Over the next two years, generative artificial intelligence (genAI) will force organizations to address a myriad of fast-evolving issues, from data security to tech review boards, new services, and — most importantly — upskilling employees.
By 2027, AI will represent 29% of organizational spend, according to IDC President Crawford Del Prete, who spoke Thursday at the IDC Directions conference in Boston.
This year alone, the average enterprise will spend $28 million dollars on genAI initiatives, based on data from a February IDC survey. In all, organizations will spend $150 billion on genAI tech initiatives by 2027, with a total economic impact of $11 trillion, according to IDC.
Similarly, through 2026, tech providers expect to allocate 50% of R&D staffing and capex investments toward AI and automation. Forty percent of services engagements will also include genAI-enabled delivery, triggering a shift in human-delivered services.
“Enterprises are going to be going through a foundational shift; that includes hardware, software and data centric platforms,” Del Prete said.
Through 2025, 75% of organizations are expected to create AI implementation review boards; 40% will be looking to increase their outsourced AI services, including AI delivery; and 40% of new applications are expected to be more intelligent, as developers incorporate genAI to enhance existing and new use cases, according to a recent CIO survey by IDC.
“Over the last year, most organizations debated creating Chief AI Officers and centers of excellence to decide how to embed AI and create new business centers for new AI-enabled products and services,” said Rick Villars, group vice president of IDC’s Worldwide Research division.
On stage at IDC Directions, from left to right: IDC Group Vice President of Worldwide Thought Leadership Philip Carter; Group Vice President of World Wide Researh Rick Villars; and IDC President Crawford Del Prete.
CIOs are also rethinking their capital investment plans and staffing needs based on AI initiatives, according to Villars, including how AI will affect an organization’s long-term revenue and profitability.
Most organizations are likely to choose a hybrid approach to building out their AI plans — that is, companies will partner with service providers while also customizing existing AI platforms such as ChatGPT, as well as building their own proprietary, but smaller, AI models for specific use cases.
“All applications you buy will become more intelligent…, but make sure to not be redundant with the applications you’re building in house,” Del Prete warned. “That will be a big, big part in efforts going forward.”
In fact, 60% of enterprises will likely underperform on their genAI initiatives by failing to engineer connections between data, AI models, and the genAI applications they adopt or create.
The top five challenges to AI adoption will be:
- Measuring the ROI of AI investments;
- AI governance;
- AI skills and talent strategy;
- Concerns around long-term costs;
- Prioritizing the right use case.
The top challenge, according to Del Prete: finding skilled workers or upskilling existing staff to address the shift in genAI internal rollouts as well as new AI-enabled products and services.
Employee knowledge or training on large language models (LLMs) and genAI tools remains a top barrier to proper implementation, according to an October Harris Poll.
That survey, conducted on behalf of Insight Enterprises, found that a majority of business leaders have been tasked with helping their company define the ROI from genAI. However, only 15% consider the costs of implementation, including technical debt due to outdated infrastructure, initial financial investments, and ongoing maintenance costs.
The results of a more recent survey by global content and tech firm Thomson Reuters found that nearly 90% of respondents expect basic AI training to become mandatory for all professionals over the next five years.
Phil Carter, group vice president of IDC’s Worldwide Thought Leadership Research, said organizations shouldn’t expect an immediate ROI from their investments. Like other major economic shifts, such as arrival of the tractors for farming, the arrival of genAI technology can take decades to achieve widespread adoption and ROI.
"Tractors [invented in the late 1800s] promised to transform the agricultural industry and liberate farmers from mules and horses,” Carter said. “By 1940, only 23% of farmers owned a tractor.”
Like farming tractors, AI will require understanding of use cases, a willingness to spend significant revenue on buying the technology, working out problems with genAI's use and training workers.
For example, travel site Expedia announced last year it had rolled out a ChatGPT bot to assist travelers in planning their trips. The bot, for instance, could recommend activities once a traveler reached a destination. While it was used by consumers, Carter said there was no material business impact, according to Expedia’s CEO Peter Maxwell Kern.
The payback from genAI, however, can be tremendous, Carter said. By 2026, companies that master their deployments will double their growth over competitors through the creation of new internal efficiencies, and customer experiences and services.
Every organization considering an genAI strategy will have to create a AI governance-by-design program, including an internal data model and usage governance model. In that way, data security and AI governance will come together, Carter said.
The problem, however: only 36% of CIOs are currently building out AI governance programs, according to IDC’s survey.
Another concern for many has been the fear that genAI will kill jobs. Carter and others, however, are adamant that while the technology may change how employees perform their work, it will alleviate workers from mundane tasks and allow greater creative freedom.
Last year, multi-national home furnishing store IKEA announced an AI-driven skills strategy to retrain more than 8,500 call center agents to serve as “interior design advisors” while allowing an AI bot to handle almost half of all common consumer inquiries. As a result, the company said it is not seeing a reduction in headcount due to AI.
In another case study Carter presented to a packed house of attendees, an unnamed company’s call center agents were able to use genAI to summarize customer call transcriptions; the technology saved each agent an average of two minutes per call, leading to a 25% productivity gain and the ability to jump onto the next call faster.
“This fear of AI taking our jobs; we hear a lot about that,” Carter said. “This is about someone who knows how to use AI better than you. We have to embrace these technologies to know how to do our job better.”
While training staff will be key to employee comfort with AI tools such as ChatGPT, Midjourney or other AI assistants, organizations need to start by training leaders in AI proficiency and change management.
Originally published at Computerworld