César Serrano says 80% of financial team time is wasted on manual data production.
1. After Work is a Spanish business radio program on Capital Radio, presenting a special edition on AI and financial management with four industry specialists.
2. Host Eduardo Castillo is joined by Alejandro Villar (Managing Director, FAICO Iberia), Pedro Plaza (AI Systems Architect, Plaza Robótica), César Serrano (Senior Solution Engineer, Lucanet), and Roberto Álvarez (AI Director for Spain and LATAM, UST).
3. The episode's core thesis is that the shift from generative to agentic AI is real and accelerating, but most companies are not implementing it wisely.
4. Alejandro Villar defines agentic AI as a 'virtual person' that applies AI examples and automates processes to produce results — from simple analysis to complex outputs beyond daily human practice.
5. Pedro Plaza argues that at Plaza Robótica, the 'human in the loop' approach remains the most effective: the AI assists, but humans make the first, intermediate, and final decisions.
6. César Serrano says the most common mistake he sees is companies wanting to automate a process before defining what that process actually is — 'you can't automate the what if you haven't defined the what.'
7. Roberto Álvarez states plainly that agentic AI is already producing a labor risk observable at UST: work done by ten people a decade ago can now be done by one, and UST is actively reducing headcount on client projects accordingly.
8. Pedro Plaza notes an emerging possibility: a first billion-dollar company run by a single person within five years — citing this as an opportunity not just for multinationals but for local businesses like dental clinics or barbershops.
9. A key danger identified is companies running isolated AI pilots without changing the underlying process — Alejandro calls this the 'pilot trap,' noting many pilots show no tangible results except for a small number of companies.
10. César, drawing on five years of CRM data from conversations with 400-500 CFOs across sectors, reveals that 80% of financial team tasks involve producing information — not analyzing it to make decisions.
11. He illustrates this with the distinction between 'a clever person who escapes a difficult situation' and 'an intelligent person who never enters one,' arguing that CFOs should use AI to redesign broken processes, not just automate them.
12. The panel warns against expecting AI to generate three-year business forecasts at the press of a button, because no AI can replace the CFO who must explain deviations to the board from first principles.
13. Alejandro frames the banking sector use case: AI fighting AI-generated fraud, arguing that when people see AI protecting their financial accounts, fear of AI decreases and trust in regulation increases.
14. European regulation — GDPR, DORA, EU AI Act, and CSRD — is identified as a structural pressure on financial-sector AI adoption, with potential fines and forced software disconnection for non-compliance.
15. César explains that CSRD forces CFOs to collect up to 1,600 ESG data points, including water use, energy consumption, carbon footprint, and EU taxonomy data, on top of existing financial reporting.
16. He describes a real CFO client who arrived at a board meeting where the CFO, COO, and HR Director each presented different values for the same data point because there was no shared source of truth.
17. Pedro Plaza identifies data mobility and confidentiality as a structural risk: data now moves via mobile, laptops, and cloud servers, and managing who can access what across AI and non-AI systems is 'quite important.'
18. Alejandro proposes four pillars of trustworthy AI in finance: robust (no hallucinations), ethical (no discriminatory bias), auditable (decisions can be reviewed), and traceable (decisions can be traced back to source).
19. Roberto Álvarez argues that AI governance should not be seen as a brake but as the enabler that allows companies to scale AI safely — requiring ethics committees, bias audits, and ongoing model performance monitoring.
20. Best for CFOs, financial directors, and business technology leaders who need a practical, regulation-aware framework for AI adoption; not for listeners seeking technical AI development content or consumer AI tips.
