GenAI adoption in the Enterprise
As an inception and early stage investor, we have found that one of the key avenues for start-ups to accelerate their journeys is to get market feedback from actual buyers.?In turn, innovative CXOs appreciate hearing from startups.?This give-get equation is reflected in our network of over 1000 lean-forward CXOs who maintain a vibrant dialogue with founders.
Last week, we hosted a conversation with Google Global Generative AI leader @Josh Gwyther to share their view of GenAI adoption in the enterprise.?Key takeaways included:
Founders, I would love to hear from you if these takeaways resonate with your own experience.
Navin Chaddha Thanks for Sharing! ?
Top Voice | Training Datasets and workflows for AI Agents
1 年It's interesting to understand what part in Data corpuses will become the moat.
CEO neurIOT Labs
1 年All valid points Navin Chaddha. On top of findout, our LLM-based solution, we've done one full deployment at one of the US' largest health insurers. In addition, we're in the process of deploying findout at 3 more customers. One key point for all of them in the enterprise setting is "zero tolerance to hallucination". They don't want static / pre-trained LLM models to throw results out of their own memory. Any false result (that's not derived from their enterprise data) can lead to wrong decisions and is therefore a big no. And so, guardrails are crucial but they also want creative contextual responses. That means going up the ladder with LLMs through SFT, RLHF, or training LLMs on their data from scratch (all expensive propositions).
Co-Founder
1 年IronCore Labs is helping unblock the Data Security/IP blockage. https://ironcorelabs.com/blog/2023/announcing-encrypted-ai-embeddings/
Repeat Founder & Investor | Advising Startups & Emerging VCs Globally with Fundraising, GTM, and M&A
1 年Existing data complexity across the enterprise, and ongoing governance post-implementation are two other big concerns. Implementing is one thing, ensuring proper use is something else, as those selling into the enterprise already know. With such a big (top-down) push for GenAI adoption in many bigCos, all these concerns are magnified.