The main news yesterday here on LinkedIn was the results of Boston Consulting Group (BCG) testing the impact of Generative AI on its consultants 👩💻
The results were quite overwhelming: consultants using GPT-4 finished 12.2% more tasks, completed tasks 25.1% more quickly, and produced 40% higher quality results.
A few weeks back, McKinsey introduced "Lilli," its latest in-house generative AI tool. Lilli is a chat application that has undergone training using over 100,000 confidential internal documents and interview transcripts. Even in its beta phase, Lilli has evidently proven to be a success within the organization, demonstrating significant impact and widespread adoption.
➡️ The tool has already dramatically cut down the time spent on research and planning from weeks to mere hours, and from hours to just minutes in others
➡️ An astounding 66% of employees now use the app multiple times a week
➡️ In the first 2 weeks of August alone, Lilli answered a whopping 50,000 questions
McKinsey concluded that a combination of cost, reliability, and security justified an in-house solution - despite the higher “tech lift”. McKinsey chose to build its own app on top of other LLM technologies:
🔵 Cohere (most likely using their Embeddings and Semantic Search products); and
🔵 OpenAI Microsoft Azure service (most likely for the “enterprise-friendly” secure GPT-4 instance)
McKinsey's declaration of being "LLM agnostic" is telling. To them, the LLM is a mere means to a desired end: achieving business goals. With Lilli's in-house build, they have the flexibility to change underlying technologies as deemed fit.
In many cases, startups will have to offer similar flexibility in terms of the underlying technology to win over enterprise customers (besides regulation, managing confidential customer data, security, PII requirements, etc.) - something that will likely require more engineering efforts and will involve continuous testing, QA, etc over multiple evolving LLMs available on the market. Something that could be overwhelming for smaller startups with smaller technical teams.
☝ However, the McKinsey's and BCG's of the world have one big advantage: data and knowledge that they've gathered from various projects for decades that is probably stored and quite well-indexed.
Data, documentation and knowledge bases will be the key moat and competitive advantage in this new era where models will be increasingly commoditized. Knowledge bases are as important to AI progress as Foundation Models and LLMs.
To compete, make sure your documentation and knowledge bases are the best on the planet. When it comes to knowledge you want to be able to store a lot of it, and you want to be able to find the right piece of knowledge at the right time. For e.g. LLMs this is typically done with a vector database (for now).
#largelanguagemodels #gpt4 #productivity