Generative AI Rapid Prototype in Days

This offering falls into step 2 (Concept Development) of the ipt GenAI process model . With this offering, we enable organizations to quickly implement a Gen AI Chatbot based on the latest large language models (LLMs) hosted in Amazon Bedrock. Use case specific Context can be integrated by feeding relevant Documents to Amazon Bedrock Knowledgebases forming a retrieval augmented generation (RAG) solution. We re-use the open source LLM App Genie framework with infrastructure as code (IaC) to rapidly (normally in 3-8 days) deliver a PoC style solution with a simple frontend. Because our solution is based on standard AWS services we can extend it quickly into a production ready solution.

Reference Architecture Diagram 

Re-Using our reference architecture allows us to quickly customize and deploy the solution.

This offer is suitable for : *Customers who want to engage with LLMs to gather first experiences and validate ideas with experiments before investing in large projects. *We focus on organizations who want to enable their employees or customers to benefit from Generative AI. *Organizations who have a dataset they want to make accessible through a natural language interface, e.g., a product catalog, or legal texts around drone usage.

Customer engagement and delivery mechanism example: This is a consulting offer by Innovation Process Technology AG. Once we have identified client needs in the first or second client call, we issue a Statement of Work (SoW) with a timeline. This SoW will be submitted to client (for approval of the overall approach) and to AWS (e.g. to the AWS APN funding portal)

This offer from Innovation Process Technology AG will be delivered using an agile mindset including fast and iterative feedback. This enforces the customer (you) to play an active role in the project. A typical project plan includes::

1. First contact and negotiation which experts (AI Experts, Data Science Experts, AWS Experts) have to be included in the PoC
2. Get to know each other, talk about the challenges and the goals with the PoC.
3. Present a project timeline with milestones, customer involvement, prerequisites, budget, backlog
4. Install and configure the AWS Services in your landing zone, the customer brings the data
5. First iteration with milestone - a first use case including a frontend prototype
6. Demo time and feedback gathering. Backlog review. Discuss and fix project timeline and adaptions
7. Final Workshop with Demo and showing a Backlog to bring the Proof of Concept into production.