Enterprises focused on clearing their debts are poised to scale with GenAI

Enterprises are transitioning from experimentation to strategic integration of AI, according to leaders who attended the first HFS OneCouncil GenAI discussions group.

Generative AI is rapidly emerging as a transformative force in the enterprise landscape. However, the journey to harnessing its full potential is fraught with significant challenges. Senior executives who attended our exclusive online discussion highlighted critical obstacles such as data and process debts, skills gaps, regulatory concerns, and sustainability issues. They shared that resolving these challenges requires strategic investments in data management, the establishment of centers of excellence (COE), cross-functional collaboration, and a focus on ethical AI practices. Furthermore, they are bumping into the data, process, skills, and technical debts identified in Exhibit 1.

Moving from experimentation to execution—the What-Why-How roadmap

We are moving beyond experimentation and are now focusing on strategic investments to integrate GenAI across our business units.

– Leader at a financial services company

HFS CEO and Chief Analyst Phil Fersht presented data from our latest GenAI research to the discussion group—accessible only to HFS OneCouncil members—sharing his take on where we are in the three-year journey GenAI has kicked off.

He told the group—to wide agreement—that 2023 was the year of ‘What’ as enterprises tried to understand the technology better and identify areas where they could experiment. They invested in multiple proofs of concept (POC) to showcase the efficacy of the technology and identify the key challenges to be addressed. 2024 is about the ‘Why’ as they look to identify areas for investment. Enterprise users reported prediction (34%) and personalization (28%) as the top use cases, followed by productivity (24%) in surveys conducted by HFS earlier this year. 2025 will be about the ‘How’ as they look to bring data, culture, skills, processes, and tech together to drive transformational change.

Addressing legacy issues is the key to unlocking GenAI’s potential

The technical debt we have accumulated over the years is proving to be a major hindrance.

– Enterprise leader at a consumer goods manufacturer

As companies grow, they accumulate numerous types of debt, including data debt, process debt, skills debt, and tech debt. When asked about the most significant challenges in implementing GenAI, data debt topped the list, closely followed by process debt. Skills debt and tech debt are the other key challenges enterprises face in their GenAI journey.

Exhibit 1: Enterprises need to navigate data and process debts to succeed

Sample: 550 enterprise leaders
Source: HFS Research, 2024

While GenAI enables faster data classification and migration, companies must invest in sorting out their data estate to maximize its benefits. Many companies still run on old processes and risk falling behind when they don’t invest in streamlining those processes to benefit from the new ways of working enabled by GenAI. Enterprises also increasingly face greater security and regulation risks as data becomes a more critical asset to safeguard and governments seek to regulate AI.

To drive successful GenAI initiatives, employees must combine business skills with technical knowledge and understand how to make AI both responsible and ethical. On the technical side, data science and machine learning, along with data analysis, interpretation, and visualization, are the most sought-after skills.

Firms turn to COEs to manage their way into the future

A COE helps in setting guardrails, tracking engagements, and investing in employee education.

– Leader at a healthcare enterprise

According to a survey of 260 enterprise leaders, HFS Research has found over half of enterprises are establishing or have already established GenAI COEs.

Exhibit 2: There is a growing appetite among enterprises to establish GenAI COEs

Sample: 260 enterprise leaders with GenAI experience
Source: HFS Research, 2024

Enterprises are looking for a COE approach because it helps them keep GenAI initiatives on track and establish guardrails for initiatives. COEs at enterprises have multiple responsibilities, including deciding on and tracking GenAI engagements, investing in employee education, and defining the framework for pursuing GenAI initiatives.

The internet revolution didn’t necessarily follow a COE approach. The technology was democratizing as it’s accessible to all—a common feature of GenAI. While COEs help with structure, focus, and prioritization, they must adapt to changes or be used mainly in the initial stages of the journey.

Sustainability and liability need to be part of the GenAI conversation

Who holds liability for content generated by GenAI remains a gray area, requiring clearer regulations and guidelines.

– Enterprise leader at an AI company

The liability for content created by GenAI is a gray area among enterprise leaders. While aggregator platforms, such as Twitter, YouTube, and Instagram, aren’t deemed responsible for content posted by users, it’s unclear who is liable for content generated by GenAI. Large language models (LLM) play an increasingly important role in customer-facing applications, so it’s unclear who will be liable for any damage caused by errors and hallucinations. Regulation and legislation typically lag technology evolution, leaving enterprises concerned about new risks.

The growth of GenAI has significantly increased the demand for computing and storage. Pure water is essential for operating semiconductor plants and cooling data centers. According to estimates, AI’s projected water usage could hit 6.6 billion cubic meters by 2027, signaling a need to tackle its water footprint. In a world where water shortage is only expected to increase, sustainability around GenAI must be given greater attention.

While voices committing to taking GenAI to scale were loud and clear among our group, we must acknowledge that those attending were a self-selecting group of pioneers and that, for many, the case for moving beyond POCs remains unproven.

The Bottom Line: As enterprises move from POCs to production, having a strategic roadmap will be essential.

Enterprises must invest in and integrate AI across business units to move beyond experimentation. This requires prioritizing data and process optimization, addressing tech and skills debts, fostering cross-functional collaboration, and establishing robust COEs to provide structure and enable governance. Additionally, a strong focus on ethical and sustainable practices will prove essential, including implementing training programs to enhance understanding of ethical AI and ensuring compliance with regulatory requirements.