If you are one of those tech leaders in a company that already has an agentic architecture with swarms of agents delivering some value and your challenge is how to optimize the cost and/or stitch these agents together across complex workflows, this blog may not be for you. But do please read it if you have a few minutes, you’ll at least get a chuckle, maybe?

However, if you are one of those leaders who is not just laughing about an old Access database in the closet running for one user but rather still relies on it for critical business processes, do any of these questions resonate with you? 

  • Can agents use the Access database that runs a critical business process? 
  •  How would an agent interact with a green screen? 
  • Can agents learn the magical combination of keystrokes to navigate through legacy workflows? 

If not, lucky you. But I’m guessing most of you cringed at least one of those questions. 

Regardless of where AI ends up, what you need is still the same

My colleagues Charlie Betz and Manuel Geitz have published a new blog called the No Regrets AI Investment Agenda where they encourage tech leaders to think about what they’re trying to enable before making big changes. They are a lot smarter than I am, so read the blog. Their advice applies no matter where you are in IT or where your company is on the AI adoption curve.

As someone who has managed these big tech portfolios with extreme levels of legacy technical debt, I wanted to give you a few additional reminders:

  • Rethink app rationalization and tech debt with AI in mind. You are currently rationalizing apps, tackling technical debt, having fun physically pulling server plugs out of the wall. That’s wonderful. When was the last time you reviewed the priority order of apps to rationalize? Perhaps AI has changed the risk profile and value of apps and it’s time to re-evaluate your needs and priorities for legacy applications.
  • Reconsider time–to–value for your workflows. Somewhere, you have hopefully estimated time-to-value for all your tech workflows. AI could radically change that time-to-value, eliminating or automating step or obsoleting entire workflows. 
  • Identify the “Voldemort” systems. Every company has those systems that “shall not be named” because they strike fear in the hearts of every leader. They are widely known to unlock customer value or internal efficiency. Do you have money and plans to eliminate them? If not, slay these dragons instead of pursuing random AI use cases with unclear returns.
  • Continually look for organizational handoff spots. While an AI-related re-org or operating model change might seem appealing, don’t give in immediately. Instead, find spots in end-end processes where work gets handed off between different chains of command in the agency or organization. These handoff spots are often where waste occurs in the form of elapsed time.  “Oh yea, Tom in Accounting sent me an email last Wednesday, I just haven’t gotten to it yet because I’m working on rearranging my sock drawer.” If a customer priority isn’t the priority of all parts of the organization needed to meet it, fix them before you even think about dropping AI into the mix. 

Lastly, become best friends with your Security and Risk teams. Truly, BFFs. Have them over for dinner if needed. There are analysts at Forrester far more qualified to speak on this topic than I; my only point of view from experience is these legacy systems often have outdated patching and other holes that hackers using AI can breach at a far greater pace now.  

AI won’t save you from your pile of Access databases. It is still your job to figure out when and how to remediate legacy tech. Remember, the future is still human, served by AI. Not the other way around.  

Share