Use Cases for AI
Use Cases and AI Strategy
Section titled “Use Cases and AI Strategy”Source: AI strategy - Cloud Adoption Framework | Microsoft Learn
How to identify AI use cases:
- Automation opportunities: find repetitive tasks, data heavy
operations, tasks with high error rates
- These tasks can be delegated to AI and/or automation to increase productivity
- Ask users how they could be more satisfied
- Assess organization and industry with survey to find challenges and inefficiencies
- Set AI performance targets. A recommended target is the AI solution must produce at least 20% productivity gains. The reason for 20% is account for minor operational variation (+/- 10%) and balance against implementation and maintenance costs, governance, opportunity cost and other management.
Common Use Cases
Section titled “Common Use Cases”Industry Enterprise use cases from Microsoft
Section titled “Industry Enterprise use cases from Microsoft”- Taiwan hospital deploys AI copilots to lighten workloads for doctors,
nurses and
pharmacists
- Data analysis and summaries across many data sources (health records, drug interactions insurance, risk reports, family background, social work reports)
- Report writing (patient care interactions, shift management)
- Personalized instructions for patients
- Tech: Azure OpenAI (GPT model), generative AI, Power BI
- To keep fish and shrimp healthy, farmers in Indonesia now have a copilot to help
- India’s Myntra innovates with generative AI to help shoppers put the right look together
- At The University of Hong Kong, a full embrace of generative AI shakes up academia
- GovTech’s LaunchPad ignites AI innovation across Singapore’s Public Sector
- Amid health-care crunch, AI applications add flexibility and provide
more time for
patients
- Health care person scheduling system that takes into account nurse availability, coverage, and specific changes
- System could reduce time spent on scheduling and make it easier to make scheduling requests and approvals
Personal use cases, Assistant
Section titled “Personal use cases, Assistant”From azure-openai-samples/quickstart at main · Azure/azure-openai-samples · GitHub
Prefer: analysis, adversarial critic, sorting through idea
- Summarize, classify, process text
- Discussion (chat)
- Analysis
- Knowledge Management
- Document retrieval and citing
- Business intelligence & analytics
- Summarizing data
- Legal and compliance
- Writing
- Programming
- Text
- Review and editing
- Report writing
- Integration of processes
- Image, Video, Media analysis
- Content creation
- Accessibility
- Classification
- Business service
- Personalized service
- Data processing
- Calculations
- Financial, cost management
- Training
- Education
- Business support: administration, marketing, communications
- Digital business, commerce
- Security and privacy
- Legal compliance
Personal Use Cases
Section titled “Personal Use Cases”- Check my calendar for next couple free afternoon, evenings and save them
- Open browser to site, open app and perform actions
- Reply with a quick email
- Check messages on platforms
Software Development
Section titled “Software Development”- Support scripting and occasional non-advanced programming:
- E.g. to craft/analyze Kubernetes manifests to host apps on AKS, DevOps pipelines, transcode docker-compose to Kubernetes resources
- Cross language software development
- Completion/comments for code, configuration files, documentation, and data sets
- Common patterns, API, and code option exploration
- Code creation from prompts and changes (ex. test cases, upgrades, version changes), translation