In this episode of Artificially Intelligent Marketing, Martin Broadhurst and Paul Avery reunite to explore how AI has transformed marketing over the past 18 months. They cover reasoning models, agentic automation, Microsoft Copilot’s evolution, open vs closed-source AI, and the rise of AI-powered hardware—sharing real-world insights and examples from their work.

Major Evolutions in AI for Marketers

• Reflection on the rapid progress of AI tools and models
• Overview of major shifts since the last episode
• How marketers are adapting to new AI capabilities

AI Reasoning Models

• Difference between chain-of-thought prompting and modern reasoning models
• Improvements in accuracy and reduced hallucinations
• Trade-offs between speed and reasoning depth
• Groq CEO’s insights on the value of ultra-fast inference

AI Tools Adoption and Platform Maturity

• Microsoft Copilot’s leap from basic to highly capable
• Key tools: Researcher agent, Analyst tool, and Copilot Studio
• Integration across Microsoft 365 (SharePoint, OneDrive, Teams)
• Comparisons with Google and OpenAI’s platforms
• Ongoing confusion over pricing and value

Model Selection: The “Model Roundabout”

• Recent advances in GPT, Claude, Gemini, and open-source models
• Balancing reasoning and instant modes
• Common use cases: coding, summarisation, planning, and copywriting
• Quirks such as GPT-5’s writing tone and output style
• Tips for reducing hallucinations and improving reliability

Open vs Closed Source AI Debate

• Rise and stall of open models like DeepSeek and Llama 4
• Meta’s shift from open development to proprietary AGI
• Open source’s future in experimentation rather than frontier innovation
• Market consolidation, privacy, and trust concerns

AI-Integrated Hardware and the Attention Economy

• Growth of wearable AI, e.g. Meta’s Ray-Ban smart glasses
• Privacy and social implications of constant recording
• Adoption driven by convenience and content habits
• Meta’s competing aims: productivity vs attention monetisation

Agentic Progress: AI Agents and Automation

• “Agentic AI” explained: systems acting autonomously to complete goals
• From document retrieval to full workflow automation
• Tools like Make.com, Zapier, and N8N enabling marketers
• Claude Code as an advanced example of self-directed agents
• Use cases: automated slide decks, proposals, and scheduled reporting

MCP (Model Context Protocol) Connectors

• Overview of MCP for connecting LLMs to CRMs and cloud tools
• Martin’s experience linking Claude to HubSpot and Google Workspace
• Examples of AI updating pipelines and deal notes automatically
• Benefits balanced against setup complexity

Current State of AI for Marketers

• Honest look at AI-generated content and “work slop”
• AI as a speed and productivity enhancer, not a replacement for experts
• Advances in visual and video generation:
  • Faster, more consistent imagery (Midjourney, DALL·E 3, Nano Banana)
  • Real-world use in proposals, events, and social media
• Emerging video models (Veo 3, Sora 2, Kling) offering realism and sound
• Reflection on low-quality AI output and the lasting importance of trusted brands

Intro In this week's episode, Paul and Martin discuss the current limitations of AI for data analysis, Google's AI overview feature, and predictions about AI's impact on jobs. They also cover the latest news and updates from Anthropic, OpenAI, ElevenLabs, and a promising new AI-powered TV show creation tool called Showrunner.

Data Analysis with AI Falls Short

Martin shares his experience using ChatGPT for data analysis, highlighting flaws in extracting accurate insights.

Issues with Google's AI Overview Feature

While widely rolled out, Google's AI overview feature has produced bizarre and inaccurate search summaries.

Anthropic Introduces Function Calling for Claude

Anthropic has made function calling available for its AI assistant Claude, allowing integration with external data sources.

Anthropic Enables Exploration of Claude's Inner Workings

The team at Anthropic mapped out the inner workings of the Claude AI model to better understand its behaviour.

GPT-4 Now Available to All ChatGPT Users

OpenAI has made GPT-4 available to all users of its ChatGPT platform, along with other new features.

ElevenLabs Launches AI-Powered Sound Effects Tool

ElevenLabs introduced a new tool that generates realistic sound effects from text descriptions.

The Simulation Unveils AI-Generated TV Shows

A startup called The Simulation showcased Showrunner, a tool for creating AI-generated TV shows and videos.

Survey Reveals Limited Use of Generative AI Tools

A Reuters Institute survey found that while awareness of AI tools is high, frequent use remains rare.

Apple Rumored to Unveil Siri 2.0 with AI Capabilities

Apple is expected to announce a new version of Siri with enhanced AI capabilities at its upcoming WWDC event.

Anthropic Chief of Staff Predicts Job Obsolescence

An Anthropic executive wrote an article predicting that AI will make most jobs obsolete within the next 5 years.

After an 8-week break, "Artificially Intelligent Marketing with Paul and Martin is back.

Here's a quick recap of recent news: OpenAI's Figure 1 robot now integrates GPT-4, and Devin, a new AI software engineer, was previewed. Claude 3 Opus briefly topped the LLM performance chart, Heygen updated its AI-created user-generated content, and Meta AI released Llama 3. Amazon launched Q for enterprises, and Microsoft is working on the Frontier model LLM named MAI-1. OpenAI's Sora video model produced its first commissioned music video, DeepMind released AlphaFold 3, and AI-enabled hardware like the Humane Pin and Rabbit R1 received poor reviews.

GPT-4o: A Unified Model for Text, Audio, and Video Inputs and Outputs

OpenAI's GPT-4o (Omni) is a versatile new AI model that processes text, audio, and video inputs and generates outputs in multiple formats. This unified approach streamlines workflows and ensures consistency in data processing. Currently, the API supports text and image inputs, with audio capabilities coming soon, making it a comprehensive tool for marketers. Learn more about GPT-4o.

Enhancements to Data Analysis in ChatGPT

ChatGPT has introduced new data analysis improvements, allowing users to upload files directly from Google Drive and Microsoft OneDrive, interact with tables and charts in an expandable view, and customise and download charts. These features, available in the GPT-4o model for ChatGPT Plus, Team, and Enterprise users, enhance efficiency and streamline workflows, making data handling simpler and more engaging.

Google Announces Cutting-Edge AI-Powered Features at I/O 2024

At I/O 2024, Google unveiled new AI features and integrations, focusing on enhanced information organisation and long context management with Gemini. Integrations in Gmail, Google Photos, Google Workspace, and Google Search improve user experiences. Gemini Pro's long context management supports up to 1 million tokens, benefiting research and content analysis. Experimental apps like NotebookLM and AI Studio offer researchers, students, and analysts innovative tools.