Read In Browser

Host Tedd Huff the Founder and CEO of Fintech Advisory firm Voalyre sits down with Duncan Barrigan, Founder and CEO of Lunos.ai, for a candid discussion about the biggest blind spot in business payments. They explore what happens when someone leaves a $2.2 billion fintech company to solve a problem most people avoid talking about: the messy, chaotic reality of collecting money from other businesses.

Duncan spent years as Chief Growth Officer at GoCardless, helping scale the company past $100 million in annual recurring revenue. He worked daily on payment processing technology, building solutions to move money from bank accounts. When you spend that much time holding one tool, everything starts looking like the same problem. Walking away from that role changed his entire perspective on what was actually broken.

The problem isn't payments. Companies can process transactions fine when both sides agree on what's owed. The real breakdown happens in the weeks and months of communication chaos that come before payment. Disputed invoices, conflicting records, missing information, negotiated terms that never got documented properly. This is where businesses lose millions in operational overhead and damaged relationships.

Tedd shares his own nightmare scenario. A client insisted they'd paid an invoice that still showed as outstanding. What should have been a quick clarification turned into hours of downloading three months of bank statements, cross-referencing every transaction, comparing records between systems that didn't match, and spending 90 minutes on the phone walking through everything line by line. The worst part: he made it worse by exporting the wrong data from his accounting system, adding even more confusion to an already chaotic situation.

These failures cost real money. Major rideshare companies employ teams of 50 or more people just to handle accounts receivable operations. Most businesses operate with spreadsheets they download from their ERP once a week, trying to track which invoices are paid and which need follow-up. When records go wrong, simple questions become complex disputes. Cash flow problems cause business failures even when customers love the product and want the company to succeed.

Duncan's realization came from stepping back and asking why 50-person AR teams exist in the first place. He'd been looking for answers he wanted rather than understanding what was actually happening. The answer wasn't another payment processing solution or another workflow automation tool. The answer was recognizing this as a communication and negotiation crisis that's persisted for thousands of years, from clay tablets in Mesopotamia to email today.

Lunos builds AI agents that act like human workers but without the limitations. They never get tired, never forget details, and scale infinitely. The company's first beta customer provided a critical lesson about real-world complexity. An engineering team spent 10 minutes trying to interpret a single customer email that referenced multiple invoices with conflicting information and invoice numbers that didn't exist. That moment forced them to embrace chaos rather than trying to eliminate it through rigid rules.

"If you don't embrace the chaos, you're not gonna be able to handle it," Duncan explains. People send messy emails. They reference wrong invoice numbers. They forget details. The AI needed to handle all of that just like a competent human employee would, not how engineers wish humans would communicate.

The system operates through three modes. Monitor provides visibility without automation. Suggest makes Lunos responsible for generating responses while requiring human approval. Act represents full automation where the AI handles communications and only escalates when necessary. One customer saw 25% of their entire accounts receivable balance, nearly a million dollars, paid within two days of switching to act mode. The AI wasn't doing anything magical. It was doing what needed to be done at the scale and consistency that manual processes can never achieve.

Lunos takes an email-first approach because that's already how most B2B payment discussions happen. Rather than forcing businesses onto new portals or workflows, the AI works within existing communication channels. When a customer responded in German, the system automatically understood the message, crafted a reply in German for the customer, and provided an English translation for the business owner. Nobody programmed this capability. It emerged from how the AI was designed to handle human communication.

Under: Transform Your Merchant Applications with Under. The Under platform revolutionizes the way ISOs handle merchant applications, offering a seamless transition to digital forms. Say goodbye to outdated processes and hello to efficiency. Discover the future of financial applications at https://under.io/ftc

Advertisment

Duncan describes how the AI thinks about integration differently than traditional platforms. Instead of trying to replace entire systems, Lunos asks what a human worker would need to do their job effectively. They need to read the CRM to check if a deal is being renegotiated. They need access to the ERP for invoice data. They need to communicate via email and get updates through Slack. The AI connects with NetSuite, QuickBooks, Xero, HubSpot, Salesforce, and other systems the same way a human employee would.

When a customer responds with conflicting information, the AI interprets the message, figures out which records need updating, drafts the reply, and changes everything else in response. If someone comes back with a dispute for the first time, the system identifies it as a dispute, determines which invoices are in question, prevents further automated action against that customer, and notifies the appropriate team member for manual intervention.

Record keeping becomes crucial in preventing disputes. Most businesses track payment promises and reconciliation in spreadsheets they update sporadically. When records go wrong, simple issues become complex problems. The AI maintains accurate records in real time, preventing many disputes from happening in the first place.

Traditional machine learning models try to predict payment timing based on broad patterns across many customers. They end up investigating whether lunar cycles matter because some people get paid on lunar cycles. Lunos focuses on relationship-specific data instead. When a customer emails saying they'll pay next Friday, that provides far more accurate forecasting than statistical models based on completely different companies.

Duncan envisions networks of AI agents negotiating payment terms across entire supply chains. Picture a company perfectly willing to pay now in exchange for a 2% discount connected to a company four steps down the chain that's struggling with cash flow. AI agents could coordinate this type of arrangement in real time, managing negotiations that would be completely impossible with human-driven processes. Great products and services could thrive because the friction that currently exists in B2B payment processes gets removed.

Your days of choosing between data security and data usability are over. Whether you're just concerned with PCI compliance or need to go further to include CCPA, GDPR, SOC2, and beyond, Sky Flow has you covered. What if you could build fast but not break privacy? With SkyFlow, you can. Visit SkyFlowSecure.com today to learn how.

Tedd asks Duncan about the future of fully autonomous AI negotiations. Duncan breaks it into two stages: single agent to single agent, and networks of agents. The first stage involves two companies doing business where agents handle the communication, data reconciliation, and negotiation about terms. What might take days or weeks with humans could happen in seconds if all the data is set up properly, or days if the agents are still chasing humans for information.

The real power comes from the network effect. When Company A owes Company B, and Company B owes Company C, agents can create chains and networks of how these relationships fit together. Cash flow issues stop being isolated problems and become system-level challenges that networks of AI agents can solve through coordinated negotiation.

Duncan's advice to fintech founders comes directly from his experience: focus obsessively on solving customer problems and let everything else take care of itself. It's easy to get caught up in metrics and business operations. Those things matter, but understanding customer problems deeply should drive everything. When you're holding a hammer, everything looks like a nail. Stepping back to see the bigger and smaller picture at the same time allows you to identify what's actually broken.

Cash flow is the leading cause of business failure. Customers want their suppliers to succeed. They love the products. But poor communication, bad record keeping, and manual processes that can't scale cause failures anyway. Solving this problem isn't about building better payment rails. It's about recognizing that humans have been doing B2B payments the same way for thousands of years because the communication and negotiation components are genuinely hard.

Technology now exists to handle these genuinely hard problems. AI agents can interpret messy emails, maintain accurate records, negotiate terms, and scale to handle thousands of customer relationships simultaneously. Early results show companies collecting payments faster, reducing manual overhead, and improving customer relationships through consistent, professional communication.

The shift from manual processes to AI-driven automation addresses challenges that have persisted since people started trading goods and services. When 98% of a $70 trillion market still relies on manual processes, the opportunity for meaningful change is massive. Companies figuring out how to automate these communications effectively while maintaining appropriate human oversight could reshape how business transactions happen across industries.

TLDR:

Host Tedd Huff the Founder and CEO of Fintech Advisory firm Voalyre sits down with Duncan Barrigan, Founder and CEO of Lunos.ai, to unpack why 98% of B2B payments in a $70 trillion market still rely on manual processes. Duncan left his role as Chief Growth Officer at GoCardless, a $2.2 billion company, to tackle what he calls the most boring problem in business: accounts receivable chaos. The real issue isn’t payments, it’s communication breakdowns that cost companies millions in operational overhead. Lunos builds AI agents that negotiate, communicate, and resolve payment disputes like humans but without the limitations. One customer collected 25% of their entire AR balance, nearly a million dollars, within two days of activating full automation. The discussion covers how AI handles messy human responses, why email serves as the foundation, and what the future holds for agent-to-agent B2B negotiations.

Don’t forget to like, share, and subscribe for more insights from industry leaders!

HAWK:AI - HAWK:AI's game-changing approach to compliance. With real-time monitoring, adaptive learning, and advanced AI, it cuts false positives, simplifying your compliance efforts. Upgrade your surveillance with ease. Visit https://gethawkai.com for more intelligent, more effective compliance.

Advertisement

Key Highlights:

Embrace The Chaos Now

Business payments fail because companies try to force order onto human communication. Real customers send messy emails with wrong invoice numbers and conflicting details. The only way to handle this reality is by accepting the chaos, not fighting it with rigid rules that break the moment humans get involved.

Manual Work Costs Millions

Companies employ teams of 50 or more people just to handle accounts receivable operations. This manual labor costs businesses millions in overhead while major rideshare companies process billions in payments using spreadsheets and email chains. The operational burden keeps growing because traditional automation tools can't handle real human behavior.

One Email Changed Everything

A beta customer sent a single email that took three engineers 10 minutes to interpret. It referenced four invoices but also mentioned invoice number five, which didn't exist. Each reference had mismatched details about different payment issues. This moment forced the entire product strategy to shift toward building agents that could handle genuine human confusion.

Cash Collected Within Days

One customer activated full automation mode and collected 25% of their entire accounts receivable balance within two days. Nearly a million dollars came in simply because the AI could handle the volume of outreach and follow-up that no human team could match. The system wasn't doing anything magical, just executing consistent follow-up at impossible scale.

Email Beats Every Portal

Forcing customers onto new portals or workflows kills deals before they start. Email serves as the foundation because that's already how B2B payment discussions happen. When a customer responded in German, the AI understood the message, replied in German, and provided an English translation for the business owner without any specific programming.

Records Prevent Most Disputes

Most businesses download spreadsheets from their ERP once a week and try to track what's reconciled manually. When records go wrong, simple questions become complex disputes that require hours of phone calls and manual reconciliation. Real-time accurate record keeping prevents many disputes from happening in the first place.

Lunar Cycles Matter Sometimes

Traditional machine learning models investigate whether lunar cycles affect payment timing because some people get paid on lunar schedules. These broad pattern models fail compared to relationship-specific data. When a customer emails saying they'll pay next Friday, that beats any statistical model based on completely different companies.

Integration Without Replacement Works

Most fintech solutions die trying to replace entire financial infrastructures. Thinking like a human worker instead of a platform solves this problem. The AI connects with ERPs for invoice data, CRMs for relationship context, and communicates through email and Slack without forcing anyone onto yet another system.

Networks Beat Individual Agents

Single agent to single agent communication handles two companies doing business. The real power comes from network effects where Company A owes Company B and Company B owes Company C. Agents can create chains across entire supply chains, coordinating payment terms that would be completely impossible with human processes.

Customer Problems Drive Success

Walking away from a chief growth officer position at a $2.2 billion company provided the clarity to see what was actually broken. When you're holding a hammer, everything looks like a nail. Stepping back to understand customer problems deeply matters more than metrics or business operations.

HAWK:AI - HAWK:AI's game-changing approach to compliance. With real-time monitoring, adaptive learning, and advanced AI, it cuts false positives, simplifying your compliance efforts. Upgrade your surveillance with ease. Visit https://gethawkai.com for more intelligent, more effective compliance.

Advertisement

Takeaways:

1️⃣ Start With Monitor Mode First:

Begin by using visibility without automation to build trust and understand your AR patterns. Monitor mode lets you see everything in a single view while keeping your current manual processes intact, giving you confidence before advancing to suggest or act mode.

2️⃣ Prevent Disputes Through Real Time Records:

Stop tracking payment promises in spreadsheets you download weekly from your ERP. Maintain accurate records in real time to prevent simple questions from becoming complex disputes that require hours of phone calls and manual reconciliation.

3️⃣ Let Email Handle Customer Communication:

Build your AR automation around email instead of forcing customers onto new portals. When customers already use email for payment discussions, working within that channel removes friction and keeps negotiations moving without login barriers or password resets.

4️⃣ Use Relationship Data For Forecasting:

Stop relying on machine learning models that predict payment timing based on broad patterns across different companies. Focus on relationship-specific signals like contract renegotiations, historical payment patterns, and direct customer communications for accurate forecasting.

5️⃣ Integrate Like A Human Worker:

Think about what data a human employee needs to do their job effectively rather than trying to replace entire financial systems. Connect with ERPs for invoice data, CRMs for relationship context, and communication tools like Slack, allowing the AI to work alongside existing infrastructure without forcing massive changes.

Links:

Duncan Barrigan

LinkedIn Profile: https://www.linkedin.com/in/duncanbarrigan/

Lunos.ai

Website: www.lunos.ai

LinkedIn: https://www.linkedin.com/company/lunos-ai/

Fintech Confidential

YouTube: https://fintechconfidential.com/watch

Podcast: https://fintechconfidential.com/listen

Notifications: https://fintechconfidential.com/access

LinkedIn: https://www.linkedin.com/company/fintechconfidential

X: https://X.com/FTconfidential

Instagram: https://www.instagram.com/fintechconfidential

Facebook: https://www.facebook.com/fintechconfidential

Time Stamps:

Transform Your Merchant Applications with Under. The Under platform revolutionizes how you handle merchant applications, offering a seamless transition to digital forms. Say goodbye to outdated processes and hello to efficiency. Discover the future of financial applications at https://under.io/ftc 

Advertisement

About The Guest:

Duncan Barrigan

Duncan Barrigan is a technology entrepreneur and executive. He's the Founder and CEO of Lunos, the AI partner that manages receivables just like you would. Well, how you would if you worked on it 24/7, never forgot anything and read every message ever sent to you - you get the idea.

He has spent more than a decade helping businesses get paid, previously as Chief Product Officer and Chief Growth Officer at GoCardless for eight years, playing a leading role in its rise from a UK-based direct debit provider to a global bank payments unicorn worth $2.2bn, with revenue growing for $1m to well over $100m ARR. Prior to that he worked as a consultant, leading projects advising PE & VC clients on fintech and financial services M&A and strategy at Oliver Wyman. He has an MA in Natural Sciences from the University of Cambridge.

Lunos.AI

Lunos is an AI-powered accounts receivable platform. It integrates with enterprise systems (such as ERP and CRM platforms) to automate the tracking, communication, and resolution of unpaid invoices. The system is designed to operate like a human accounts receivable agent, managing customer interactions and adapting to responses in real time.

About the Host:

Tedd Huff: 

Tedd Huff is the Founder of Voalyre a Fintech professional services and Advisory firm focused on global money movement, payments and banking. He is also the founder host and executive producer of the Fintech Confidential network.

Over the past 25+ years, he has contributed to FinTech startups as an Advisory Board Member, Co-Founder, and Chief Experience Officer, providing strategic and tactical direction for Global Payments OpenEdge, Heartland Payments, Nuvei, and TSYS, among others, focusing on growth while delivering innovation, process improvements and user experience-driven value to simplify the complexity of payments.

DD3, Media: 

DD3 Media is a media creation, management, and production company delivering engaging content globally

Others you may enjoy

Why we moved to Beehiiv

The Newsletter Platform Built for Growth

When we started the newsletter, there were SO many choices. But until now, there hasn’t been a publishing tool built to help us grow our publications as quickly and sustainably as possible!

beehiiv was founded by some of the earliest employees of the Morning Brew, and they know what it takes to grow a newsletter from zero to millions.

It is an all-in-one publishing suite that comes with built-in growth tools, customization, and best-in-class analytics that actually move the needle - all in an easy-to-use interface.

We are excited to engage with you through — responsive audience polls (find out what you want to hear about most), custom referral programs( get rewarded for referring people to the Fintech Confidential newsletter), SEO-optimized webpages (make it easy to find the content you are looking for), and so much more.

If you have or are considering to starting a newsletter, there’s no better place to get started and no better time than now.

Try beehiiv absolutely free with no credit card required.