AI, GenAI, most talked-about tech, still far from full enterprise adoption

Lack of frameworks in strategy, data, and tech slows AI adoption, says a latest report by Infosys Knowledge Institute

Artificial Intelligence (AI) and generative AI (GenAI) continue to capture global attention. However, research suggests that full adoption of AI in the enterprise world may take three to five years.

 

One of the main reasons for this, according to research by Infosys, is that companies lack an all-encompassing framework for implementing AI. According to the ‘Enterprise AI Readiness Radar’ by Infosys Knowledge Institute, only 2 per cent of leaders are confident in their readiness on a broad framework.

 

The report identifies five essential dimensions of an AI framework for companies: strategy, governance, talent, data, and technology.

 

Five dimensions of AI readiness

 

Strategy – This dimension assesses companies' readiness to approve AI projects with a validated business case and return on investment (RoI), comply with policies, ensure resource availability, define clear use cases with target customers, and secure sign-off from a governance council or similar body. Only 15 per cent of respondents were confident their AI projects had these elements in place.

 

Governance – Governance measures how well companies establish processes to mitigate AI risks, including hallucinations, bias, data misuse, and security concerns. This dimension also includes legal, ethical, and brand governance processes. Only 10 per cent of respondents were confident in their governance processes.

 

Talent – About 35 per cent of respondents expressed confidence that their employees are prepared for AI. However, 57 per cent were not, and 8 per cent were undecided. This dimension assesses employee proficiency with AI tools and techniques in their roles.

 

Data – The data dimension evaluates a company’s ability to locate and access data, including data accuracy and permissions. Only 10 per cent of respondents indicated that data was easy to locate and readily available for AI projects.

 

Technology – The research highlights that the most significant gap lies in the technology dimension of enterprise AI readiness, with only 9 per cent of companies prepared on this front.

 

Infosys’ chief technology officer, Mohammed Rafee Tarafdar, remains optimistic, expecting the percentage of confident companies to increase to 15–20 per cent within a year.

 

“When it comes to AI/GenAI deployment there are waves of deployment that happen, among enterprises it will not be one big-bang deployment. So what we are seeing now is the first wave of deployment. These are mostly AI tools being used for productivity and efficiency and quality. We are seeing this happen at core software engineering, customer service and engagement and experience etc.,” Tarafdar told Business Standard.

 

Tarafdar added that the next wave, which will encompass areas like IT operations, business processes, and process reimagination, will take more time.

 

Importance of AI readiness for competitive advantage

 

Tarafdar noted that companies need to prepare for AI to differentiate themselves in the evolving business landscape. He also emphasised that change management among leadership and employees is crucial for successful AI adoption. “From a strategy point or percentage of acceptance from a leadership standpoint is good. They are focused on having the right talent and they are putting together measures that are required for governance. Areas that lack are data and tech. From a CEO, CXO and board point I continue to see push from them,” he said.

 

According to Tarafdar, return on investment will drive enterprise AI adoption. He pointed out that AI costs have decreased over the past two years, suggesting that enterprise AI may be at an inflection point.

 

“If you look at the past two years, the cost of a token per million has come down significantly. It has come down by a factor of 100 to 200. Secondly there is a lot of focus now on small to mid-sized language models, and these are smaller in size. I think we are at an inflection point in enterprise AI to kick-in because the cost is coming down, the model sizes are reducing and enterprises are more aware of finding value and what works for them. From here it will slowly go up,” he said.

 

The Infosys research is based on a survey of over 1,500 executives across 12 sectors, including automotive, consumer packaged goods, and financial services, in Australia, France, Germany, New Zealand, the UK, and the US, conducted in June and July 2024. While challenges remain, the survey underscores that all companies will eventually need to adopt enterprise AI and GenAI.