GenAI’s shine is waning: why aren’t companies going all in?

The clock is ticking for organizations to create significant value through their generative Artificial Intelligence (AI) (genAI) initiatives. Promising pilots have led to more investments, escalating expectations and new challenges. During this pivotal phase, C-suites and boards are beginning to look for returns on investment, and value-led use cases with strong ROI and a clear path to scale will be essential.

Meanwhile, according to the latest edition of Deloitte’s “State of Generative AI in the Enterprise” report, many Generative AI (GenAI) efforts are still at the pilot or proof-of-concept stage, with a large majority of respondents (68%) saying their organization has moved 30% or fewer of their GenAI experiments fully into production.

Our report dives into why this is the case, what barriers organizations are facing, and how they can overcome them.

The data dilemma

Data-related issues have caused 55% of the organizations we surveyed to avoid certain GenAI use cases. That could be because of data quality issues, intellectual property concerns, not having the right data, or worries about using certain kinds of data to train large language models (LLMs). Even those who consider themselves highly prepared for implementing GenAI will likely encounter unforeseen issues with Data Lifecycle Management (DLM) when moving from proof of concept to scale.

GenAI demands for data architecture and management means organizations need more robust data governance, especially for data that doesn’t already exist inside the organization. As more people within an organization leverage data, access frameworks and literacy require more attention. This may change a company’s approach toward cloud or on-premises (on-prem) data services and focus efforts on DLM to reduce risk. More advanced Large Language Model (LLM) users may eventually work with synthetic data, creating a new set of risks and governance challenges.

Yet, three of the top four barriers to successful development and deployment of GenAI tools and applications are worries about regulatory compliance (36%); difficulty managing risks (30%); and lack of a governance model (29%). Likely driving these concerns are risks specific to GenAI, like model bias, hallucinations, novel privacy concerns, trust, and protecting new attack surfaces.

Actions shaping the path forward

The top actions organizations are taking to improve data management are enhancing data security (54%); improving data quality practices (48%); and updating data governance frameworks and/or developing new data policies (45%).

The value from GenAI initiatives will increasingly come from organizations leveraging their differentiated data in new ways – whether for fine-tuning LLMs, building an LLM from scratch, or using enterprise solutions. For GenAI to deliver the kind of impact executives expect, companies will likely need to increase their comfort with using proprietary data sets, which may be subject to existing and emerging regulations. Without these actions, a lack of trust will likely continue to exist in the early stages of applications.

To help build trust and ensure the responsible use of genAI-powered tools and applications, organizations are generally working to establish new guardrails, educate their workforce, conduct assessments, and build oversight capabilities. Specific actions organizations are taking include establishing a governance framework for using GenAI tools and applications (51%); monitoring regulatory requirements and ensuring compliance (49%); and conducting internal audits/testing on GenAI tools and applications (43%).

The look ahead

Many organizations are learning that they can’t even get started with GenAI until they address their data deficiencies. Activities such as LLM tuning and training require high-quality data that is free of issues related to privacy, confidentiality, and intellectual property. As such, DLM should be at the top of every organization’s GenAI priority list, while simultaneously mitigating risks and preparing for regulation.

Will organizations demonstrate the patience and perseverance needed to unlock the transformational potential of GenAI? That is the outstanding question, as there is a chance that their interest in GenAI could wane if initiatives don’t pay off as much, or as soon, as expected.

The Deloitte Download

From the front-line experience of the world’s largest professional services organization, Deloitte leaders offer a practical take on the complex technologies that set businesses apart. Our authors go beyond the buzz to relate what they’Visualization Engine (VE) learned in real, timely cases that can help drive bottom-line results.