Key Takeaways
- Strategic planning is essential: Organizations looking to implement generative AI must have a clear plan to avoid past pitfalls, such as uncontrolled costs seen during the dot-com boom and initial cloud adoption.
- Buy versus build dilemma: The decision to build AI in-house or buy third-party solutions involves weighing factors like cost, talent availability, future readiness, scalability, time, and security.
- Talent and skills consideration: With the rapid evolution of AI, businesses need skilled talent for successful implementation, whether through training existing staff or hiring new talent.
- Vendor evaluation: For those opting to buy or partner, it is crucial to ask potential AI vendors the right questions about costs, licensing terms, support, and security measures.
If you’re a company that wants to start deploying AI in your business, you’re not alone. According to McKinsey and Co., 65% of survey respondents this year said their organizations are regularly using generative AI, almost double the percentage from a similar survey 10 months earlier. The bandwagon is quickly filling up.
And for good reason. From increasing efficiencies to improving the customer experience to cost savings, there are many benefits to adopting the rapidly emerging technology. It makes sense to join in.
However, organizations need a plan of attack. It wasn’t that long ago, during the dot-com boom (and then bust), when companies rushed to hire developers and create websites before finding they weren’t able to cover all the costs. Then the cloud came along, and organizations pushed data and applications into public clouds before seeing those costs skyrocket. A trend now is bringing back—or “repatriating”—some of those cloud workloads on premises.
The move to generative AI is moving even more quickly, but the desire to reap its benefits and stay competitive needs to be tempered by the need to plan. And the first decision that needs to be made is, whether you should buy or build it? There is a lot to consider, from what your business needs are to the expected outcomes. The direction chosen will set the stage for every decision that comes after.
Deciding whether to build or buy AI
There are benefits to doing generative AI in-house, including independence, control—particularly as more sensitive corporate data is being used to train AI models and needs to be secured—and the ability to develop your own models. But some aspects need to be considered:
Cost:
There are significant upfront costs to building an AI environment in-house, whereas a third-party solution can be had through a monthly licensing fee. That might make sense at first, but over time, the overall financial hit of building it could be blunted by not having to license the technology from third parties.
Talent:
Does your company have the skills to develop and run AI programs? A study found that while 81% of IT pros feel they can use AI, only 12% have the necessary skills. Another found that 95% of organizations say AI is a business priority, but 51% said they don’t have the skilled AI talent to put their strategies in motion. You may have to invest to train or hire that talent, while a third party likely will already have those skills in place.
Future-ready:
The innovation around generative AI continues to accelerate. Two years ago, ChatGPT introduced us to the idea of typing in a prompt and getting an answer in natural language. Now, we’re moving into an era of autonomous AI agents that can solve problems, make decisions, follow complex instructions, and work together on their own. How can you best keep up with it all?
Scalability:
AI solutions need to be flexible and adaptable. Developing your own solution means you can right-size it to fit your organization, but it also means having to scale it as needed and maintain it over the long haul. Third-party AI solutions run in cloud environments, which are more easily scaled up or down and where the maintenance responsibility falls on the partner, not you.
Time:
It can take a while—sometimes months or years, depending on the project—to build an AI solution, but if you have the time, it could make sense in the long run. However, if the need is urgent and opportunities will be missed by building a solution, buying or licensing the technology will lead to a faster deployment.
Security:
Bad guys are not only using AI for their own means but also targeting AI systems through such attacks as data poisoning and manipulation, prompt injections, and model theft. Do you want to take on the responsibility of securing the models and their data or put more of that onto a third-party partner?
Key questions to ask AI vendors or partners
Depending on the direction chosen, the shopping for AI technology or an AI partner begins, and the key to the search is asking the right questions. Some are ones a business will ask a vendor of any technology, such as the cost for buying or licensing the tools, terms for their use, or the support that comes with them.
But there are some that are AI-specific, and many cover the situation whether you’re building or buying. Here are a few important ones:
- How does your technology or service address our particular enterprise use cases? Is it scalable or customizable to meet our needs?
- Can your generative AI models be trained on or include our proprietary data? If so, how does your technology store and secure that data? What are the data retention policies for the data?
- How easily can your AI solution be integrated into our IT stack? How long will the integration take? What hardware and software do we need to have in place?
- How user-friendly is your AI technology? What training will be needed for employees, and how long will it be?
These are good questions to start with, and more will flow as you talk to potential sellers or partners and start to pare down your options.
Integrating generative AI: Key considerations for business success
The overall point is that generative AI is here, and it will be important to incorporate it into your business for operational and competitive reasons. A key is having the right information as you move forward. That includes understanding your needs, what an AI solution involves, and whether you want to do it yourself or go with a partner.
Once that decision is made, understanding what you need to know about the AI solutions you’re going to build, buy, or lease—and the questions to ask to get that information—is key. You don’t want to have buyer's remorse for such critical technology.
Related
The power of plain language in health care marketing.
Pharmaceuticals
The remarkable benefits of older employees in the workforce.
Diversity, Equity & Inclusion, Managing & Leading Teams, Recruiting & Hiring
Universities need to reconnect with COVID alumni now.
Marketing & Creative, Higher Education