 |
Don't Panic! It's Just DataDon't Panic! It's just Data Author: EM360Tech
Turn data overwhelm into data-driven success. Led by an ensemble cast of expert interviewers offering in-depth analysis and practical advice to make the most of your organization's data. Language: en Genres: Business, Management, Technology Contact email: Get it Feed URL: Get it iTunes ID: Get it |
Listen Now...
AI’s Silent Killer: Technical Debt & the Hidden Costs of Bad Data
Episode 36
Thursday, 3 July, 2025
Enterprises face one common problem: the hidden costs of AI-based technical debt.“There’s a lot of hype around AI, but many initiatives aren't founded in a business value proposition,” says Paul Brownell, CTO, Growth Acceleration Partners (GAP). “People wander in without an intentional path for ROI.” In this episode of the Don't Panic, It's Just Data podcast, host Douglas Laney, BARC Research and Advisory Fellow, and author of Infonomics and Data Juice, speaks with Paul Brownell from GAP and Frank Lavigne, Advisory Board Member of CloudArmy.The speakers ultimately agree that AI promises greater returns on investment (ROI). However, it’s imperative to note – without a strong data foundation and strategy, AI can quickly turn into a financial nightmare.AI’s Potential to Cause Technical Debt Alluding to a significant “leak in the bucket” for AI initiatives, Brownell says, "a lot of these projects aren't founded in a business value proposition." This can often lead to organisations "wandering in without an intentional path." Both Brownell and Lavigne agreed that the most overlooked and costly area is data engineering. Lavigne exemplified this by referring to a meme depicting a sleek F-35 jet labelled "your AI" flying above a pockmarked, potholed road labelled "your data infrastructure.""I think that pretty much says it all," Lavigne stated, highlighting the critical and often unglamorous role of data engineering. Brownell resonated with this, calling it "mundane, routine, detail, hard pick and shovel work."Without mighty data quality, data governance, and traceability, AI projects are built on unsteady ground. Such AI initiatives occasionally result in inaccuracies and create a lack of trust.Scientific Path to AI Initiatives in DataBrownell advocated for a scientific approach to AI initiatives to overcome the hidden costs and maximise ROI. He said, "Come up with a hypothesis around where the business value is going to be, then apply some prototyping. Do real-life experiments to prove out your theory." Such an approach allows organisations to adjust course quickly. "The larger the ship, the harder it is to turn. So if you have these smaller kinds of proofs of concept, you can kind of find out in smaller increments how far we're off course,” explained Lavigne. This lowers risk and paves the way for more experimentation.TakeawaysAI investments can create hidden financial burdens.Data readiness is crucial for successful AI initiatives.A hypothesis-driven approach can guide AI projects.Iterative experimentation leads to better outcomes.Data engineering is essential but often overlooked.Generative AI can assist in data pipeline management.Selecting AI tools requires flexibility and speed.Purpose-built AI models may outperform generative models.Organisations must foster a culture of continuous learning.Understanding the total cost of ownership for AI is vital.Chapters00:00 Uncovering AI Technical Debt04:56 Data Readiness for AI Initiatives09:55 Selecting the Right AI Tools13:06 Generative AI vs Predictive AI18:14 The Future of AI Development
