal teams, they determined the form itself was unnecessary. Eliminating it reduced processing time from two months to two hours while improving accuracy, with human verification still in place.

Keshavan noted that this outcome—dramatic improvement without AI—illustrates the value of questioning the underlying process rather than defaulting to technology.

AI where it delivers clear outcomes

Where AI does play a role at Voya, it's embedded selectively based on specific business cases. Within the Wealth Management division, AI supports code development, identifies potential errors earlier in the software development lifecycle, and powers virtual agents that handle routine customer inquiries.

In 2025, AI-powered self-service in Wealth Management deflected 2.81 million calls, while virtual agent interactions received 96 percent positive customer feedback, according to a company spokesperson.

The company has also built digital-first tools for retail wealth clients, including faster account rollovers, with AI working behind the scenes to reduce friction.

Treating AI tokens like headcount

Scaling AI responsibly requires cost discipline and clear prioritization. Voya now applies the same scrutiny to AI token usage as it does to human capital requests. Teams proposing AI-powered projects must justify the investment with business cases that estimate token consumption and tie it to measurable outcomes.

That discipline extends to evaluating whether a less expensive model, a conventional API, or traditional automation can achieve the same result. Without these controls, Keshavan said, budgets can be exhausted quickly.

The company also invested in AI literacy across its workforce. More than 99 percent of Voya's global employees completed AI literacy certification within three months, ensuring teams speak a common language around the technology even as its trajectory remains uncertain.

Why it matters

Voya's approach offers a counterpoint to enterprises racing to deploy generative AI without addressing underlying system complexity or workflow inefficiencies. By consolidating platforms, insourcing critical capabilities, and redesigning processes before layering in AI, the company has achieved measurable improvements in speed, accuracy, and customer satisfaction. The discipline around cost controls and business case requirements provides a practical model for scaling AI sustainably as the technology matures and token costs become a significant line item.

Details of Voya's modernization and AI strategy were first reported by Steven Norton for Forbes.