I had a fascinating conversation today with someone leading a major technology transformation. We were talking about issues and challenges. Fixing-the-plumbing issues, alignment challenges, and the need to focus on getting the core elements right. And then we talked about innovation, and in particular the stakeholders who want to know how they are going to build Generative AI into the already-very-complex product. I expect there are a lot of people at the moment with those stakeholders. It reminded me of this blog by Hugh Wallace at Research Data Scotland on the Trouble with Innovation: https://lnkd.in/eE_Qpbh5 💡 The trouble with innovation being - among other things - that it distracts us from fixing the plumbing. It distracts from the mission and often fails to deliver. It can steal funding from the things we really need to do, because it makes the business cases for important things look boring. That's so often the case with buzzwordy innovative technologies. I'm sure it's happening with Gen AI right now in all sorts of organisations. But with Gen AI I think there's a chink of light, because I have never seen a hype tech before which made such a strong case for fixing your data architecture, quality and stewardship. The more I use it and work on strategies for deploying it, the more I think: this requires data-plumbing-fixing activities! 🔧 So I wonder, maybe the way to write a snazzy business case for doing the basics right is to express it as the critical path to Gen AI? 🤔
Sam Wright
1y
Hey Pete! Love this, and it's definitely something we see as well. AI is great, but to do anything really interesting you need to input good, clean data. To do that all the plumbing needs to be right in the first place. People in our field underestimate the importance of data engineering when it comes to AI all the time. It's interesting reading this and seeing that it's the case elsewhere too.
Yes, Peter Chamberlin - I share your thoughts! We're having a lot of conversations with clients recently like this. Gen AI is sooo shiny shiny that it may actually be enough of a distraction to fix the plumping while no one is looking.
Hugh Wallace
1y
Thanks for the name check and link Pete, and great to get your hopeful ‘reasons to fix the plumbing’ take on it. Maybe we *can* have it all!
So, so true Pete. I can't thing of another technology we've seen that so perfectly illuminates the very old adage of garbage in, garbage out :-) We need responsible technologists to present a clear eyed view that GenAI will make things WORSE if the time isn't taken to clean up the data.
Dan Frydman
1y
Had this conversation today with my colleague. That no matter how many new small business sites could be created with generative AI, it will be a long time before a plugin or AI service will be able to fix information architecture, UX and existing coding issues in a live site.
What about a business case to 'Fix the plumbing' with the assistance of AI? 😉 Use the capabilities of AI to detect what could be improved.
Was on a similar line of topic today in how important reading data correctly is, the old survivor bias topic etc
Partner, Public Digital
1yGreat thoughts (both Pete & Hugh!). To add to what you've said... an argument I've been making recently, but not sure I've completely figured out is this... If you want to use GenAI for stuff that is both exciting, and at the heart of your mission - you probably need to fix the plumbing, and you're going to have to wait a while to see if it really pays off. If you want to use GenAI for tactical, behind the scenes, localised quick-fixes, you might be able to use it immediately, but it won't be glamorous, and you'll have to tolerate a lot of imperfections.