CU 2.0 Podcast Episode 378 AI vs Big Data: The Anne Legg and Saroop Bharwani Show

November 12, 2025·46 min·3 clips

Is there a war inside credit unions between big data and AI, or can the two flourish in harmony?

1. The CU2.0 Podcast features Anne Legg of Thrive 3.0 and Saroop Bharwani of Senso debating whether big data and AI compete or complement each other inside credit unions. 2. Anne Legg is founder and CEO of Thrive 3.0, specializing in helping credit unions leverage data; Saroop Bharwani is co-founder and CEO of Senso, which has been bringing AI to credit unions for several years. 3. Both guests agree on the episode's central thesis: big data is the fuel for AI, not a rival to it, and credit unions that separate the two will underperform. 4. Bharwani reframes AI context requirements as analogous to human context: just as a presenter performs better with background information, an AI agent produces better outputs when fed both structured core data and unstructured call transcripts. 5. He argues that credit unions relying only on core system data are looking at 'one half of the whole picture,' and that call transcripts and internal chat logs represent an underutilized data set. 6. Legg links this to her decade-earlier critique of the Hadoop era: credit unions had the data but lacked a strategy, and the same risk applies to AI adoption — tools without governance and use-case discipline produce little value. 7. Bharwani traces the technical failure of Hadoop-era NLP to cost, latency, and the absence of the transformer architecture, which Google published openly in 2017 in the 'Attention Is All You Need' paper on arXiv. 8. He credits Google's open publication of the transformer paper as foundational to today's AI landscape, and expresses confidence that Google remains a major force despite its slower initial response to ChatGPT's November 2022 launch. 9. The practical AI example both guests highlight is One Nevada Credit Union's member-service representative knowledge base: fragmented SharePoint PDFs with version-control conflicts were replaced by a unified context layer queried in seconds. 10. Bharwani reports that Senso's credit union customers combined now average 91% response quality, up from a starting point of 30%, measured through a four-step improvement loop of evaluation, attribution, remediation, and tracking. 11. Suffolk Credit Union is cited as another implementation, where the context-layer project was tied to a core system conversion overseen by a staff member named Michelle Dean. 12. TrueStone Credit Union (Steve and Gary) and Arrowhead Credit Union are also named as organizations that have completed the foundational knowledge-unification step. 13. On the 1%-versus-100% call-analysis comparison, Bharwani describes how five human analysts at $70,000–$100,000 per year spot-checking calls have been replaced by the Echo AI fleet — with those same staff now using real-time pattern analysis to drive operational change. 14. The collections use case is raised by McGarvey via Altra's product at a Virginia credit union: an executive praised the AI-fielded collections calls because staff who previously endured hostile call recipients were redeployed to other roles. 15. Both guests oppose mass staff reduction as a strategy, with Bharwani arguing that credit unions willing to 'level up' alongside automation can grow without bound because AI-driven insight surfaces new work rather than simply reducing old work. 16. Legg uses the metaphor of skipping from bicycle-building to rocket-building to caution credit unions against overreaching; she advises starting with a use case that is already embedded in existing strategy and easily measurable. 17. McGarvey raises the competitive threat from JP Morgan Chase and Bank of America, noting that their AI-driven operational efficiency gains will widen an already-existing efficiency gap with credit unions if credit unions delay. 18. The tone is conversational and collaborative, with Legg and Bharwani riffing off each other's points; McGarvey functions as an engaged layperson who asks the questions a non-technical listener would ask. 19. Credit union executives, data officers, and technology leaders weighing AI vendor decisions or trying to build an internal AI strategy will find specific, named examples and measurable outcomes here. 20. Listeners seeking vendor-neutral AI strategy, consumer fintech perspectives, or technical deep-dives on model architecture will find the episode focused narrowly on credit union operational use cases.

As heard by us

Big data and AI are framed as complementary parts of the same credit-union problem.

Anne Legg and Sarup Barwani make a straightforward case that credit unions do not need to choose between big data and AI. The conversation stays practical: data quality, context, and the plumbing that lets AI work.

Read the full review in PlayNext →

Why you'd press play

If you want a blunt case for pairing big data and AI in credit unions, start here.

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from The CU2.0 Podcast