Deriv Boosts Productivity and Reduces Onboarding Time by 45% with Amazon Q Business

Millions of customers use Deriv’s financial platform, and over the past 25 years, Deriv employees created a massive amount of content to support its services and operations. This data was hosted in a variety of locations, such as Google Drive, GitHub, internal wiki pages, and discussions and threads in Slack. This made it difficult to locate documents and information about processes, causing inefficiencies and delays in getting new hires up to speed. With more than 1,400 employees in 21 countries, it was challenging to coordinate schedules across multiple locales to carry out onboarding trainings and project meetings.

Deriv originally reached out to Amazon Web Services (AWS) to discuss potential uses for Amazon Bedrock, a service that helps companies build generative AI applications with foundation models. It was during this discussion that AWS mentioned Amazon Q Business, a generative AI-powered application that safely and securely uses a company’s internal data to help employees get work done. The no-code solution allows users to have a conversation with the tool to retrieve data from their repositories, with no machine learning (ML) experience required. Deriv was excited by the potential for employees to have immediate, uncomplicated access to its vast array of data sources.

Deriv had no previous experience working with AI, and had delayed exploring it, assuming it would require hiring a team of senior engineers and researchers to implement and manage. However, as a fully managed service that doesn’t require ML experience, working with Amazon Q Business proved to be extremely intuitive. With a team consisting of a single developer, Deriv set up the data sources it wanted Amazon Q Business to index, then tested a variety of use cases. After just half a day of testing, the solution worked smoothly and launched for employees the very next day. Arun Venkataraman, principal engineer of operations at Deriv, noted, “We were surprised how quickly we were able to implement and start using Amazon Q Business. No one thought working with AI would be this easy.”

Amazon Q Business can personalize its interactions to each individual user based on an organization’s existing identities, roles, and permissions. It comes with built-in guardrails to ensure the responses it provides are appropriate and employees get the right level of information. “Like many businesses, some of our departments have more sensitive information than others,” stated Venkataraman. “Because of built-in guardrails in Amazon Q Business and the ease of setting up access control lists, we’re able to make generative AI available to all of our employees in a safe way.”

Deriv integrated Amazon Q Business across its customer support, marketing, content, and recruiting departments. For new hires joining its customer support team, the company provides extensive manuals that document its contact center processes. However, this has made training labor-intensive. By utilizing Amazon Q Business to index all customer support materials, new employees can ask the solution specific questions and find answers 45 percent faster than before. Rather than manually searching through documentation to find information about a process, Amazon Q Business quickly locates the most recent, relevant information. Previously, it would take new hires a month to get up to speed. After implementing Amazon Q Business, it takes them about one week.

Deriv’s marketing team also saw major benefits from the generative AI solution. With much of their campaign data stored in different locations, they used to schedule calls with multiple stakeholders across various time zones to stay informed. Users can now ask Amazon Q Business about a campaign at any time, and it searches associated data repositories to provide a summarized status. Deriv’s content team is also exploring Amazon Q Business for creating marketing materials, such as technical blog posts. Amazon Q Business can quickly and accurately generate articles based on Deriv’s resources, while leveraging the company’s best practices and guidelines to ensure content quality.

In addition, Deriv extended Amazon Q Business functionality to its recruiting team. Part of that team’s hiring process involves reviewing 10-page candidate questionnaires, which take 20 minutes to read and evaluate. Now, when a recruiter uploads one of the questionnaires to Amazon Q Business, the tool reads and summarizes the data in less than a minute. This efficiency reduced the time needed to review content by 50 percent.

In just a few weeks’ time, Amazon Q Business made a significant impact across Deriv’s departments. The company saw productivity gains in all areas where the solution was implemented, saving employees time by reducing manual document searches and eliminating many meetings. Users also became more comfortable using AI technology, and increased their search efficiency by learning how to create good prompts.

Deriv implemented AWS Global Accelerator and Amazon CloudFront in tandem with Amazon Q Business. These tools helped Deriv improve availability, performance, and security for its applications, and reduced workload latency by 45 percent—from 1.5 seconds to less than 200 milliseconds. This setup helped deliver content to Deriv’s customers faster, and combined with Amazon Q Business, accelerated data delivery to employees.

Currently, only Deriv’s non-technical teams are using Amazon Q Business, but that’s about to change. Its tech stack is due for an update, so Deriv plans to explore Amazon Q Developer to help architect a solution. Using the tool with Amazon Bedrock, it aims to build a code transformer to automatically convert its code and move it to a new stack. Deriv is in the process of building a team for its code migration, and already has Amazon Q Business creating job descriptions. It estimates that in the next few months, its entire organization will be upskilled in generative AI and use it as an integral part of daily work.