Episode notes
Does every business need AI agents, or could a simpler workflow do the job better?
In this episode of The AI Lyceum™, Samraj Matharu speaks with Dr Quentin Reul about AI agents, knowledge graphs and what it takes to turn enterprise AI experiments into useful business outcomes.
Quentin explains why predictable tasks may not need autonomous agents, while deep research can benefit from systems that explore sources and adapt their approach. The conversation considers how to choose the right technology, assess its cost and measure whether it improves the problem you set out to solve.
We explore AI hallucinations, critical thinking, explainability, smaller language models and proprietary business knowledge. Quentin illustrates why context matters through two companies with the same name: without distinguishing the businesses, an AI system can produce a convincing answer about the wrong organisation.
We also discuss how knowledge graphs can help validate information extracted by language models, why AI answers need checking, and how human review, integration and maintenance affect the real cost of adoption.
“Fall in love with the problem, not the solution.” Dr Quentin Reul
EPISODE HIGHLIGHTS
0:00 ➤ Recording Setup and Welcome
1:21 ➤ Introducing Dr Quentin Reul
2:37 ➤ Entities, Names and AI Hallucinations
5:08 ➤ How Language Models Generate Answers
8:37 ➤ Human Learning and AI
12:04 ➤ Do We Need AI Agents?
13:40 ➤ Deep Research and Agentic Workflows
17:03 ➤ Critical Thinking and Checking Sources
20:46 ➤ Tokens and the Limits of AI Reasoning
24:10 ➤ Knowledge Graphs and Validation
27:02 ➤ Can AI Be Explainable?
32:07 ➤ Additive and Subtractive Fine-Tuning
35:39 ➤ Synthetic Data and Model Costs
39:30 ➤ Chips, Infrastructure and AI Economics
45:37 ➤ Smaller Models and Business Context
46:58 ➤ The Hidden Costs of Operating AI
49:01 ➤ GPUs and the Infrastructure Behind AI
54:00 ➤ AI Agents and the Semantic Web
55:30 ➤ Defining the Business Problem
56:48 ➤ Measuring AI Value and ROI
59:34 ➤ Innovation and Practical Adoption
ABOUT DR QUENTIN REUL
Dr Quentin Reul is an AI strategist focused on knowledge graphs, generative AI, agentic workflows and responsible adoption. His work connects technical implementation with business needs, emphasising relevant context, validated outputs and measurable value.
ABOUT THE AI LYCEUM™
The AI Lyceum™ is an independent community examining artificial intelligence through philosophy, science and public discussion.
Hosted by Samraj Matharu | Certified AI Ethicist (University of Oxford) | Visiting Lecturer (Durham)
YouTube: https://www.youtube.com/channel/UCzZOR1oz-h8QglWZUCi-HhQ
Spotify: https://open.spotify.com/show/034vux8EWzb9M5Gn6QDMza
Apple Podcasts: https://podcasts.apple.com/us/podcast/the-ai-lyceum/id1837737167 Amazon Music: https://music.amazon.com/podcasts/5a67f821-89f8-4b95-b873-2933ab977cd3/the-ai-lyceum
Website: https://theailyceum.com
Personal website: https://samrajmatharu.com