Businesses Want AI, But Don’t Know How To Assess Its Impact
A TCS study finds that companies want to become "AI-ready”, but business executives are worried about how to show evidential performance.
Companies are enthusiastic about adding artificial intelligence to their businesses, but nearly three-fourths (72%) say they need a better way to assess its implementation, according to a Tata Consultancy Services business study.
The companies’ desire to become “AI-ready” is high, the IT giant’s survey showed. However, they’ve still not figured out key aspects like revamping their operating models and how best to measure the success of their AI implementations.
For their study, the company has defined AI as generative AI, as well as more established AI tools such as predictive analytics, forecasting, machine learning, simulation, robotics, and other similar technologies.
Business executives say they need to show that their investments in generative AI have to be evidential via performance indicators, or they risk losing their budget. It’s worth pointing out that, given the blistering pace of technology development, the lack of standardised or adequate assessment metrics isn’t surprising.
Businesses understand that AI is coming and is set to change how they operate. This is why 55% of businesses surveyed by TCS say that they are changing their business models, their offerings, and how they sell them.
But that’s separate from adapting AI for business. In that regard, only 17% of companies are discussing the technology and making enterprise-wide plans for it.
Companies across the world are racing to be on the cutting edge of AI, given the rewards they’ll be able to reap in the coming years. The revenue potential for the genAI industry is set to see extreme growth over the next ten years or less. GenAI’s market size is expected to grow to $1.3 trillion, compared to $40 billion in 2022, according to a report from Bloomberg Intelligence last year.
But building a large-scale language model is an expensive affair. Stanford University’s AI Index Report 2024 found that OpenAI’s ChatGPT-4 cost $78 million worth of compute to train, while Google’s Gemini Ultra cost around $191 million in compute. Despite this, 51% of companies say that they’re planning to build their own enterprise-specific models.
To get around costs, companies are taking the help of companies that can leverage their expertise in the technology. Of the companies surveyed, TCS found that 23% of non-tech entities are using external vendors for all or most of their AI implementations. That number climbs to 27% when it comes to technology companies.
It’s no shock, then, that plenty of AI-focused startups are popping up. Private investment in the field has octupled since 2022 to reach $25.2 billion, according to Stanford University’s report.
TCS released its TCS AI for Business Study shortly after its first-quarter fiscal 2025 results and spoke to 1,272 companies across 12 industries and 24 countries. Of the people interviewed for the survey, 16% were CEOs, 35% were division/business unit heads, and 49% were product owners at the vice president or senior vice president level.
