Why 95% of AI Pilots Produce Zero ROI | Yasmeen Ahmad, Google Cloud
In this episode, Yasmeen Ahmad, Managing Director of Product Management for Data & AI Cloud at Google Cloud, reveals why 80–90% of enterprise data is "dark" and untouched — and how Google Cloud is building the tools to finally unlock it. Yasmeen shares how BigQuery's new Knowledge Engine captures the invisible business context that human analysts have always carried in their heads, and why this semantic layer is the real unlock for enterprise AI in 2026.
Yasmeen breaks down how enterprises are scaling from 50 to 2,000 autonomous AI agents, why continuous evaluation (not unit testing) is the only way to keep agents trustworthy, and what Google learned from seeing 50% of its own code now written by AI. She also explains why 95% of AI pilots produce zero measurable ROI — and why companies that partner with a platform like Google Cloud see dramatically different results. Plus, her contrarian take on governance: it's not the brake, it's what lets you drive 150 mph into the bend with confidence.
Key Topics Covered
Why 80–90% of enterprise data is "dark" unstructured data that GenAI can finally unlock
How BigQuery's Knowledge Engine captures the invisible business context analysts carry in their heads
The semantic layer: why the next big unlock is context, not just more powerful models
How enterprises are scaling from 50 to 2,000 autonomous AI agents
Intent-driven agentic AI: giving agents outcomes instead of step-by-step instructions
Why continuous evaluation is replacing traditional unit testing for AI agents
Google's internal AI adoption: 50% of code written by AI, 10% engineering efficiency gains
Why 95% of AI pilots produce zero ROI and what changes that outcome
AI governance as an accelerator — the "brakes that let you drive 150 mph" framework
Why culture and founder mentality matter more than technology budget for AI success
Episode Timestamps
00:00 - Introduction and welcome
00:50 - Being Scottish in Silicon Valley and the power of community
03:13 - The career thread: curiosity, pivots, and getting outside your comfort zone
06:20 - What makes data fascinating: the hidden stories inside numbers
07:48 - Why data is the lifeblood of enterprise AI
10:19 - 80–90% of enterprise data is "dark" and untouched
12:59 - What BigQuery actually does (explained simply)
14:50 - The invisible work: knowledge layers and business semantics
17:38 - The agentic AI moment: agents that think, plan, and execute
20:41 - From 50 to 2,000 autonomous agents inside enterprises
22:07 - Why you can't evaluate AI agents like traditional software
25:46 - Signals of AI readiness: Google's 50% AI-written code and Honeywell's 30% efficiency gains
30:21 - Why 95% of AI pilots produce zero ROI
35:37 - Governance as a speed accelerator, not a brake
39:53 - Who's best poised to win: culture over budget
45:59 - Why do you do what you do?
Yasmeen's Socials:
LinkedIn — https://www.linkedin.com/in/yasmeenahmaduk/
Partner Links
Book Enterprise Training — https://www.upscaile.com/
Subscribe to our free newsletter — https://www.theaireport.ai/subscribe
Learn more about your ad choices. Visit megaphone.fm/adchoices
Yasmeen breaks down how enterprises are scaling from 50 to 2,000 autonomous AI agents, why continuous evaluation (not unit testing) is the only way to keep agents trustworthy, and what Google learned from seeing 50% of its own code now written by AI. She also explains why 95% of AI pilots produce zero measurable ROI — and why companies that partner with a platform like Google Cloud see dramatically different results. Plus, her contrarian take on governance: it's not the brake, it's what lets you drive 150 mph into the bend with confidence.
Key Topics Covered
Why 80–90% of enterprise data is "dark" unstructured data that GenAI can finally unlock
How BigQuery's Knowledge Engine captures the invisible business context analysts carry in their heads
The semantic layer: why the next big unlock is context, not just more powerful models
How enterprises are scaling from 50 to 2,000 autonomous AI agents
Intent-driven agentic AI: giving agents outcomes instead of step-by-step instructions
Why continuous evaluation is replacing traditional unit testing for AI agents
Google's internal AI adoption: 50% of code written by AI, 10% engineering efficiency gains
Why 95% of AI pilots produce zero ROI and what changes that outcome
AI governance as an accelerator — the "brakes that let you drive 150 mph" framework
Why culture and founder mentality matter more than technology budget for AI success
Episode Timestamps
00:00 - Introduction and welcome
00:50 - Being Scottish in Silicon Valley and the power of community
03:13 - The career thread: curiosity, pivots, and getting outside your comfort zone
06:20 - What makes data fascinating: the hidden stories inside numbers
07:48 - Why data is the lifeblood of enterprise AI
10:19 - 80–90% of enterprise data is "dark" and untouched
12:59 - What BigQuery actually does (explained simply)
14:50 - The invisible work: knowledge layers and business semantics
17:38 - The agentic AI moment: agents that think, plan, and execute
20:41 - From 50 to 2,000 autonomous agents inside enterprises
22:07 - Why you can't evaluate AI agents like traditional software
25:46 - Signals of AI readiness: Google's 50% AI-written code and Honeywell's 30% efficiency gains
30:21 - Why 95% of AI pilots produce zero ROI
35:37 - Governance as a speed accelerator, not a brake
39:53 - Who's best poised to win: culture over budget
45:59 - Why do you do what you do?
Yasmeen's Socials:
LinkedIn — https://www.linkedin.com/in/yasmeenahmaduk/
Partner Links
Book Enterprise Training — https://www.upscaile.com/
Subscribe to our free newsletter — https://www.theaireport.ai/subscribe
Learn more about your ad choices. Visit megaphone.fm/adchoices
Loading content...
Read the full transcript
Matterfact has every podcast on Apple Podcasts and Spotify: 700K+ channels, 120M+ episodes.
The candid version is on a podcast,
not the earnings call.
Matterfact turns the conversations your coverage is already having into a searchable, citable record — across every episode, not just this one.