Smarsh strengthens AI governance to power operational velocity

Building a centralized, composable foundation.

To break through these operational bottlenecks, Smarsh set out to implement a robust integration strategy led by IT operations. The core goals were simple yet vital:

• Build a standardized, uniform integration layer to replace messy, one-off setups.
• Automate customer setup, usage tracking, and financial reporting across 12 different platforms.
• Establish strict data rules and cleanliness to unlock the business’s massive potential for AI acceleration.

“Our data governance wasn’t where it needed to be,” says Tom Hamilton, Business Systems Engineering Manager at Smarsh. “As a business, we recognized that a strong data governance model, a clear data strategy, and new AI workflows would be the differentiators that keep us ahead of the market.”

Harmonizing data and AI orchestration with MuleSoft Anypoint Platform.

Smarsh selected MuleSoft Anypoint Platform as its enterprise integration standard, leaning into its industry-leading scalability and tight alignment with the broader Salesforce ecosystem already in place.

Initially brought in to solve core setup and tracking challenges, MuleSoft quickly became the operational engine for Smarsh’s internal data strategy. By standing up a centralized connection layer, Smarsh successfully unified its internal metrics, pumping automated usage and billing details directly from individual software systems into a centralized Snowflake data warehouse.

Furthermore, as Smarsh scaled its use of generative AI company-wide, MuleSoft provided critical infrastructure tools to implement cost controls and AI usage management.

“We’ve rolled out Claude across the organization, and MuleSoft enables us to track usage, enforce governance, and see who’s using what tokens,” Hamilton adds. “It’s quickly moved beyond an integration tool.”

Utilizing MuleSoft Model Context Protocol (MCP) support, Smarsh’s IT team can rapidly prototype secure, reusable AI software creation processes. This allows the business to safely turn backend systems into secure, AI-ready tools, giving developers the ability to build advanced automation while preserving enterprise guardrails.

“We’ve been very clear that AI is not here to remove headcount,” Hamilton says. “AI will help us take what we have and accelerate it. Even where 98% of the code is being written by AI, every bit of the architecture, the design, the oversight, and the long-term maintenance of all that code is the responsibility of a human.”

Reclaiming leaked revenue and slashing time to value.

Today, the operational maturity powered by MuleSoft has converted Smarsh’s IT team into an outright revenue-generating partner for the business.

By automating usage tracking through MuleSoft in support of its new data strategy, Smarsh eliminated systemic data blind spots. This seamlessly uncovered $2.2 million in missing, previously unbilled Annual Recurring Revenue (ARR).

For the vendor risk management (VRM) platform, a customer onboarding process that once required over five days of manual work now triggers automatically via MuleSoft connections, completing end-to-end setup in just three seconds, a 99% reduction in customer time-to-value.

Month-end financial close processing time has dropped from an unpredictable five-plus days to one consistent day to process invoices across the entire corporate portfolio.

Backed by a 4x asset reuse ratio, Smarsh’s six-person IT team achieves staggering project velocity. Leveraging its new AI software creation process powered by MuleSoft, a one-time payment connection project that normally required 900 hours was delivered in under 90 hours, with just 10 hours spent on actual code development.