Accenture uses Hazy synthetic data for innovation hub candidates
Blog

Case Study: Accenture uses Hazy synthetic data to evaluate third-party innovation partners

By on 28 May 2020.

Patrick Hegarty

Patrick Hegarty is the Ecosystem & Partnerships Lead for The Dock, Accenture’s Global innovation and R&D centre. It’s Patrick’s job to identify and build partnerships that discover new ways to fulfill human needs, in areas like in finance and healthcare.

Hazy helped Patrick start a data-based innovation project from idea inception to project kickoff in three weeks. Without Hazy, the project would have to have waited until a customer provided a sample data set that may have been incomplete and taken months to provision.

Patrick first heard about Hazy through the Accenture financial services account team via the Accenture FinLab program and immediately saw the potential for Hazy to solve one of his team’s biggest challenges.

“One of our biggest challenges as we develop solutions is obtaining customer data,” Patrick said

The Accenture Dock is probably better than most at this as they partner with a client from the beginning, but still, even with the trusted Accenture brand, they can find it hard to access their customer’s data to build and train solutions on.

Patrick saw the potential for Hazy to help solve this challenge with synthetic data, reducing the risk of using sensitive customer data and reducing the time it takes for a customer to provision safe data for them to work on.

“Hazy can help accelerate our work with synthetic datasets,” he said.

Patrick saw this challenge surface in the Fostering Better Finance project, a collaboration between Accenture’s Banking business and pilot clients to identify vulnerable customers and intervene to promote better financial decision making.

The Accenture Dock team wanted to “build a solution that identifies interventions based on the research and testing we’ve undertaken, like irregular customer spending habits or a number of different triggers,” he explained.

This solution was based on banking transaction data, which is a particularly sensitive data set that banks keep under lock and key. No banking customers were willing to share the raw dataset with the Accenture Team without significant governance processes. This caused long delays.

“We had identified the first pilot client but had not begun to put them into development yet. We were waiting for the data to do that,” Patrick said.

He saw the opportunity to work with a Hazy synthetic transaction dataset to get started on the project before the real customer data became available.

“We went from the first take-off meeting to receiving the first synthetic data set and training our intervention identification model in under three weeks,” Patrick said.

“We’re using the Hazy dataset to verify different sequences and different assumptions we made around customers using qualitative and quantitative assumptions before, but Hazy allows us to test that on as realistic customer data.”

Patrick Hegarty, Accenture

When asked how they knew if the Hazy data was accurate or high quality, he responded that The Dock’s analytics team got involved in assessing the distributions of events and verified how consistent the transactions were, and that there weren’t any anomalies.

“We are looking at delivering the next generation of digital banking. It’s a behaviourally driven digital banking platform that seeks to promote financial wellness,” he said.

“For us, Hazy was solving a pain point we deal with often at The Dock, running around obtaining accurate customer data,” Patrick explained.

Using Hazy, Patrick was able to kick the technical project off quickly and demonstrate its efficacy to potential clients before ever having to access their raw data.

Subscribe

Synthetic data newsletter

Sign up for our sporadic newsletter to keep up to date on synthetic data, privacy matters and machine learning.