Introduction

As someone who has been working in enterprise technology for nearly two decades, specifically in large-scale transformations and implementation programs, I’ve witnessed firsthand the evolution of customer experience solutions. Recently, I’ve been struck by the immense potential of generative AI in addressing industry-specific challenges. At Pale Blue Consulting, we’ve assembled a small team to explore this potential through a series of demonstrations and prototypes.

Our Journey into Generative AI

Our foray into generative AI was inspired by the impressive capabilities of models like GPT-4, which I first experienced in March 2023. Convinced that these tools will become a core element in future solutions, we embarked on a journey to gain hands-on experience and explore their potential in various industry contexts.

Our Prototypes

To date, we’ve developed three main prototypes, each targeting a specific industry problem:

  1. Hosaka: A content personalization tool for publishers that generates tailored podcasts, newsletters, and web content from existing material. Built using GPT-4, Pinecone, Langchain/LlamaIndex, Elevenlabs, and HTPlay.

  2. Halcyon: A lead enrichment solution that validates and enhances prospect data using minimal input, aiming to provide a cost-effective way to improve lead quality. Developed using Claude 3 Haiku, Supabase, Langchain/LlamaIndex, Exa, and various API providers.

  3. Aether: An automated note-taking tool that transcribes, summarizes, and stores information from various sources, making it searchable and easily accessible. Built with Claude 3 Haiku, Whisper, Unstructured.io, and Supabase.

  4. Expenses: An internal tool for taking unstructured expense data and bank statements and performing a monthly reconciliation. Built with Haiku, Langchain/LlamaIndex, and Google Sheets.

  5. Luminarium: A prototype for an events interview tool which takes exhibitor input along with web and document inputs then interviews the exhibitor to fill out their profile - this also can then be used to provide a way for attendees to find exhibitors to connect with. Built with Haiku, Langchain/LlamaIndex and Vapi.

Challenges and Lessons Learned

Developing these prototypes has been an exciting journey, but not without its challenges. The rapid evolution of generative AI means that capabilities are constantly increasing, which is both an opportunity and a challenge. We’ve learned that it’s crucial to focus on areas where solutions can be improved and enhanced by advancements in models and capabilities, rather than being replaced entirely.

Future Plans

We’re currently working on a more fully-featured prototype of our interview product, leveraging the latest capabilities, as well as a more complete tool-using AI assistant. Stay tuned for updates on these exciting projects!

Collaboration and Engagement

We’re eager to connect with like-minded professionals and discuss the potential of generative AI. Follow our blog for future posts, and engage with us on LinkedIn and X. We’re always open to collaboration opportunities and would love to hear your thoughts and experiences in this rapidly evolving field.

Join us as we explore the practical applications of generative AI and its potential to streamline processes and enhance customer experiences across various industries.

#GenerativeAI #IndustrySpecificSolutions #Prototypes #CustomerExperience #EnterpriseTechnology