Implementing any IT project requires time, planning, and effort and AIOps probably requires even more planning and stakeholder involvement, because of the breadth of coverage and potential to bring profits to multiple IT domains/functions (ex: ITOps/ITSM/NOCOps). Customers have high expectations from AIOps, but, even after taking such major projects, most AIOps vendors are only able to support a few core AIOps use cases, which severely limits the utility and potential of AIOps.
Core AIOps Use Cases
Following are some core and common AIOps use cases:
- Alert noise reduction using event correlation
- Incident MTTR reduction
- Root cause inferencing
- Anomaly detection
- Predictive analytics (forecasting)
- Trend and baseline establishment
Definitely, these use cases are core to AIOPs and have proven to bring benefits to many customers.
How can you go beyond these core AIOps use cases? How can you maximize the potential and ROI of your AIOps?
Expanded AIOps Use Cases
This is where Robotic Data Automation Fabric (RDAF) comes into the picture. RDAF makes AIOps more open and extensible and allows you to work with data even more efficiently, allowing you to implement a lot more use cases.
Following are few expanded AIOps use cases that RDA enables:
- Change detection for EC2 VMs, App connections, etc.
- Log Clustering: Group on-prem and cloud logs and visualize in Kibana
- CMDB Synchronization: Keep CMDB up to date using AIOps inventory
- E-Bonding: Synchronize incidents b/w two ITSM systems (Ex: Jira<->ServiceNow)
- NLP Text Analytics: Incident text insights (concepts, categories, keywords, summarization)
- Incident Sentiment Analysis: Polarity (degree or positivity of negativity, subjectivity, identification of sentences that are mostly negative or positive)
- Named Entity Recognition: Like Regions or Country or City names, people names, Dates/Times, Language names. Landmark names, money related numbers
- Incident Assignment Group Prediction: Using supervised classification
- External AI service integration: Integration with OpenAI GPT-3, IBM Watson NLU, etc.
- In-place incident enrichment: In ITSM ticket console with NLP insights and recommendations
- Container security and vulnerability analysis and change detection, reporting using external security tool integrations (Ex: Cloud Defense)
- Software error message contextualization using Open AI
- Incident enrichment – traceability with relevant logs, traces, build changes attached to the incident
- Incident enrichment – supportability with relevant communication trails of support tickets with vendors, field notices, CVEs, PSIRTs, etc.
- Knowledge mining and articles suggestions based on supervised learning and NLP insights
All the above use cases can be easily implemented using no-code RDAF pipelines in AIOps Studio, which is like a visual designer or IDE to implement RDAF. The no-code and inline bots approach makes a broad appeal to a lot of practitioners and field teams without requiring any kind of specialized programming or domain skillset.
By the time I write the next blog, we would have probably come across few more use cases and scenarios, which we will keep sharing frequently. The real promise of AIOps 2.0 can only be achieved with an AIOps platform that is open, extensible and has data integration and preparation capabilities.
To learn more about RDAF or to try it out, visit: https://www.cloudfabrix.com/signup/