What is the right platform\cloud services with of the shelf tools can be used to create enterprise GenAI SaaS in organization with multiple companies and subsidiaries? In the target SaaS model, multiple APIs should be available to consume external LLM services, front end, prompt library that can be used by the end user, predefined data sources, Guardials and security controls.
Sort by:
CIO in Banking6 months ago
Azure has a OpenAI Studio and AI Foundry that make it simple to spin up private access to all the major LLMs and pay based on usage. If you want to build a specific front end to store a prompt library and other things, you may need to build the web front end, but they make it straight forward to do so.
Director of IT in Insurance (except health)6 months ago
Hyperscalers provide services is maturing fast. Primary cloud AWS/Azure/GCP service could be a safe choice.
My personal thoughts as Industry Leader for Development of an enterprise Generative AI/GenAI says platform for a multi company or subsidiaries requires a full grown cloud platform that offers a various out of the box tools and services. These address scalability, availability , compliance and platform inter interoperability.
Some of the recommended Cloud Platform are :
AWS - Amazon webservices is rich in AI/ML functionality, AWS is aggressively moving in AI field to produce efficient and speedy application development.
Key Services by AWS are :
AWS has capability of AWS bedrock for GenAI integration.Bedrock has capability of interacting with different LLM models like Anthropic, Cohere.
AWS Sagemaker can be used to create model of custom AI/ML.
AWS Lambda is used for serverless function development.
Amazon Cognito used for user authentication and security.
Amazon GaurdDuty can be used for real time AI based threat intelligence.
Azure - Azure OpenAI service can be used in building seamless integrations with LLM’s like ChatGPT. Azure provide strong support for building API’s, frontend applications and data source management.
Key Services by Azure are :
Azure OpenAI for access to a variety of models developed by OpenAI
Azure Functions for building functions as service.
Azure Synapse Analytics is used for large scale enterprise data handling.
Azure Key Vault to perform secure storage of keys and credentials.
Hybrid Cloud - Platform such has Redhat Openshift or VMVare Tanzu provide ability to have hybrid cloud strategy . This will enable organizations to have more control over on-prem infrastructure also providing flavor of Cloud Native development. Although Hybrid cloud has some limitations compare to public cloud, its upto the organizations to choose Public Cloud or Hybrid Cloud strategies.
Additional Recommendations could be :
• Multitenancy: Azure Kubernetes Service, Amazon EKS, or GKE shall be used for managing multitenant applications.
• Observability: APM tooling should be placed-Dynatrace and Datadog, or the native monitoring services-can be used to track performance.
• Cost Optimization: Use pricing calculators from each provider to design an affordable solution. Sometimes this is the crucial factor in choosing the platform.
• Developer Ecosystem: Enable teams with pre-built SDKs, CLI tools, and cloud-native IDEs-for example, VS Code with cloud extensions.
The right platform selection is matter of the existing infrastructure in the organization, skillset of technologists and regulatory requirements. Based on above criteria Organization should select their Enterprise GenAI SaaS model.