How Wise automated 40% their daily tasks with Credal

Founded in 2011, Wise (formerly TransferWise) is a global leader in international money transfers and financial services. With operations across multiple countries and rapid growth, Wise sought an AI solution that could meet their scaling organizational needs.

We spoke with Wise’s ML platform lead, Mark Harley, and product manager Emil Chetty about their journey implementing Credal across the 1,000+ employees working at Wise.

When Wise started looking for a solution, they wanted a secure platform to provide AI productivity that could integrate into day-to-day workflows. 

As a rapidly growing fintech company, Wise wanted the platform to meet three criteria:

1. Model-agnosticism. They needed a solution that wasn't tied to a single AI provider. Considering how fast AI is evolving, there's the risk that one model won't always be the best forever. They also didn't want to go through the process of negotiating contracts with OpenAI, Anthropic, and Google separately.

2. Developer-friendly infrastructure. As an API-first company, any solution needed to have robust API access for easy integration with audit logging and cost tracking.

3. User interface. Offering both a developer RAG platform while also allowing non-technical users to spin up assistants without needing a huge internal build out was critical to getting every employee onto the platform quickly. 

After evaluating various options, Wise found Credal to be the perfect match that served every need.

While Wise initially focused on Credal's platform capabilities, they were blown away by how organically teams started using and deploying copilots throughout the organization.

"There are people in there who are not at all technical. People in operations teams will bring up in passing things like, 'Oh, I created this copilot to do this for my team.’ The entire legal team has got their whole operation built on top of Copilots too,” Mark says.

What was the secret sauce? For Wise, they framed it as how Credal provides an end-to-end journey contained in a singular platform.

A Wise employee would first get access to Credal and start talking to documentation. Then, they would discover files in the Confluence and Google Docs they didn't realize existed to answer their questions. The employee would then create a reusable copilot, referencing those same data sources, come up with a specific set of instructions, and deploy that to a Slack channel for their team. 

The user interface paired with the robust API for integration was unexpectedly the approach that worked for teams to actually adopt AI.

Customer Service QA 

Wise faced a significant challenge scaling its customer service teams to meet growing demand while maintaining high-quality support. The volume of customer inquiries was outpacing the company's ability to hire and train new employees.

In response, the Wise team created a Credal copilot that performs QA for customer support interactions across 14 different Wise-specific evaluation criteria. The copilot was trained on hundreds of different support cases with sample reviews.

This solution enabled to identify correct and incorrect responses, assess customer impact, provide evaluations, and offer specific improvement suggestions.

The business impact was substantial: Wise dramatically reduced quality review time, customer satisfaction scores rose as agents received more targeted coaching, and first-contact resolution rates showed marked improvement. The copilot solution ultimately allowed Wise to scale its customer service operations while maintaining the same operational costs.

Generating Fraud Reports

Another major use case was streamlining the creation and formatting of Suspicious Activity Reports (SARs). When Wise’s automated systems flags a suspicious transaction, a financial operator provides a detailed account of the situation for internal use. However, an external report must also be filed to formally report the suspicious user to law enforcement. This requires completing a structured questionnaire in a specific format for regulatory review.

To simplify this process, Wise developed a Credal copilot to automate the transformation of internal reports into fully completed SAR questionnaires, while keeping a human in the loop.

By automating the conversion of internal reports into structured Suspicious Activity Reports (SARs), Wise significantly reduced the manual effort required for compliance reporting. Financial operators saved hours per report, allowing them to focus on high-value investigative work rather than formatting reports. This also improved the consistency and accuracy of SAR submissions.

Wise employees across the board saw tangible ROI from implementing Credal. In a company-wide survey measuring Credal's impact, nearly half of employees reported saving 20-40% of their daily work time. Even more impressive, 15% of employees achieved time savings of over 40% each day. These concrete results provided the quantifiable evidence to continue investing in AI.

Key operational areas saw particularly strong improvements. Compliance teams saved hours per Suspicious Activity Report (SAR) by automating the transformation of internal reports into properly formatted submissions, Meanwhile, customer service teams benefited from faster quality reviews, more targeted coaching, and higher first-contact resolution rates.

Beyond these individual wins, Credal has fundamentally changed how Wise works. With an end-to-end AI platform that combines powerful APIs for developers and an intuitive UI for all users, Wise has democratized AI adoption across the company. Both technical and non-technical teams can now spin up assistants in minutes, automating routine tasks and delivering better customer experiences.

What stood out to Wise beyond Credal's technology was the genuine partnership that delivered both fast feature development and exceptional support.

"You guys were always willing to put us on the roadmap and get features out the door really quickly," the Wise team emphasized.

The Credal team didn't just build features fast - they were right there when Wise needed help. As Mark put it: "Having a channel where the Credal team is directly talking to our engineers when they're debugging something or building an integration makes everything super quick. No support tickets - things just get done."

Have questions on the above or want to know how these apply these use cases more specifically to your needs? We’re happy to chat. Drop us a line at sales@credal.ai!