Intelligent Automation is the orchestrated use of AI technologies to automate IT processes.
This is different from the traditional rules-based automation that automates processes based on fixed decision parameters and metrics thresholds.
How intelligent automation works
Unlike other forms of automation, intelligent automation introduces cognitive decision-making capabilities and contextual knowledge to automation controls. These controls have several downstream applications ranging from IT operations (ITOps) to cybersecurity.
Intelligent Automation for enterprise IT infrastructure operations and business processes improves IT performance by offering the following key capabilities:
- Supports real-time decision-making process with cognitive intelligence.
- Augments the role of ITOps and IT service management (ITSM) by intelligently automating processes and capabilities such as conversational AI.
- Redistributes knowledge of complex tasks across business functions and IT operations that would otherwise remain in silos. This knowledge is consumable and actionable by all users, irrespective of their IT expertise.
Intelligent automation vs. traditional IT automation
So how is intelligent automation different from traditional IT automation?
A key difference between Intelligent Automation and traditional IT automation relates to the modeling approach.
Traditional automation
Traditional automation defines the behavior of the IT environment, systems and operational workflows based on a fixed set of rules. When a system query or alert is triggered, the rule-based automation engine traverses relevant logical options and converges to a solution that best fits the described parameters.
A simple example of an automation system may rely on classical ML mechanisms such as a rules-based expert system or decision trees. Intelligent Automation is different in the sense that it relies on advanced ML technologies such as deep learning algorithms that are trained to emulate the true system behavior using real system data and parameters.
Intelligent automation
Now consider a case example for the comparison from incident management. A rules-based system may be designed to trigger alert and an automated control action such as deprovisioning of Web servers when a network traffic sensor log exceeds the predefined magnitude threshold.
This may be the case to protect against a DDoS attack, but the server downtime comes with its own opportunity cost — such as performance degradation of the dependent services.
It may be the case that an external event (a marketing moment, a social post, etc.) may have caused an excessive but organic rise in network traffic. This is later verified by analyzing contextual information such as the network-wide traffic patterns.
An Intelligent Automation system does just that: it analyzes large volumes of traffic in real-time to acquire contextual knowledge before executing a decision control.
Advanced ML algorithms emulate cognitive intelligence in the decision-making process. Large Language Models (LLMs) similar to ChatGPT can engage in realistic and human-like conversations for front-line ITSM support. Conversational AI enhances the end-user and customer experience, which is valuable for your digital transformation efforts.
Drivers of intelligent automation
Intelligent Automation is driven by real-time information streams. Process mining is used to analyze how teams and systems operate. It helps identify business processes and the real-time constraints applicable to them.
While traditional automation takes a predefined process workflow and automates tasks for the end-user, Intelligent Automation discovers the gap between the designed workflow and its execution. It uses these insights to identify:
- Manual processes that can be automated.
- How bottlenecks are introduced with the existing automation workflows.
It unlocks the value of data to enable real-time decision making. This data is usually available across multiple domains and unstructured formats.
Traditional robotic process automation (RPA), along with process mining, serves as a transport mechanism to extract useful information in its raw format. It requires additional layers of processing:
- Data transformation
- Security
- Storage
- Analysis
An end-to-end and centralized data platform is required to unify the decision making process from siloed information sources and business functions. (This is also an important component of a self-service ITSM initiative and a unified knowledge base for the end-user.)
A conversational interface in an Intelligent Automation system is only useful when the provided guidance and support is accurate and efficient. Speed to insight is key to large-scale adoption of Intelligent Automation. It is also one of the key objectives for organizations seeking productivity improvement by embedding AI into IT automation.
Standardize complexity
These technologies may be complex but Intelligent Automation delivers high value at scale and speed by standardizing, simplifying and optimizing processes as a pretext to automation. Simply automating waste processes introduces new process bottlenecks, risk of downtime as well as a high learning curve for new users who may not be accustomed to your operational workflows.
Like any technology-led transformation initiative, Intelligent Automation also introduces significant change and a strong governance process, training routine and executive support. In order to mitigate these risks, start with a proof of concept and set the right expectations for your transformation journey.
Intelligent automation should have a robust solution focus: use AI capabilities to replace human ITSM agents and ITOps managers. Choose which processes are automated first: start with a business value centric model. Plan for sustainable growth and scalability: automated workflows should not make processes and operations more complex.
