Salesforce has been making big AI announcements at Dreamforce and the hype is at an all-time high. AI is everywhere, and at this point, every organisation is either developing or implementing their AI strategy.

As a Salesforce Architect, helping businesses be successful with AI means not only being aware of capabilities and features, but also understanding key risks, trade-offs and limitations.

Common AI use cases

AI is generally used for the following key applications:

‍

Key risks and considerations when using AI

AI gets better every day, however, at their core all AI algorithms are trained on data. Generally speaking, any issue that exists with the training data, may re-appear or be amplified by the AI algorithm.  

Be wary of the following considerations when using AI, which may help you identify key business risks.

‍

An example: Salesforce CodeGen

Imagine this scenario. Your Customer is working on a large digital transformation initiative and is excited to use Salesforce CodeGen to accelerate developments and reduce costs. They have done a Proof of Concept and are excited about using CodeGen as an “AI assistant” to quickly cross-skilling their developers on the Salesforce Platform, and are looking for advice on the matter. 

Key risks of using CodeGen

Now let’s use the framework above to see what material risks apply

‍

Salesforce’s Position on Trusted, Ethical and Humane AI

These are common concerns, and Salesforce are taking a position on the topic.

Conclusion

AI is exciting and transformative, however it’s important we understand, together with its potentials, its key risks and limitations to help customers adopt AI in a way that’s not only helpful, but also ethical and safe.