Conference Presentation, Panel, Roundtable
'AI Blueprint Deploying GenAI for Enterprise Transformation' | RAISE Summit 2024 | Paris
RAISE SummitGabriel Hubert, Biljana Kaitovic, Bruno Zerbib, Laetitia Cailleteau, Philippe Battel, Jérémy Arosh
- Capture and translation capabilities will enable real-time internal language processing to trigger automated customer notifications during live feature updates.
- An exploratory mindset is identified as the primary predictor for rapid generative AI adoption, superseding technical proficiency, though distinct training is required to reassure those with uncertainty.
- Technology is projected to automate specific repetitive tasks without replacing human employees, shifting the knowledge workforce from deterministic "calculator" models to stochastic environments where identical inputs yield variable yet acceptable outputs.
- Product design strategies will distinguish between latency requirements for immediate conversational responses and offline processes delivering results in minutes or by end-of-day, with B2C products prioritizing speed more heavily than B2B offerings.
- Evaluation feedback mechanisms will evolve from binary ratings to granular star systems to isolate errors between knowledge bases and the AI assistant itself.
- Generative AI will be utilized to identify safety trends and risks in industrial data that humans previously could not detect, aiming to save lives.
- Organizations must balance support for exploratory users with top-down training and reassurance for those hesitant to adopt the technology.
- The stochastic nature of AI outputs may cause discomfort among users unfamiliar with the technology, necessitating a corporate culture of adaptability and growth mindset.
- Deployment of algorithms will require pre-screening to categorize models for compliance with regulations such as the EU AI Act before workforce-wide rollout.
- By 2030, Orange projects that most current human-performed tasks will be executed by AI, a transition expected to span 10 to 15 years for full realization.
- The same company has committed to a 40% CO2 reduction by 2030 despite a projected fivefold increase in resource demand from large language models, aiming to develop a sustainable version of AI.
- Future business models will be co-created by integrating technology and business teams to navigate the shift from an "age of innocence" to an "age of anxiety."
- The younger generation will drive the transition from "prompt engineering" to "agent AI" and "auto-prompting," while companies integrate existing data assets rather than starting from scratch.
- Operations in highly regulated sectors will increasingly favor on-premise deployment or private cloud options like DRCC to satisfy financial laws and the EU AI Act.
- The World Economic Forum estimates that 47% of working hours will be impacted by generative AI in the short term, while 95% of workers see opportunities but worry about improper scaling.
- Generative AI is expected to shift from incremental changes to a disruptive force within a few years, paralleling the evolution of electricity from replacing candles to building cities.
- Collective expectations include a shared journey to scale technology ethically, requiring collaboration across technology, business, and philosophy to address societal challenges like climate change.