The Conversational AI Podcast

So, today I got access to ChatGPT's Agent feature and I wanted to record something for posterity. This one is best viewed on Spotify or YouTube if you'd like to see the live screenshare as I'm talking, otherwise I do my best to describe things as we go.
My chosen task for Agent was this: Login to my LinkedIn, accept some recent connection requests and say "Thank you for the add, [Firstname]" to each connection.
Simple, I know, but I wanted to keep things simple and easy. It handled things really, really well -- so much so, that I felt myself being 'upgraded' (or wanting to upgrade) to the $200 ChatGPT subscription as I watched it work. See what you make of it!

In this episode, I'm talking through a case study about how NatWest Bank in the United Kingdom is using Generative AI to analyse the output of their Conversational AI (chatbot) system, Cora.
The source is a post by Wendy Redshaw on LinkedIn. Wendy is the bank's Chief Digital & Information Officer. You can find her post right here.
Takeaways
I'm a fan of NatWest and follows their executive output.
NatWest is using generative AI in a careful and controlled manner.
Generative AI adds value primarily through conversation.
Cora, NatWest's chatbot, is a key use case for generative AI.
7,000 conversations have been summarized using generative AI.
Human review is essential for evaluating AI performance.
The average rating for AI-generated summaries is 4.95 out of 5.
Understanding customer complaints is crucial for financial institutions.
NatWest uses the Lama 3.3 model from Meta through Watson X.
Chapters
00:00 Introduction to NatWest and Generative AI
02:51 Exploring NatWest's Chatbot Cora and Its Impact
05:44 Evaluating AI Performance and Customer Insights
09:08 Technology Behind NatWest's AI Initiatives

I'm taking a little bit of a departure in this episode because I'm talking about some of my activities today to create an app that I wanted to use for myself. And I did this using Claude across the space of about two hours -- and although I've been doing this very regularly, I can't help but be incredibly impressed and enormously reflective on what the power and capability of these tools means for commerce, enterprise and society. So this is a Stream of Consciousness episode.
And in case you were wondering, I used my own little piece of software to dictate this text. Watch or listen to the episode to find out more.

In this episode I had the pleasure of interviewing Paul Walker, Global Solutions Director at iManage. With 25 years of experience in legal tech, Paul offered valuable insights into how AI has evolved in the legal and professional services sectors.
We discussed how iManage specialises in knowledge work industries, particularly legal and accounting, and explored the revolutionary impact of generative AI. Paul explained how AI has transformed the way organisations manage unstructured data, significantly reducing the lead time for extracting answers from this information.
A key point Paul highlighted was how AI now provides crucial context to unstructured data, substantially enhancing its value. Throughout our conversation, he emphasised that curiosity is essential for effectively adopting and leveraging AI technologies.
Looking to the future, we discussed how AI interfaces might evolve beyond traditional prompts toward more intuitive interactions.
One of the most practical takeaways was how AI can boost productivity by automating laborious tasks, with Paul predicting that the integration of AI into workflows will become increasingly seamless over time.
Check out Paul's LinkedIn profile and find out more about iManage at www.imanage.com.

In this episode, I'm looking at a very, very simple query I put to Google's Gemini regarding my email inbox. Oh dear.

Hold the Agentic AI Revolution for a few minutes.