SaaS series - PagerDuty: Gen AI & automation imperatives shaping cloud connectivity
This is a guest post for the Computer Weekly Developer Network written by Manu Raj in his capacity as senior director for enterprise data, analytics & AI/data science at PagerDuty.
Pager Duty is known for its Digital Operations Management platform that aims to integrate machine data alongside human intelligence on the road to building interactive applications that benefit from optimised response orchestration, continuous development and delivery.
Raj writes in full as follows…
Automation is the heart of digitisation, alleviating increased workloads, remedying more tasks and enabling connectivity between different systems essential for SaaS. Key to the success of SaaS and expected to grow to $908bn by 2030, is robotic process automation (RPA), its market forecast to grow to more than $13bn by 2030.
RPA automates data and process flows together at a granular level to enable inter-application data and process integration between your users or any source and target systems.
It can use AI to make sense of unstructured data and help guide automation designs, enabling them to become smarter as they execute and augment the skills of IT professionals with intelligent recommendations, predictions and decisions. In this regard, traditional RPA is changing to a combination of RPA, GenAI, AIOps and incident management together, enabling digital workforces to augment their organisations and complement human skills to make IT teams more efficient.
Automation & cloud
Automation is already a critical part of cloud connectivity in the space that APIs used to dominate. With connections between SaaS services being made in several ways – notably via APIs, event-based triggers, or webhooks, there’s a greater role for automation to solve and predict issues in real-time, using GenAI. This allows engineers to become higher-level troubleshooters rather than the ‘ticket punchers’ of times past. This is because GenAI can create a wide variety of data to generate new and unique outputs which will be critical for overseeing the exponential evolution of cloud.
It used to be the case that RPA was mainly used to help with two broad types of issues. The first was high impact but predictable problems solved by simple automated steps. This would typically be automating the provisioning of users, answering customer queries, processing orders, etc.
The second category of issue related to data quality alerts, statistical monitoring and automated business processes. In these kinds of issues, teams can apply their knowledge of the business to create custom playbooks. However, this could be time consuming and difficult to keep up with the pace of change at SaaS providers. Additionally, solutions to automation and playbooks may not always be fully automatable and may require human input to integrate automated solutions with the SaaS provider. GenAI and RPA can be a prominent solution here to identify, build playbooks, automate repetitive actions, build and implement pre-scripted solutions from known or even create new repositories.
Humans are not enough
The growing and complex SaaS landscape now needs more than a human response if IT teams are to return a highly available, quality service. A key question for DevOps teams is how many use cases can they manage to write a playbook about? Now they must identify use cases and have engineers write out the scenarios.
This is where GenAI comes in to monitor common issues and automate writing the scripted response. Engineers can then investigate, validate and use the scripts to prescribe solutions. This can be applied to a third type of systems failure that we already see that requires automation to offer predictive solutions.
This occurs when say data quality alerts may be affected by issues that may not be obvious. Here, GenAI can be used to predict automated solutions to application errors or corrupted calculations. This type of process automation helps with statistical controls and for large business operations. Today, we are faced with more event-based triggers between SaaS systems and the challenge of identifying and managing these in real-time calls for an AIOps platform to support the engineering and data operations team. A plethora of SaaS connections are happening in real time. Issues can’t be solved with a ticket-based, human response.
What’s needed is a modern platform to bring down the downtimes automatically, an area where regular ticket-based processes often fail. AIOps, automation and GenAI-driven playbooks allow engineers to respond effectively at pace.
The data ecosystem is evolving
Manu Raj, senior director for enterprise data, analytics & AI/data science at PagerDuty.
All three methods mentioned above talk to external systems and APIs alone are not the only method of enabling this connectivity. Large providers now gather and create data and application marketplaces and share it with other cloud providers. It’s much more secure and can do so outside of traditional channels within the data centre. This mechanism supports event-based triggers across SaaS cloud providers and makes most current mechanisms redundant.
Companies that do not provide new methods of data and business process shares will not succeed in the market where speed of service is a key customer requirement. These include event-based triggers, data shares, or GenAI-based playbooks to handle more complex environments that are not scalable.
Adding GenAI automation on top will become a massive differentiator for the service providers that consistently deliver great cloud and SaaS experiences, defining the next battleground for customers in the digital ecosystem.
Digital transformation has increased the burden on all engineering and operational teams, forcing enterprises to appreciate the value of process automation. Engineering skills and resources are in short supply. DevOps tools used to manage infrastructure, services, providers and workloads are often isolated, disconnected and lack cohesion when they must work together to solve problems.
Consequently, leaders must view operations cloud, AIOps, playbooks and RPA as an opportunity to define the next chapter of cloud in supporting and automating solutions for the evolving landscape. We are on the cusp of a major evolution in GenAI in using both reactive and predictive solutions to manage cloud connectivity. The potential for the technology to define SaaS is only just being realised.
