Hype vs reality: Is artificial intelligence delivering the promised ROI?

AI promises trillions, but organisations fail to capture value

Don't want to miss the best from Business Standard?

Regular conversations with customers highlight increasing expectations from artificial intelligence (AI). It has captured the imagination of business leaders like no other technology in recent memory. Be it CEOs exploring new revenue streams or CIOs rethinking their operating models, AI dominates every boardroom conversation.

 

Positioned as a critical driver of innovation and a potential growth engine, according to FICCI and BCG, it could inject more than $15 trillion into the global economy by 2030. This number has captured the imagination of investors and executives alike. AI promises to transform business operations, unlock new revenue streams and reimagine customer value. Delivering strategic advantage, AI is poised to become the new operating template of enterprises. The scale and speed of its adoption will be unprecedented and unpredictable, but ubiquitous, accelerated and deeply embedded across functions.

 

Yet, the early adoption narrative paints a sobering picture. Many of the leaders who speak passionately about AI also share a common frustration- they have yet to see real returns.

 

According to a BCG report, nearly three-quarters of companies (74 per cent) struggle to capture value from AI. MIT recently reported that 95 per cent of generative AI projects failed to achieve their desired return on investment (ROI) targets.

 

Gulf between aspiration and outcome

 

Currently, AI appears to be hype without scale. This is because most enterprises approach AI as an exciting pilot rather than a strategic enabler. Without linking it to core business objectives, customer journeys and data foundations, the technology risks remaining an experiment and not becoming a growth engine.

 

AI is often positioned and sold as a plug-and-play solution. But like any new technology, it brings its own set of complexities. Various customer conversations indicate four key obstacles:

 

- Undefined strategy: In the past few years, several fragmented and siloed AI projects in the industry failed to deliver measurable outcomes, just as expected. This is because many companies jump into pilot projects without integrating these experiments into the broader enterprise strategy.

 

- Data governance and security: AI models rely on massive data sets. Interestingly, most enterprises do not struggle with data scarcity, but the complexity surrounding its consumption and governance, including privacy, compliance and ethical use. This hinders the effective usage of AI models, preventing the move from pilot to production-grade AI.

 

- Legacy drag: Legacy architectures were never built for AI. They typically have complex, legacy IT environments with diverse platforms and technologies that compound the complexity. This becomes even more pronounced with Agentic AI, which demands fundamental re-engineering of digital infrastructure to support real-time data flows, dynamic orchestration and scalable compute environments. Only then can autonomous agents continuously learn, adapt and deliver evolved decision-making at the enterprise level.

 

From pilot to scale

 

Despite this, there are incidents where AI demonstrates value. Customers are reporting success with Agentic AI, especially in sales and marketing. In sales, it autonomously qualifies and prioritises leads, helping teams focus on high-value prospects and shorten sales cycles. Post-sale, it enables intuitive, personalised customer care to drive long-term loyalty and repeat buying. In marketing, real-time campaign optimisation maximises ROI and reduces ad spend waste. Further, it segments audiences, offering personalised content based on evolving behaviours, leading to higher conversion and engagement. But how do organisations replicate this across the enterprise?

 

Businesses must adopt a purposeful approach to AI. Rather than being treated as an add-on, it should be embedded as a foundational element influencing how these enterprises build, deliver and secure their services.

 

Organisations must understand their position on AI maturity to prioritise investments, close capability gaps and focus on initiatives that deliver measurable business value. A few principles stand out in these journeys, including those that resonate with customers:

 

- Anchor AI to strategy: As a first step, clearly define AI goals. Every AI initiative must be tied to clear business priorities, whether it’s revenue growth, customer experience or efficiency gains. Success should be measured in terms of cost savings and by how AI drives long-term differentiation and robustness in the organisation. Enterprises that approach AI intentionally will capture real value.

 

- Prioritise data: Data challenges are both at the technical and organisational levels. To address these issues, a unified data strategy is needed that spans the entire enterprise and prioritises secure, clean, connected and contextually rich data sets. To realise the full potential of AI models, data should be seen as a strategic asset that requires continuous management, governance and alignment with real business value.

 

- Inherent security: With the EU’s AI Act and the US National Security memorandum on AI, regulation is making security inseparable from AI deployment. For enterprises, this means embedding security frameworks from the very first stage of AI development and implementation. This calls for strict data protection at the operational level and the integration of transparency, control and resilience into AI systems.

 

- Onboard the right partner: Scaling AI is not a solo journey. Firms need a strategic partner to help drive meaningful outcomes from AI investments, not merely deliver one-off implementations. A partner that bridges infrastructure gaps, streamlines data pipelines, safeguards networks and guides them on ethical AI use, governance and long-term scalability. Dedicated FDEs (Forward Deployed Engineers) play a critical role as they work with customers, understanding their objectives, assessing real-time outcomes of AI applications and, ultimately, customising AI solutions aligned with business needs. The right partner will help quickly identify and address deployment hurdles – whether its data quality issues, integration complexities, infrastructure challenges or security barriers. This accelerates the path from pilot to scale, ensuring that outcomes remain the central priority.

 

From hype to outcome

 

Governments worldwide are accelerating AI adoption with various initiatives – whether it’s the EU’s digital innovation hubs or India’s National AI Strategy. But policy alone will not drive impact. Businesses need to look beyond hype and approach AI through a strategic lens, aligning their vision, people and processes.

 

In a nutshell, AI can act as a multiplier of human creativity, enterprise resilience and customer trust. The winners in the AI era will not be those who launch the most pilots, instead those who weave AI into the fabric of their operations, thereby converting hype into scale and experiments into sustainable enterprise value.

 

The writer is chief business officer, Tata Communications 

Disclaimer: These are the personal opinions of the writer. They do not reflect the views of www.business-standard.com or the Business Standard newspaper