Findings from Silicon Valley part 3: SalesForce's take on #AI

Still in same building as on my previous article, but now drilling into Salesforce 's views just day before Dreamforce . I think at least for me this approach was better as we were able to walk through the topics in nice non-crowded surroundings. We also managed to move out of San Francisco on same day to Palo Alto.

Show casing some points which were catching my eye during the visit and naturally commenting those from my own opinion point of view.

I think this is an interesting point also in general level: The focus of discussion is moving as AI is evolving and new domains / ways of using it are presented. The above presented illustration is focusing into end-user point of view, which is quite natural for platforms like SalesForce represents. Although I have been thinking that we still have plenty of more deep-down located challenges, which could gain from help provided by AI algorithms like neural networks. If the attention is on end user level, companies focus on employee efficiency and might loose the opportunities to transform processes or even business with more "embedded" type of AI utilization.

I believe that there is still a lot to gain from deep learning and predictive AI as soon as generative AI will start to lose momentum in public discussion. Naturally current use cases show generative AI's potential as tool for enhancing efficiency in specific areas like customer service and simple summarization type tasks. More challenging and specific tasks naturally challenge the model training and capability to create company specific models with clean and relevant source data.

Data is critical success factor for all AI use. We were discussing about data driven marketing with hyper personalized content. That has been done already before era of AI at certain level. In this topic the scary part earlier was, if the marketing planner was able to identify relevant needs so that results would be appealing also for customer. Now we transfer this responsibility to hands of AI, which naturally can deliver as good results as good the source data used for model is.

Based on earlier press releases SalesForce seem to be utilizing IBM watsonx data platform. This co-operation sounds clever both parties are focusing they key competence areas and utilize these capabilities for common benefit provided to customer companies. Also ability to connect different data sources is an advantage complemented with data flow mapping and modelling.

I think the above illustrated picture nicely pulls together the role of generative AI's challenge and merit: you need to keep balance between both. Use it in clever way in those tasks where it fits to support your employees. "Clever" includes here understanding what AI is capable and what it is not to avoid situations where it hallucinates or brings in unwanted bias.

It was nice to meet Einstein and hear the story behind. I believe the story will continue towards creation of even more clever tool to help businesses and people working within them. Also story from this Boardman Oy visit to Silicon Valley continues.