ReginaGPT

• 

  ReginaGPT in a modern corporate environment, managing governance, risk, and compliance

Inspiration

Creating policies and standards can be cumbersome, and finding guidance on the approach creates a whole different level of stress. Therefore why not automate this process and use the tool as a guide to create a well tuned document that ensures compliance at every level.

What it does

This app will generate a policy or standard information based on industry standards using known standards such as ITIL, COBIT, NIST, FFIEC, GLBA, and GDPR framework.

How we built it

This app will built exclusively with AWS PartyRocks using the Claude Foundational Model.

Challenges we ran into

Here are some potential challenges a user may face when using this cloud governance policy generation app:

• Complexity of inputs - The app requires providing quite a bit of contextual information like industry, policy type, applicable standards/frameworks, etc. This level of input could be difficult for some users.

• Lack of customization - The generated policy/standard is based on predefined templates and may not fully address an organization's unique needs or preferences in terms of structure, language, etc. Customization options could be limited.

• Understanding outputs - The generated documents may use technical or legal jargon that is difficult for non-experts to understand fully. More plain language and explanations could help.

• Keeping inputs up-to-date - As an organization, industry and regulatory landscape changes over time, the contextual inputs provided to the app may become outdated. Users would need to periodically review and update inputs.

• Integration challenges - Getting the generated policy/standard integrated into the actual governance program could require additional manual effort for things like approvals, distribution, enforcement, etc.

• Reliance on automation - Some users may hesitate to fully rely on an automated process and prefer more human involvement in policy development. The app's outputs would need to be thoroughly reviewed.

• Limited support - As an AI/software tool, the level of support provided to users may be more limited than working directly with governance experts. Clarifying help/instructions is important.

• Output quality - Early versions may have issues/bugs that impact the quality, completeness, or compliance of generated outputs that must be addressed.

Accomplishments that we're proud of

Accessible to All - The software-as-a-service model means anyone can access the app's capabilities, not just those with in-house expertise or budget

On a personal note, anytime I build an application. It feels good to see the final product.

What we learned

What's next for ReginaGPT

Would love to migrate this to Bedrocks, build an application using different models, and train the model using different data points.

Built With

• claud
• partyrock