As businesses increasingly adopt artificial intelligence (AI) and Generative AI (GenAI) to enhance operations and drive innovation, many are encountering significant roadblocks in their integration efforts. Challenges such as skill shortages, data security concerns, and limited access to high-quality data are posing hurdles, according to Anshul Pandey, senior director and head of Applied Innovation Exchange Mumbai – India, Capgemini.
Speaking with FE, Pandey said: “The maturity of the technology itself is one of the main challenges… Enterprises are adopting, but it is slow adoption. It’s not very fast-paced. So most of them are into the POC (proof-of-concept) phase.”
One of the most pressing obstacles is the skill gap within organisations. Although there has been progress in areas like big data, the lack of sufficient expertise remains a barrier to GenAI adoption. “Skills are not yet mature enough in the organisations,” she said.
Another significant challenge is data availability and management. Pandey highlighted that some clients, particularly those in research and development, are hesitant to move their data to the cloud due to privacy and security concerns. This reluctance limits their ability to leverage scalable AI infrastructure, which could otherwise lower costs and speed up results.
“But I would suggest that clients should move to the cloud, at least with simple use cases. It helps in costing and will help in their fast outcomes as well,” she said.
Further Pandey highlighted the importance of a comprehensive data strategy to fully harness AI’s potential. A robust data strategy includes capturing, curating, and consuming data in alignment with business objectives.
Tailored approaches are particularly necessary for sectors such as public services and financial services, where data needs are distinct. For instance, Pandey shared an example of a financial client transitioning from SAP HANA to Snowflake using a data mesh framework. Capgemini also helped integrate large language models (LLMs) like ChatGPT with Neo4j knowledge graph, achieving a successful proof of concept.
Capgemini is tackling these challenges through tailored solutions and strategic partnerships. The company promotes the use of industry-specific clouds and works closely with partners to help clients prioritize their data strategies for maximum value, Pandey said.
The use of synthetic data is also emerging as a solution to data availability limitations. Synthetic data allows clients to conduct POCs without risking sensitive information, making the process more efficient and cost-effective.
Pandey said that tackling these barriers is essential for businesses to unlock the transformative potential of AI, ensuring faster outcomes and maximising efficiency in their operations.
