Sign up to save your podcasts
Or
Share The Riskiest Agentic AI Use Case in Finance Is Lending
Share to email
Share to Facebook
Share to X
Share to email
Share to Facebook
Share to X
Or
The Riskiest Agentic AI Use Case in Finance Is Lending
Agentic AI is everywhere in finance right now — agentic commerce, agentic
payments, stablecoin protocols, Anthropic's ten new analyst agents. But
agentic lending is the one that touches creditworthiness, and that is why
the EU AI Act classifies it as high-risk under Annex III, alongside law
enforcement and border control.
In this episode of Conversations & Insights, I sit down with Tamara Laine,
Founder and CEO of Empower, who is building agentic AI infrastructure for
U.S. lenders. Tamara is an Emmy-winning investigative journalist turned
fintech founder, and she comes at the bias and explainability problem from
a perspective most fintech CEOs don't have.
We cover:
▪️ Why agentic lending is the riskiest agentic AI use case in finance
▪️ Two decades of fintech innovation: what alternative credit scoring and
machine learning actually fixed, and what they didn't
▪️ The gig economy data point that pulled Tamara into this work: 6 out of
10 gig workers in the US say they have been denied a loan
▪️ SR 11-7 in plain language: the U.S. model risk guidance that examiners
still use to validate AI in credit decisions
▪️ The EU AI Act, Annex III, and why August 2026 matters
▪️ Why Empower does not use AI for the actual credit decision (and why
that line is harder to hold than it sounds)
▪️ "AI prepares, humans decide" — soft version of Brian Armstrong's
one-person team vision, or genuinely different?
▪️ Why 80% of AI agents that reached production were pulled back,
according to a recent Sinch report
▪️ Five-year outlook: do the LOS providers like Encompass and nCino
eventually absorb the agentic layer, or do specialist agent companies
win?
The deeper question underneath all of this: when AI is making decisions
that determine whether a person can buy a home, a car, or grow a business,
where exactly does the human stay in the loop? And who decides where the
line gets drawn?
Guest Bio:
Tamara Laine https://www.linkedin.com/in/tamaralaine/ is an Emmy Award-winning journalist, fintech founder, and business growth strategist. As the founder and CEO of MPWR, she is transforming how credit and capital are accessed by gig workers, women, and underserved communities through AI and alternative data. A former gig worker herself, Tamara brings lived experience to her mission of building a more inclusive financial system. She has launched and scaled eight-figure tech ventures and is a recognized expert in financial inclusion, ethical tech, and economic innovation. Tamara also co-hosts What the Frac?, a podcast on startup growth.
View all episodes
By Efi PylarinouAgentic AI is everywhere in finance right now — agentic commerce, agentic
payments, stablecoin protocols, Anthropic's ten new analyst agents. But
agentic lending is the one that touches creditworthiness, and that is why
the EU AI Act classifies it as high-risk under Annex III, alongside law
enforcement and border control.
In this episode of Conversations & Insights, I sit down with Tamara Laine,
Founder and CEO of Empower, who is building agentic AI infrastructure for
U.S. lenders. Tamara is an Emmy-winning investigative journalist turned
fintech founder, and she comes at the bias and explainability problem from
a perspective most fintech CEOs don't have.
We cover:
▪️ Why agentic lending is the riskiest agentic AI use case in finance
▪️ Two decades of fintech innovation: what alternative credit scoring and
machine learning actually fixed, and what they didn't
▪️ The gig economy data point that pulled Tamara into this work: 6 out of
10 gig workers in the US say they have been denied a loan
▪️ SR 11-7 in plain language: the U.S. model risk guidance that examiners
still use to validate AI in credit decisions
▪️ The EU AI Act, Annex III, and why August 2026 matters
▪️ Why Empower does not use AI for the actual credit decision (and why
that line is harder to hold than it sounds)
▪️ "AI prepares, humans decide" — soft version of Brian Armstrong's
one-person team vision, or genuinely different?
▪️ Why 80% of AI agents that reached production were pulled back,
according to a recent Sinch report
▪️ Five-year outlook: do the LOS providers like Encompass and nCino
eventually absorb the agentic layer, or do specialist agent companies
win?
The deeper question underneath all of this: when AI is making decisions
that determine whether a person can buy a home, a car, or grow a business,
where exactly does the human stay in the loop? And who decides where the
line gets drawn?
Guest Bio:
Tamara Laine https://www.linkedin.com/in/tamaralaine/ is an Emmy Award-winning journalist, fintech founder, and business growth strategist. As the founder and CEO of MPWR, she is transforming how credit and capital are accessed by gig workers, women, and underserved communities through AI and alternative data. A former gig worker herself, Tamara brings lived experience to her mission of building a more inclusive financial system. She has launched and scaled eight-figure tech ventures and is a recognized expert in financial inclusion, ethical tech, and economic innovation. Tamara also co-hosts What the Frac?, a podcast on startup growth.