"The Analyst’s View of AI: Judgment, Costs, and Real‑World Use" featuring Bruce Daley

Episode Summary AI is changing how knowledge work gets done, but it is not replacing experts the way many people fear. In this episode, Clint and Greg talk with longtime software analyst Bruce Daley about what really happens when experienced professionals start using AI tools in their daily work. Bruce began experimenting with large language models because he believed AI might eventually replace analysts. Instead, he discovered something more interesting. Analysts who learn to use AI gain a major advantage over those who do not. The conversation also digs into the mechanics of building an AI product. Bruce shares lessons from launching Analyst Copilot, including why token costs matter more than most founders expect and why curated data is the key to avoiding hallucinations. Finally, the episode closes with a practical takeaway. If you are early in your career or looking to stay competitive, learning how to design effective prompts may become one of the most valuable skills you can develop. Guest Introduction Bruce Daley is a longtime software industry analyst who has spent decades studying how major technology shifts reshape the enterprise software market. Over the course of his career, he has covered hundreds of technology companies and has been quoted in outlets such as The Wall Street Journal and the Financial Times. Bruce started his career as a software developer in the early days of enterprise computing before moving into industry analysis, where he spent years helping investors and business leaders understand how technology markets evolve. Recently, he made a major career shift. Instead of only analyzing the AI wave, he decided to build in it. Bruce is now the founder of Analysts Copilot, an AI-driven platform designed to help analysts, investors, and business leaders track and understand large sets of companies using large language models. Join Bruce for a live demo of Analysts Copilot when he is a guest on Carter Lusher's Tool Talk podcast on March 12, 2026.https://www.linkedin.com/posts/carterlusher_archat-analystrelations-artooltalk-activity-7436406958565404672-BmHv/  AI Challenge: Prompt Engineering Taste Test This week’s AI Challenge is designed to show just how much prompt design affects the quality of AI output. Take a simple question like:      “Explain the CRM market.” And then run it through an AI tool three different ways. First, use a basic prompt. Second, add structure. Specify tone, format, and constraints. Third, create an expert prompt that defines a role, audience, and desired output. Same model. Same question. But the answers will be dramatically different. That experiment will show you why prompt engineering is quickly becoming one of the most valuable skills in the AI era. You can find the full instructions here: https://www.promptthis.ai/blog/how-to-test-out-prompt-engineering  And if you want to share your results or your own AI experiments, reach out to us at: https://www.promptthis.ai/contact  Chapter Breakdown 00:00 – The AI hype problem The show opens with the mission of Prompt This: cutting through AI hype and focusing on real business applications. 01:11 – Introducing Bruce Daley Clint and Greg introduce Bruce and explain how he moved from software analyst to AI founder. 03:17 – Can AI replace analysts? Bruce shares why he initially tested AI to see if it could replace him and what he discovered instead. 04:56 – Using AI to track hundreds of companies How Bruce used large language models to scale his research coverage from dozens to hundreds of companies. 06:45 – What AI cannot replace in the analyst role Industry relationships, informal information networks, and professional judgment remain uniquely human advantages. 09:11 – Why AI improves research quality AI handles first drafts and repetitive work, giving analysts more time to refine insights and perspective. 09:23 – Comparing AI to past tech waves Bruce compares the AI moment to the rise of the PC and the internet. 11:09 – Build versus buy in the AI era Why infrastructure and developer tools come first, followed by applications that solve real business problems. 14:17 – Understanding tokens and AI cost models Bruce explains how token usage affects cost and why model selection matters. 15:07 – Lessons from building Analyst Copilot Early mistakes, unexpected token bills, and why cost discipline matters when building LLM products. 16:02 – The importance of data quality Why curated data is the best defense against hallucinations. 18:00 – The rise of prompt engineering Bruce explains why prompt engineering could become a major career skill. 18:49 – Real-world use cases for Analyst Copilot How sales teams, investors, and analysts use AI to stay informed and prepared. 20:44 – Why specialized AI tools beat generic prompts The advantage of curated data, structured prompts, and purpose-built interfaces. 23:14 – AI Challenge: Prompt Engineering Taste Test Clint and Greg introduce this week’s hands-on AI experiment. 24:30 – Where to find Bruce and Analyst Copilot Bruce shares where listeners can learn more about his work. 25:00 – Episode wrap-up Clint and Greg close out another episode of Prompt This.