The mantra for computing is to use other people’s technology, so it’s surprising that generative AI doesn’t seem to follow that pattern for many enterprises.
Generative AI is pushing some critical decisions and doing so very fast. Every organization faces the crucial decision of whether to build a custom generative AI platform internally or buy a prepackaged solution from an AI vendor, generally delivered as a cloud service.
The numbers and the opportunities are working in favor of DIY. That’s very strange, but the reasons may surprise you. They may even lead you to rethink your enterprise genAI strategy.
Complete customization and control
Building a generative AI platform from scratch gives an enterprise total control over its features and functions. The AI tech can precisely adapt to the organization’s requirements. This ensures compliance with the company’s unique workflows and provides a bespoke user experience. Remember that DIY generative AI can be done on public, private, or traditional platforms. Today, we’re focused on using specific genAI technology, primarily open source, either on premises or on a public cloud.
Natural language interactions do provide a more “human” approach to dealing with static business processes. However, people are concerned that these systems may become core to the business quickly and that unless they have complete control over all features and functions, they risk the system not providing total value. Namely, if a purchased AI platform with all the bells and whistles changes direction or even goes away, they are stuck with a failed system and a failed business.
More money, more time, more risks
Building a complex generative AI platform requires a team of experts with specialized knowledge, and it’s very difficult to find enough of them in the existing talent pool. You need data scientists and AI engineers to work with platform engineers, cloud and non-cloud, to develop customized genAI solutions built to the exact specifications of the business.
This can lead to increased complexity and the need to hire expensive talent. I have a CIO friend who is sending his staff out to graduation ceremonies at good technical universities and recruiting people directly before they hit the open job market by approaching them in the parking lots at the school. That’s disturbing but innovative at the same time.
Most enterprises must be creative to find enough people. Some enterprises are running into this talent roadblock and are delaying their projects or deciding to buy a system rather than build one.
The value of buying
Buying a system offers rapid deployment and functionality out of the box. This includes prebuilt solutions that allow for quick implementation. You get immediate value and an accelerated time to market. More importantly, buying a generative AI service guarantees ongoing support, updates, and improvements. Although the DIY approach can come with some help for some parts, you’re mostly on your own if you opt to build.
Think of the cost of building and supporting your database versus buying one from a database vendor. AI systems are much more complex and have many more components, but the analogy is apt.
The value of the build approach wholly depends on your need to create a one-off solution custom-built to the business needs. You’re betting that the additional cost, time, and risk will pay off in having complete control of the core system, which, for many, will become the business, not just automate the business. It’s likely the correct and strategic use of genAI in the upcoming years will make or break a business; a great deal is at stake.
Weigh all the factors
When deciding between building or buying a generative AI platform, consider all the pros and cons. First, the cost of building generative AI internally can be substantial. In contrast, off-the-shelf solutions offer practicality and cost-effectiveness. Second, building generative AI internally requires assembling a proficient team, whereas an off-the-shelf solution gives you access to the expertise of the AI vendor that made the system. This means pushing the risk and the cost to the vendor or provider.
Finally, creating AI solutions from scratch means complete creativity and control over the technical process. This allows for incorporating compliance measures and the exact functionality to meet the requirements from the outset. We all know how building goes. Customization leads to numerous iterations and time-intensive development. Also, support and maintenance are crucial for in-house gen AI. If this does not provide enough value to justify a DIY approach, look at buying, which removes risk, time, and cost.
I suspect we’ll see many poor decisions end up with the business going bust. Maybe they couldn’t provide the technological value in their industry because they bought when they should have built. Maybe they couldn’t build something of value because they lacked talent and had a limited budget. No pressure.
By: David Linthicum
Originally published at: InfoWorld
