UK Startup & Scaleup Week

This summary captures the key insights from the conversation between Aidan Dunphy (Co-Founder, Frntir.ai) & Paul Lancaster (Founder, UK Startup & Scaleup Week | PLATFORM) about the future of AI in business, specifically focusing on Agentic AI & Synths.

Moving Beyond the AI Hype: Business First, Tech Second

Aidan Dunphy & Cyril Le Roux Co-Founded Frntir.ai with a ‘business first, tech second’ approach. Their goal is to move past the industry jargon & hype to focus on delivering quantifiable outcomes for businesses using AI, particularly those in the 50 to 500 employee range.

The fundamental problem they address is the challenge of institutional knowledge – valuable information buried in documents, emails, CRMs, or, worse, locked inside people’s heads.

Introducing the ‘Synth’: AI as a Colleague

Frntir.ai uses the term ‘Synths’ to describe their AI agents. Unlike traditional agents that simply follow instructions, Synths are designed to be more like semi-autonomous human colleagues.

• Proactive & Persona-Driven: Synths are given a job description & a persona (aligned with the company’s culture), allowing them to act proactively rather than just waiting for a prompt.
• Human-Paced Interaction: The AI is built to interface with humans using understandable language & behaviour, attending meetings & learning the business – fitting around the human world, not the other way around.
• Confidentiality Baked In: They are built to manage different kinds of memory (episodic, persistent) & respect confidentiality & sensitivity, such as internal information & salary data.

Agentic AI vs. Traditional LLMs

Aidan highlighted the crucial difference between large language models (LLMs) & true agentic AI:

Type	Function	Key Capability
LLM/Chatbot	Answers questions based on known data (like a “super Google search”).	Reasoning based on training data.
Agent (Traditional)	Executes instructions using tools.	Tools, Memory, Carrying out instructions.
Synth (Frntir.ai)	Works a job based on a persona & business needs.	Proactivity, Initiative (e.g. reaching out to a colleague for missing info), Cultural Fit.

Real-World Value & Practical Applications

Aidan says that Synths are designed to automate “soul-crushingly boring” but necessary work, delivering immediate business value:

• Solving the Knowledge Gap: For example, with a MedTech company that Frntir.ai is working with, a Synth can search through internal emails, CRM data & documents to answer a question like, “Have we ever formulated this before?” in seconds, a task that might have previously taken days.
• Automating Complex Pricing: Extracting constantly changing cost information from supplier PDFs & emails to automate complex job quotes that currently take days to process manually.
• Capturing Unwritten Knowledge: Instead of having staff sit down to write lengthy documents, the AI can attend a meeting or interview a human to gain knowledge & generate the necessary documentation itself.

Ultimately, Agentic AI changes the nature of knowledge work: businesses will shift from paying people for the knowledge they store to paying them for who they are and what they can do – leveraging their unique human skills while the AI manages the institutional memory.

Learn more about how Aidan & Cyril (Frntir.ai) can potentially help your business & download a free ‘Practical Guide to Agentic AI’ at https://frntir.ai

ps You can also listen to this interview (& more) on Paul’s ‘SuperConnector Show’ podcast which is available on Apple Podcasts, Spotify, Amazon Music, Audible or Spreaker via our ‘Podcasts‘ page.