Welcome to the FIRST episode of The Break to Build podcast 🎉
I’m really excited to get this out in the world, but more importantly, hear from YOU.
✍️ Is this information useful?
🚀Can you leverage it to level up productivity, intelligence, and GTM speed within your marketing or sales team?
🙋🏽♀️Yes, or no, I’d love to hear from you. Feel free to send me a DM on LinkedIn.
Let’s get into it.
Who is Andy Roy?
Andy Roy is a digital transformation expert and the CEO of Improvize, an AI consultancy focused on backend and CX-enhancing AI solutions. He has designed experiences for large brands like Kraft Foods, US Bank, and Cheesecake Factory while working for leading enterprises like Razorfish (Publicis), Microsoft, Experian, and TransUnion.
He has built natural language interfaces since 2016, including sentiment analysis, unsupervised learning, and multiple conversational AI chatbots. He works intensively with large language models, including Anthropic’s Claude and OpenAI’s ChatGPT, to build data-driven digital agents (e.g., a Salesforce CRM advisor) and automated document processing (e.g., processing loan applications).
You can find Andy via his LinkedIn and check out what he and his team do on the Improvize homepage.
Who This Conversation Is For
Heads of Digital Marketing
Sales and Marketing Leaders
Product leads (especially those looking to assess open-source vs closed LLMs)
Strategic and tactical GTM managers/planners
Heads of Customer Experience
Digital Marketing Managers
Anyone involved in creating new sources of premium first and zero-party data for their brand.
Top Insights
Generative AI and language models are experiencing growth comparable to the iPhone or the internet, with profound implications on the ‘new marketing funnel’ powered AI search.
Conversational AI transcends traditional rule-based chatbot utilities, evolving into digital agents that bear knowledge and take action based on sophisticated cloud-based services.
The deprecation of third-party cookies emphasizes the importance of first and zero-party data collection. It makes a strong case for conversational AI as a source of rich insights into customer intent.
Despite their potential, large language models pose risks such as 'hallucinations' and misinformation, necessitating reliable sources and responsible deployment.
What are the main differences between opting for a closed LLM and an open-source one?
What We Cover In This Episode
0:03:21: The overwhelming speed of AI advancements leads to confusion and fear among enterprise leaders today.
0:07:52: Language models revolutionize the marketing funnel and content creation.
0:11:23: Digital agents combine knowledge and action for tailored interactions.
0:13:31: Digital agents use code to access cloud services and automate work.
0:17:56: Enterprise leaders should focus on what AI enables, not the specific language model.
0:21:13: Conversational interfaces provide an opportunity to gather first-party and zero-party data.
0:22:04: The four steps to building an AI chatbot: topics, knowledge, conversation, and channels.
0:26:59: Use cases in healthcare can include clinical decision-making and administrative support.
0:27:30: AI applications in healthcare: note-taking, scheduling, patient education.
0:28:13: Concerns about AI hallucinations and mitigating against irresponsible AI
0:32:19: Open-source LLMs vs Closed LLMs
0:37:53: Leveraging generative AI for GTM personalization and shortened sales cycles.
0:40:13: The value of zero-party data and emotional connection with customers.
0:42:27: Preview of next episode on cookie-less advertising.
What You Can Do With These Insights
Present your marketing/sales/product team with ideas to build conversational AI chatbots that can present the below outcomes:
"GuideBot," interacts with new users during the onboarding process on our platform. By asking targeted questions about the user’s business size, industry, and specific challenges, GuideBot collects valuable first-party data.
"FeedbackFriend," a chatbot designed to collect user feedback on new features directly within our app. It engages users in a conversational manner, asking for their thoughts on usability and any additional features they desire, fostering a sense of co-creation with your customer base.
By analyzing the data collected from your chatbots, you can develop predictive models that identify customers needing additional support or at risk of churning.
Start a conversation with your Head of Data Privacy about what type of LLM and security governance will be used if AI change management is ramping up.
Discuss with your Head of Digital how Conversational AI can be used to collect premium first-party and zero-party data at scale to prepare for a cookieless advertising future.
Additional context and ideas here from Andy Roy and the team at Thought Bakery: https://thought-bakery.com/conversational-ai/
References/Resources
https://blog.hubspot.com/sales/conversational-ai-for-sales - HubSpot’s top 10 conversational AI platforms for sales
https://zapier.com/blog/custom-chatgpt/ - Learn how to build a custom GPT to start using conversational AI within your team.
https://sites.research.google/med-palm/ - Learn more about Google’s medical LLM with updated case studies provided
https://www.cbinsights.com/research/open-source-closed-source-ai/ - The generative AI divide: Open-source vs. closed-source LLMs
https://www.ibm.com/blog/how-ai-powered-chatbots-are-transforming-marketing-and-sales-operations/ - Don’t take our word for it. Check out how IBM’s Watson is championing sales and marketing using Conversational AI.
More about Chae O’Brien
I’m not interested in contributing to the digital noise our feeds get flooded with daily.
If there are topics/trends you’d like me to cover or other areas of these conversations you’d like to dive deeper into, drop me a message via one of the below. I’m a research junkie, and nothing is more fun than finding answers to the doubted or unknown ✌🏼
Website: https://www.chaeobrien.com/
LinkedIn: https://www.linkedin.com/in/chaenara/
Email: hello@chaeobrien.com
Share this post