OutSystems Agent Workbench
After many years and thousands of digital transformation initiatives, I have seen that an organization's digital transformation is fundamentally a macro expression of the digitalization that happens in each team. Today, digital transformation comes with an opportunity to leverage artificial intelligence (AI). So, if you want to propagate AI across an entire organization, a good starting point is to implement it in a team.
Very much like how the value chain of a company is composed of different clients and suppliers, internal teams in an organization have a network of dependencies: a given team receives requests it must answer, and it depends on the services it requests from other teams. This process is usually streamlined by implementing help desk or "ticketing" workflows–the equivalent of organizations using digital systems to support their customers. Now, where can AI start making a difference?
Optimize high-frequency tasks with AI automation
When a team identifies their most repetitive tasks and standardized work, they can begin to optimize it. This is what we might call the ground zero of automation–the team creates a self-service portal that replaces human interaction for less complex, high-frequency tasks. This way, the organization has internal "customers" taking care of their own needs in real-time and being able to continue their tasks much faster, without adding toil to "suppliers" from different teams.
In this case, you can leverage AI to take efficiency up a notch by analyzing the most frequent tasks on the portal, monitoring SLAs, and inferring optimizations. By understanding which 5% of the portal transactions are bringing teams 80% of the results, an AI system will help you discover which tasks you should optimize.
Support deflection and UX optimization
Even if a self-service portal serves multiple internal needs, documentation complexity may leave people with questions that require additional assistance. In this case, they fall back on teams that serve as knowledge clusters. An HR team, for instance, is constantly engaged in answering specific employee-related questions. These requests require them to sift through personal memories and documents whenever they need to answer a question.
A better solution would be to have a chatbot, powered by a large language model (LLM) trained with company data, ready to answer various employee questions in seconds. In the past, this required AI expertise and effort. But now there is OutSystems AI Agent Builder. Teams can use it to build generative AI (GenAI) agents into new or existing applications quickly, without any prior coding or AI expertise. Plus, it enables them to govern the AI capabilities and services used in those applications. These agents can serve multiple purposes, such as ticket deflection, creating ChatGPT-like apps, performing data extraction and summarization, and generating personalized content or images.
GenAI systems are ready to tackle nearly all aspects of onboarding, knowledge transfer, or customer care and dramatically improve user experience. For instance, to submit an internal request today, there is a digital form to fill out. But we are already seeing free-text inputs, sometimes complemented with dictation interfaces, that GenAI models automatically restructure into form fields, saving hours of effort and significantly improving user experience.
Workflow orchestration
One final example we can pick up from digital transformation processes is workflow orchestration between teams. When an internal request requires connecting with people outside a team–to ask, for instance, for additional information or to complete a necessary sub-task–previous generations of digital systems would resort to enterprise-wide workflows to gather everyone around a "case" or "ticket.” Organizations could also use a composite application that would interact with multiple other systems via APIs.
Today, this is another area for AI disruption. AI models can now perform workflows and API compositions in real time, replacing deterministic systems limited to previous definitions. AI can connect the dots between internal systems, transforming entire interdepartmental workflows into a composite API.
AI will transform the nature of work
The challenge behind all of these examples is that they change the nature of how you design and control these systems. There used to be a clear difference between interacting with biological systems and digital systems. But AI models have shown us that they act more like people, instead of resembling traditional software. They will make mistakes that might require auditing and supervision or might have the equivalent of breakdowns by hallucinating. Unlike software, they cannot be bug-corrected– they must be trained.
This stresses the importance of having the right tools for leveraging AI in your organization. It’s also why we're deeply invested in helping our customers experiment with AI. Consider starting at the team level. This approach will allow you to identify organizational patterns, opportunities, and risks as you move forward. Ultimately, and much like digital transformation did for organizations in the 1990s, AI automation will fundamentally revolutionize the nature of work – and the time to begin is now.
