As CIOs scramble to get ahead in the AI arms race, copilot technology will become increasingly important to enhancing efficiency, collaboration and accuracy. Gartner forecasts that by 2027, 70% of professional developers will be using AI-powered coding tools, up from less than 10% in September 2023.1
Meanwhile developers using the GitHub Copilot are seeing productivity boosts by as much as 20%.2
Development teams must therefore understand how best to use AI co-pilots, such as Microsoft’s industry-leading, AI-powered developer tools, to accelerate innovation as boardrooms urgently seek to gain competitive advantages.
The productivity dividend
Technology firm EPAM is helping development teams leverage Microsoft Copilot.
As Roman Pavlov, Solutions Architect at EPAM, explains: “While this varies according to seniority levels, on average developers only spend around an hour a day coding. Using copilots to address the rest of a developer’s workload, which includes attending meetings and troubleshooting issues, will therefore have a bigger impact on overall productivity.”
Key developer use cases for copilots
EPAM cites a wide range of compelling copilot use cases for the developer community. These include:
- Boosting productivity. Usually, copilots increase developer productivity by 10-15%.
- Providing quality assurance. Helping developers run unit tests faster to ensure the quality of their applications before they go into production.
- Boosting creativity. Providing a good starting point at the beginning of the development process, helping developers generate and validate ideas.
- Assisting administration. Helping to summarize content from Teams, JIRA, Confluence and other collaboration tools, enabling developers to stay on top of workflow.
- Training. Helping new developers get up to speed on corporate guidelines, processes, and preferred ways of working.
- Accelerating PoCs. Allowing developers to explore PoC (proof of concept) pathways faster and ramp up the number of PoCs and PoVs (proof of value) they are able to produce.
Considerations for success
As businesses look to leverage Microsoft Copilot and GitHub Copilot, there are a number of considerations to keep in mind. One is who exactly should use copilots within the development team.
According to Alexey Taranenko, Azure Technology Consultant at EPAM, senior developers benefit the most: “We see more of a productivity boost for mid and senior developers than we do for juniors, because they have the grounding in code needed to use these tools to their fullest potential. Junior members should be given space to learn these skills so they can apply copilots successfully later in their careers.”
Another important issue is around transparency and interpretability. Copilots must not operate within a black box, particularly in heavily regulated industries like healthcare and financial services. Says Pavlov: “Developers and the wider business must have visibility of how copilots come to decisions and outcomes as they will be held accountable if anything goes wrong.”
Microsoft Copilot can help provide this transparency, as they each record a comprehensive history of activities, providing transparency into user interactions, actions taken, and feedback received, allowing developers to see what decisions the Copilot made and correct them if necessary.3
Alexander Lipkin, Director Technology Solutions at EPAM, adds another note of caution: “In some cases our clients do not wish to use large language models or copilots for their models, and we ensure that our engineers adhere to these policies through the use of internal tools, processes, and training. Even when clients are in favor of using copilot, there’s often a stipulation that their data isn’t used for model training. Microsoft enables us to address this requirement through strict privacy and security measures.”
Where businesses strike gold and see a use case deliver value, the next action is to scale that across the enterprise.
Adds Lipkin: “If people start using copilot for particular use cases, we have to reward those people and promote the use cases across organization. That’s exactly how EPAM adopted copilot internally, and we are trying to scale this experience to our customers.”
Learn more from EPAM on what the AI future has in store for enterprises.
3 Microsoft, “Microsoft Copilot Studio: Building copilots with agent capabilities.” May 2024