Defining use cases

1. What specific problem are we trying to solve?
2. Have we assessed if GenAI is the right tool for the problem?
3. Have we spoken to vendors and conducted research?
4. Have we engaged stakeholders across the organisation?
5. Does the GenAI tool align with our strategy and policies?
6. Can we collaborate with other organisations for shared resources?
7. Have we identified a non-critical area for initial testing?

Data management

1. Is our data high-quality and well-prepared for training AI?
2. Do we have data management systems in place?
3. Are we ensuring that no confidential or personal data is included?
4. How do we know when we are training the model versus when we are not training the model? 
5. Are we compliant with GDPR and other relevant legislation?

Model selection and customisation

1. Have we evaluated pre-trained versus custom models?
2. Have we conducted consequence scanning? 
3. Have we trialled and tested models before committing?
4. Do the contractual terms (T&Cs, IP, data rights) work for us?
5. Have we obtained legal advice on the contractual terms? 
6. Have we considered costs – both upfront and ongoing?
7. Do we have an exit plan in place if the model no longer serves us?

Deployment and integration

1. Have we included staff in our decisions on using this model/tool? 
2. Have we trained staff on effective prompting and ethical AI use?
3. Have we defined what data can be input into the tool?
4. Do we have a process for reviewing AI outputs for accuracy?
5. Are we piloting the tool before full-scale deployment?
6. Have we created an AI policy? 
7. How does the AI tool integrate with our existing systems?
8. Are we transparent with clients about AI use in our services?

Monitoring and risk management

1. Do we have continuous monitoring in place (issues around complacency)?
2. Have we established feedback channels for performance assessment?
3. Are we meeting legal and regulatory compliance?
4. Have we addressed cybersecurity concerns?
5. Do we have insurance coverage for AI-related risks?
6. Is there a business continuity plan in case of failure?
7. Have we implemented responsible AI policies to mitigate bias and misinformation?

Future-proofing

1. Do we have a life cycle management process in place?
2. Is our AI strategy agile enough to adapt to evolving technology?
3. Do we have a plan for upskilling staff or partnering for AI expertise?
4. Are we staying updated on regulatory changes and industry trends?
5. Are we conducting horizon scanning? 

Go back to the main course content