As the initial excitement around AI wears off a few notches, business leaders are pondering less pressing but more practical questions about their AI projects. It's, “How much is this going to cost?”
The answer to that question depends on the size and purpose of a company's AI project. Building and training your own large-scale language AI models is a very expensive endeavor, with significant costs in hardware, energy, and engineering talent.
Fortunately, most businesses don't need to build their own models. There is already a wide industry of vendors helping businesses take advantage of the latest AI innovations, including Microsoft, Google, Amazon, and various AI startups. Many of our offerings also include open source AI models, bringing the power of these powerful generative AI tools to organizations that don't have the technical resources or skills to deploy them on their own.
“The cost of the AI tools themselves is actually not really prohibitive,” said Chief Transformation Officer for AI and Sustainable Materials at TE Connectivity, a maker of EV sensors and other electronic components. says Phil Gilchrist. “What is even more difficult is that we have to recognize that we are going to live in a world that is going to be an AI world, and we have to organize ourselves.”
For most companies considering implementing generative AI, there are three main pricing models to consider. Subscription-based pricing where businesses pay a recurring monthly fee per user. The other is an outcome-based model that sets the price when the customer deems the task successful.
Subscription-based models include services like Microsoft's Copilot AI assistant, which charges $30 per month for enterprise users, who can create data visualizations in Excel or use the overstuffed Outlook Empty your inbox faster. However, companies that want to deploy customized chatbots to handle customer support will likely pay different fees depending on how often the bot is used.
Boston Consulting Group's John Pineda advises clients to think carefully about whether the AI being deployed is intended to augment human work or completely replace a task. . If tasks are expected to be fully automated, “it starts to make sense to price either consumption-based or more outcome-based,” Pineda says. Conversely, if your use case is to make the way people work more efficient and support workflows, a subscription user-based model may be a better fit.
But Pineda also advises that companies shouldn't focus so much on the cost of AI that it stifles innovation. “Experimentation creates value,” he says. “Let people use it, try it, test it, [to] Come up with your own ideas about what you can do with the technology. ”
One size does not fit all
As companies realize the value of generative AI across their organizations, different teams and departments are likely to create their own projects, each requiring different combinations of AI models and different pricing plans. Become.
“Enterprises need to employ the right model for the right task,” said Ritika Gunnar, general manager of data and AI at technology giant IBM. According to her, 80% of companies now utilize multimodal approaches, relying on open source and commercial AI products. As companies continue to pilot and deploy AI into production, she speculates that they will continue to utilize different models to determine what is best.
The first step for any business is to determine whether AI is actually the best tool for the job. “Focus on the right use cases,” he says Gunnar. “Because AI and generative AI capabilities are the pieces that help accelerate the outcomes that businesses are trying to focus on.”
TE Connectivity's upcoming large-scale project will involve classifying approximately 200 million documents across dozens of databases. Gilchrist determines which AI models will help extract that information to increase TE Connectivity's competitive advantage, trains employees on how to implement and use the technology, and decisions will be needed to help employees become accustomed to using and trusting AI.
TE Connectivity is constantly developing new products, whether it's an AI chatbot created to help customers find information on a company's website or a tool designed in-house for engineers. We are piloting the technology.
“I'll try it first and prove it to myself.” [it] It really delivers the value advertised,” says Gilchrist. “We're very much a 'prove it to me' company.”
Abe Kuruvilla, chief technology officer at software maker ACI Worldwide, said many business leaders today are focusing on the power of AI and thinking about productivity in terms of reducing costs, including headcount. However, he is more focused on enabling his ACI employees to process tasks at a faster pace.
“For me, I'm still figuring out a monetization strategy around speed,” Kuruvilla says. “What is the value proposition for our clients compared to a typical traditional pricing model?”
He said ACI Worldwide and other companies are still trying to figure out the total cost of building products and using AI in the cloud. “The question is, 'How much is the market willing to pay for that speed?'” Kuruvilla asks.
If there's one thing that's clear so far, it's that the range of AI services is likely to continue to expand. According to Bloomberg Intelligence, spending on AI software, services and other products generated over the next 10 years is expected to soar to $1.3 trillion.
