Companies can add $460 bn to profits with better AI implementation: Study

Global spending on artificial intelligence (AI) -centric systems would approach $118 billion in 2022 and grow to more than $300 billion by 2026, said a recent study. But all this spending is not paying dividends. Infosys Data+AI Radar found in a study that companies could generate over $460 billion in incremental profits if AI implementation was improved.

The three things companies need for these gains are improved data practices, trust in advanced AI, and AI integration with business operations. However, despite high expectations for data and AI, most companies fail to act on these areas to convert data science to business value.

According to the Infosys Data+AI Radar ‘Making AI Real’ report, although three of four companies want to operate AI across their firms, most are new to AI and face daunting challenges to scale. At least 81 per cent of respondents deployed their first true AI systems in only the past four years, and 50 per cent in the past two.

Balakrishna D R, Infosys’ executive vice-president - global head AI & automation and ECS, points out that there is a mismatch in expectations. “What we found is that most enterprises have dabbled with AI. Many have done POCs and they have use cases. But their expectation has been to leverage AI and make steep changes to the way their organisation’s working, efficiency gains, etc. And that has not happened,” Balakrishna said about the findings of the report.

The report also found that 63 per cent of AI models functioned only at basic capability, were driven by humans, and often fell short on data verification, data practices, and data strategies. Only 26 per cent of practitioners were highly satisfied with their data and AI tools. Despite the siren song of AI, something is clearly missing.

Satish HC, executive vice-president and co-head for Delivery, Infosys, said: “Companies that build foundations to trust and share their data are more agile and scale their AI. Companies that don’t trust their data risk a vicious cycle of “pilot purgatory” and only use data and AI to solve small problems. Data management and trust in AI form dual solutions to increase business capability and financial rewards.”

Balakrishna also pointed out that having a data centre or a team of data scientists was a good thing to have but senior executives also needed to be involved in aligning business units to leverage the value.

The survey, which covered 2,500 AI practitioners, found that 81 per cent deployed their first AI systems in the past four years. However, most companies (85 per cent) had not achieved advanced capabilities, and most AI models (63 per cent) are still driven by humans. Compounding this, outcomes were middling at best: Users were highly satisfied with their data and AI results only about a quarter of the time.

Balakrishna also pointed out that collection of data and how clean it was also presented a challenge within organisations which then impacted results. “Trying to create a central repository for data can be a humongous and time-consuming exercise. We instead recommend a hub-and-spoke model that will bring the best of both worlds.”

The reports said the old process – extract, transform, and load data into a private warehouse – faced limits. Followers of that procedure could only apply AI to the data contained in the four walls of their warehouse. Data management strategies that fostered data sharing, both importing in and sharing out, expanded the universe of available data.

Infosys Knowledge Institute found that high-performing companies think differently about AI and data, and these leaders focus in three areas:

• Transform data management to data sharing. Companies that embrace the data-sharing economy generate greater value from their data. Data increases in value when treated like currency and circulated through hub-and-spoke data management models ($105 billion incremental value). Companies that refresh data with low latency generate more profit, revenue, and subjective measures of value.

• Move from data compliance to data trust. Companies highly satisfied with their AI (currently only 21%) have consistently trustworthy, ethical, and responsible data practices. These prerequisites tackle challenges of data verification and bias, build trust, and enable practitioners to use deep learning and other advanced algorithms.

• Extend the AI team beyond data scientists. Businesses that apply data science to practical requirements create value. The report found that business—data scientist integration accelerates efficiencies and value extraction (additional $45 billion profit growth). For intelligent data, business and IT are much better together.

Combined, these areas not only scale AI usage but unlock its potential value – transforming AI dreams to insights and operational effectiveness and improving the human experience. Infosys research found the financial services industry recorded the strongest satisfaction with its data and AI uses, followed by retail and hospitality, healthcare, and high tech.