Why integrating AI is proving to be a struggle for businesses

Companies are eager to integrate AI into their operations, but making it work effectively is turning out to be much harder than expected.

S&P Global’s 2023 AI report found that 69% of businesses have at least one AI project in production.

Many see AI as a way to improve customer satisfaction, enhance product quality, and increase revenue through innovation.

Others focus on using AI to cut costs by addressing inefficiencies in operations and IT.

Despite the enthusiasm, most businesses are not fully prepared to optimise their AI efforts.

AI depends on large, consistent datasets to work well, but businesses often deal with fragmented data in various formats—structured, semi-structured, and unstructured.

This lack of standardisation creates a major challenge.

Outdated IT systems and weak data management further complicate scaling AI.

Even when businesses move forward, transitioning to production environments for large datasets presents additional hurdles.

Performance and security issues also slow progress.

Many companies struggle with achieving fast, efficient processing and ensuring systems are secure enough to handle sensitive data.

AI’s carbon footprint adds another layer of complexity.

Training ChatGPT-3 in 2022 emitted over 500 tonnes of CO₂—far more than the average person emits in a year.

However, solutions are emerging. Google’s DeepMind showed AI could cut energy use in cooling systems by 9% to 13% at two sites.

Businesses with modern IT systems can reduce AI’s environmental impact while continuing to innovate.

What you should know:

The shortage of skilled workers is another major challenge.

MIT’s 2022 survey found that hiring is the biggest issue for companies with revenues between $1 billion and $500 billion.

It’s not enough to hire AI specialists.

Businesses also need translators who can align technical solutions with business goals and non-technical employees who feel confident using AI tools.

Aleksander Mądry of MIT highlights the importance of building trust among users to ensure AI systems are effective.

To succeed with AI, businesses need to prioritise data and talent.

Without organised systems and skilled workers, progress will be limited.

Companies that invest in both areas while balancing innovation with sustainability are the ones that will lead the way.

So, we can build AI that chats, but we can't manage our messy spreadsheets?