OutSystems drives CX by automating 35% of support operations with generative AI

Challenge

Generative AI: Countless digital innovation opportunities

As CIO at OutSystems, Tiago Azevedo is responsible for leading the Digital Team in building a world-class business architecture and working closely with business partners to understand challenges and equip their teams with innovative solutions that support their strategies.

So, when generative generative AI took off, Tiago and his team—backed by the CEO who recognized the importance of fast innovation—embraced the challenge to see how OutSystems could quickly tap into generative AI’s transformative potential.

Like most companies getting started with generative AI, OutSystems faced a few obstacles: “The challenge seemed daunting at first, but we gathered a few people with knowledge of gen AI internally, segregated functions across the board, and created a cross-functional collaboration framework that empowered each individual to contribute their expertise,” Tiago said. But with endless opportunities to consider, it made it difficult to know which use case to tackle first.

Given the company’s commitment to customer success and the fact the support team needed a better way to scale to accommodate rapid customer growth, delivering a generative AI solution that could help automate support operations and free up the team’s time to focus on more complex support tickets was a no-brainer. And that’s exactly what they did.

Solution

Generative AI-powered chat agent: Automating support, elevating the customer experience

After hosting brainstorming sessions to identify the first use cases, the team started on their generative AI journey with the goal of, streamlining support operations and automating as many responses to support questions and incidents as possible. Using the OutSystems platform, the team delivered their first customer-facing conversational generative AI agent—the AI-Powered Support Assistant.

The AI agent is live on the Support Portal and autonomously provides requested information by integrating with large language models (LLMs) that pull data from various sources, including the OutSystems website, internal documents, and training materials. This, along with prompt engineering, contributes to the production of accurate and useful responses to customer inquiries.

During that time, the team also worked with other departments to identify additional business needs that could be solved with generative AI—with a focus on small use cases that could make a big impact. From launching the generative AI customer support agent to exploring new use cases, the team saw many similarities that could help them build a generative AI framework and deliver solutions at scale.

“We focused on two small use cases that were low-hanging fruit and quickly saw many similarities. That’s when we saw patterns emerging,” João Araújo, Tech Lead on the Generative AI Foundations team, said.

From there, João and his team identified four patterns that could help them execute their generative AI strategy and quickly produce new solutions.“Now, when a team submits a use case, we can quickly analyze it and attribute it to a pattern. This way, we already know how to solve the issue. It allows us to be incredibly efficient,” João said.

And as the team tackles more generative AI use cases, the patterns are continuously evolving to ensure they can continue to scale generative AI efforts with ease.

Results

Better customer experiences, reduced costs

The impact of the generative AI customer support agent has been incredible. In the first year since it was rolled out to customers, the AI agent has successfully answered over 62% of support questions and 21% of support incidents, helping enhance the overall customer experience.

Since the AI-Powered Support Assistant leverages the entire OutSystems knowledge base to provide the best answer, customers don’t have to search through all the documentation and can get a more immediate answer. “The generative AI-powered support assistant was able to significantly increase the number of cases that did not require the assistance of a representative, resolving support cases at an unprecedented rate,” Tiago said.

In addition, the generative AI support agent has helped reduce new ticket creation by 34%, saving the company over $600,000. Overall, OutSystems has been able to automate 35% of its support operations, allowing the company to keep up with ongoing growth without the need to expand resources.

And by using the combination of low-code and AI with the idea of patterns, OutSystems builds generative AI solutions much faster and with less effort. “With this framework, we can reuse the same application and just change the context and the information feeding it to suit another use case. This allows us to scale our generative AI strategy and continue to rapidly innovate to support our customers and employees,” Tiago said.

For the team, it’s all about bringing AI to where the user is. For example, OutSystems has reused the customer support portal agent to create a similar solution on top of its human capital management platform Workday to answer employees’ questions. And now, the team has delivered several generative AI solutions across use cases, including marketing and sales.

With generative AI’s undeniable power to generate transformation, there’s a lot more on the horizon for OutSystems. What started as a question of “where do we begin” is opening up a whole new world of possibilities to continue delivering exceptional customer experiences, boost productivity, and maximize ROI.

For companies still unsure, Tiago shared this advice: “Integrating low-code platforms with generative AI is no longer a trend—it’s a game-changer. This combination empowers businesses to innovate faster, streamline processes, and deliver value at an unprecedented scale. My recommendation is to embrace low-code and generative AI and become a low-code shop. This will enable your developers to learn generative AI on the job and replicate use cases, allowing you to scale your generative AI strategy quickly across the enterprise.”