Enterprise Times attended the launch of Salesforce AI centre in London. The centre will be a learning lab for experimenting and developing new ideas, which will be shared with industry experts all over the world. Enterprise Times took the opportunity to meet Paul O’Sullivan, Head of Salesforce UK AI Centre and UK&I Chief Technology Officer. There’s a lot of hype about AI, generative AI, and predictive AI, businesses are truly excited by the potential of the technology. O’Sullivan shared his top 4 tips for the successful implementation of AI within an enterprise.
1. Retail flexibility
Enterprises still need to have a certain level of flexibility from their platform. For instance, on Saleforce’s Einstein One, the solution can work with language models that offer zero retention. It will not store data and enterprises can alternate various language models during the course of time. Organisations need to ensure that whatever the technical decision is taken, they incorporate a flexible approach. Because the current status of AI and the foundational language models will be very different in five years time. Technology is moving at such a rapid pace, and enterprises do not want to be locked into a language model.
2. Grounding AI in contextual data
Ground your AI in the contextual data relevant to your business and what the enterprise is trying to achieve. Enterprises decide how to shape and write prompts, the data that will fuel the accuracy of responses required. Part of the grounding process will be looking at the quality of contextual data within the business. Bringing all that data held in separate silos together to drive activation.
3. Deal with unstructured data
Tackle unstructured data that exists in the organisation. Salesforce’s research has estimated that 90% of customer data is trapped in unstructured formats. Businesses need to unlock insights from previously inaccessible unstructured data, including in PDFs, emails, transcripts, and other unstructured formats.
4. Embrace Change Management
The change management component is absolutely key. It will be important to take everybody in the business on the same AI journey. Companies have to educate their workforce about the AI changes taking place within the enterprise. If they fail to educate their teams, they could very innocently think they are moving forward incorporating ChatGPT or Gemini (formerly known as Bard, is a generative artificial intelligence chatbot developed by Google). Migrating enterprise data into a language model which is more consumer orientated.
