top of page

"Quilting" AI applications

Updated: Mar 17

The year was 2002, and I remember when I first stepped into the “rack room”. That classification being a stretch, because the company only had a single 42U rack, located in the closet across from the break room. We were a dozen engineers, using our own “desktop replacement” computers and applications. Seven years later, two changes of control and exit at last; but before we could complete the transaction, we had to clean up the tech-stack.


Fast forward two decades, replace the rack room with SaaS applications, and let the latest technology be AI; if your business teams are not coordinated and managed, your business is likely to end with a siloed patchwork of applications and services making up your tech-stack. As Murphy’s law well states, integrating that “will be a problem at the most inopportune time.”



A collection og GenAI application logos
The GenAI ecosystem is a powerful patchwork. Make it work for you!

Avoid the operational challenges that are inherent to the “tech-stack patchwork" by planning your technology and business integration process from the start.


I get it, businesses want to move on AI as quickly as possible. No disagreement on that at all - let’s go! Taking a moment to slow for a bit and plan ahead will help your business implement and adopt, repeatedly and effectively, the “then latest” AI technologies.


Here are three factors to consider:




Who will do the implementation?

Delegating the implementation of AI applications to individual departments leads you immediately to the patchwork you should avoid. You are likely to end with multiple disparate applications for a similar purpose, or using an AI application that does the work well for one department be “close enough” and maybe forced onto another department.


Delegating the implementation of AI technology to any one department is also fraught from the word go. If you choose IT, remember that they already have plenty to do, and while they most certainly need to be involved, they will stop work when the network needs help, and will likely implement technologies; not integrated solutions.


Imagine marketing using Excel to write their copy… or engineering running their calculations in Word. Imagine IT working through their stack of trouble tickets after working to roll out the wanted AI apps. You laugh, but history repeats itself, and I’ve seen the above happen more than once.


Assign the implementation of Ai technology to an entity that is focused on the end-to-end business integration. One entity that is empowered to manage all stakeholders and keep business harmony from day one.


How do you future-proof the implementation?

A fact of our business lives is that technology will change periodically and when least expected. If you know the inevitable, please plan for it.


Avoid getting locked into using any one app. Whether it may be by the contract you execute, or the implementation mode you choose - don’t. 


If all possible look for a way to bring the best capabilities and features that you seek from the GenAI world through APIs, or perhaps an aggregator of AI that allows you to replace AI apps as they disappear or are absorbed into another with zero impact on your business tech-stack. 


Use an appliance or application that presents a constant interface to the tech-stack no matter what happens the AI that is “upstream”.


Which brings me to the next element to consider…


How do you protect your current tech-stack investments and avoid them from locking you down to limited AI capabilities?

Bringing in the functions and features that you need from any AI application to any specific tech-stack will help your team overcome the challenge of being stuck using the limited “meets the standard” native AI capabilities of current apps.


Unpopular opinion, most AI capabilities that are native to your tech-stack applications are narrow, and focused on the “ooh” factor of the individual app. Nothing wrong with that; but it just doesn’t help the rest of the team or business.


Good planning ahead, and smart business process focused integration helps avoid this problem if you use an appliance or application that presents a constant interface for the AI applications connecting to the tech-stack.


In short - 

  1. An independent tiger team (internal or external)

  2. Using a common interface for AI and current tech-stack apps.

  3. Integrating end to end best-in-class features with “business first” top of mind.




8 views0 comments

Recent Posts

See All

Comments


bottom of page