ОТНОСТНО ТОЗИ ЕПИЗОД
In this episode, Minyang Jiang, Chief Strategy Officer at Credibly, shares how the company has approached generative AI adoption in financial services, from early experimentation to building sophisticated human and AI workflows. She explains how Credibly used generative AI to improve industry classification, risk analysis, data extraction, and emerging applications such as agentic pricing, while also learning where traditional automation is a better fit. The conversation explores the risks of bias, hallucinations, security, and overreliance on AI, along with the importance of human oversight and clear organizational guardrails. Jiang also discusses how leadership, employee education, experimentation, and access to enterprise AI tools can drive adoption while reducing shadow AI. Looking ahead, she argues that the strongest organizations will not simply use AI to cut costs, but will deliberately design human and AI hybrid operating models that strengthen creativity, decision-making, innovation, and long-term growth.
Takeaways
- Generative AI adoption should begin with experimentation and learning, rather than waiting for the technology to become fully mature.
- Credibly's early AI work focused on improving business industry classification and creating risk-adjusted categories that were more useful for lending decisions.
- Generative AI can create new value by connecting and interpreting data, rather than simply reproducing existing human workflows.
- Traditional rules-based automation is often better for repetitive, predictable tasks, while generative AI is better suited to work requiring reasoning, context, and flexibility.
- Trying to make generative AI exactly replicate human processes can limit its effectiveness and lead organizations toward the wrong use cases.
- AI is particularly strong at extracting and analyzing large amounts of information, but prioritization and final decision-making often require human judgment.
- Organizations need safeguards against bias, hallucinations, incorrect interpretations, model drift, security risks, and overly persuasive AI outputs.
- Successful AI adoption requires a combination of employee experimentation, dedicated technical or innovation teams, and strategic leadership from executives.
- Giving employees secure access to enterprise AI tools can lower barriers to experimentation and reduce the incentive to use unauthorized personal accounts.
- The future of work will likely center on human and AI hybrid models where people provide judgment, creativity, strategy, taste, and organizational context while AI expands their analytical and creative capabilities.
Today's guest, Minyang Jiang, can be found at:
LinkedIn: https://www.linkedin.com/in/minyang-jiang-6981a148/
Your host, Hunter Jensen, can be found online at:
Website: https://www.barefootsolutions.com/
LinkedIn:https://www.linkedin.com/in/hunterjensen/
Keywords
Generative AI, artificial intelligence, fintech, financial services, business lending, small business lending, risk assessment, industry classification, data extraction, data analysis, agentic AI, agentic pricing, AI agents, automation, robotic process automation, large language models, LLMs, GPT, Gemini, human AI collaboration, human AI hybrid, AI adoption, AI strategy, AI governance, AI guardrails, model bias, hallucinations, model drift, human oversight, decision-making, enterprise AI, shadow AI, change management, employee training, workforce development, future of work, innovation, SkunkWorks, AI experimentation, AI implementation, risk management, security, business intelligence, workflow automation, AI literacy, organizational culture, employee adoption, productivity, efficiency, creativity, strategic leadership, digital transformation, customer experience, underwriting, cash flow analysis, data parsing, retrieval augmented generation, vector databases, multimodal AI, model validation, enterprise technology, AI ethics
