I spent years of my career getting paid well to patch applications at companies that had refused to invest in their engineering teams. The pattern was always the same, and I am watching it play out again with AI.
All of this stemmed from the company’s unwillingness to invest in the skills of their team and in some cases force engineers to learn new technologies. It is often the case that the company itself, especially at the senior executive level does not care what tech stack is used as long as the application runs well and the company can sell its products and services. There might be some directive like using a certain vendor for ecommerce but that rarely prevents day-to-day operations shifting with industry trends. Instead the cause of these issues would be engineers and managers of those teams finding some reason to dislike SPAs and newer web technologies, often irrationally, then once that happens for a couple of years it becomes very hard to jump back onto the train. Eventually it becomes harder to hire, your most excited forward-thinking team members leave and you are stuck with a decaying application hiring contractors to patch holes.
Companies will face the same problem with generative AI and AI tooling.
I have two companies that I am familiar with that have fully banned the use of AI for any application development, meetings or documents. They do not believe in the technology or that it can help their business. I know a third that has reduced their engineering team to rely on a model ill-suited to software engineering because they only wanted to have one model used across the whole company with marketers and business analysts outnumbering engineers and democratically choosing something for their purpose regardless of engineering’s needs.
The best companies are recognizing we are in a space with many models being created, foundational models shifting and what is best changing weekly. Having a stable of 2-3 models that can be used at the employee’s discretion helps plug gaps and upskill the entire team on what is best practice. Developing skills with AI practice now means you can train models internally on your existing data, create chatbots for internal wikis and get your entire workforce familiar on tooling. The investment in these programs now means that you are marginally faster at every task leading to significant improvements in your speed to market and release schedule over time. How much time does your team lose to finding the right person to talk to or understanding why a feature works a certain way? How many meetings have been booked to gather context when instead a chatbot could immediately gather that context?
These are challenges for businesses who are looking at a wide range of tools and not knowing what to bet on for their future success. Executives need to spend the money and commit to upskilling their teams. Trusting vendors in the AI space can be difficult when there are risks of data ownership and concerns around them training on company data. This is a risk you must be willing to take. These concerns are common and most model providers now have agreements to not train on provided data. The real risk is that by not building these internal tools using foundational models now you will be far behind the ball compared to your competitors. Trying to build these tools years from now once your best engineers have moved on will be a major project for consultants costing far more than it would if you started today.
Your internal tooling does not have to be perfect on its first release. Try building something small integrating existing data and boosting team capabilities in minor ways to ultimately build something big. Your IT team might be skeptical of AI especially on the security side which of course is what they are hired for but now is not the time to hide in our shells, we must be bold and make these technologies work for us. You likely have AI advocates throughout your organization, not just in your technology departments. Seek these people out and make committees, book clubs and study groups to get a broad understanding of how problems are being solved with different models.