At least 30% of GenAI projects will be dropped after POCs by the end of 2025 due to such issues as poor data quality, insufficient risk controls, fast-growing costs, or an inability to realize desired business value.
Gartner said GenAI is beginning to enter the trough of disillusionment in its latest Hype Cycle for Emerging Technology, 2024.
Respondents reported that their GenAI deployments have helped companies notch 15.8% revenue increases, 15.2% cost savings, and 22.6% productivity improvements.
Rigorously quantify business value, ensure data quality, privacy and security, and strengthen human-GenAI collaboration are key principles to guide the evaluation, selection, and enablement of use cases with GenAI.
Partners can help enterprises develop detailed business cases by holding workshops to understand overall goals, current state of data processes and infrastructure. They evaluate potential use cases, solve business pains, estimate ROI, and develop KPIs.
Enterprises must prepare data to ensure AI models generate accurate and reliable outputs. They must implement guardrails and tools to protect sensitive information, including model outputs, from exposure.
While GenAI will automate some processes, most of the time, it will assist humans in making better decisions. GenAI can provide scenarios and recommendations for decision-makers to consider so that they can optimize outcomes.
Leaders should take time to train teams on the latest capabilities, define roles and responsibilities clearly, provide guardrails and escalation paths when GenAI doesn’t perform as expected, and stress that they are augmenting human capabilities rather than replacing them.
Allstate has implemented a GenAI-powered chatbot that delivers real-time, multilingual support and seeks to improve the performance of previous models threefold by identifying those customer journeys that require agent support.
Leaders can use these three core principles – developing a sound business case, addressing data requirements, and helping teams collaborate with AI – to make new GenAI initiatives successful.