SFDC Penguin: AI Now Tour London Cover

AI In Action: 3 Key Lessons from the AI NOW Tour

This past year has seen Salesforce make rapid progress in its AI offerings. Alongside this, Salesforce World Tour – London showed how Data Cloud can help enrich prompts to generate highly personalised outputs. With all this change, you may be wondering where to start? If so, read-on about Salesforce’s AI Now Tour! 

What is the Salesforce AI NOW Tour?

The AI NOW Tour is part of Salesforce’s goal to train over 100,000 developers in the use of its AI tools. Main topics include Einstein Copilot, Prompt Builder and Data Cloud. You’ll also get the chance to use Code Builder, Salesforce’s web-based integrated development environment (IDE). Current and upcoming locations of the AI NOW Tour are shown in the map below. 

Summary map of Salesforce AI NOW Tour Locations.

Current and upcoming locations for Salesforce's AI NOW Tour.

Who is it intended for?

The session is geared towards Admins, Developers, Consultants, Architects, IT Managers and Business Analysts. All you need is a laptop and familiarity with the setup of the Salesforce platform. The latter is important as you will spend most of the day navigating Salesforce and setup. 

How much does it cost?

I attended the event on the 18th of June, as part of the opening of Salesforce’s first AI Centre. The event was free to attend. Be sure to check the official website for more information. 

My 3 Key Lessons from the day

Here are my top 3 lessons that I took away from the AI NOW Tour in London. 

1. A greater (practical) understanding of how Data Cloud and Einstein 1 work together

The main take-home was a more practical understanding of how Data Cloud and Einstein 1 work together. For example, you start by connecting Data Cloud to Amazon S3. After this, you ingest data and create a unified profile. Once completed, Data Cloud and CRM data is used to enrich prompts, which are then used in Copilot.  

Examples such as these show how the tools work together to create an end-to-end solution. It also highlighted how specific components, such as Flex Prompt Templates, can be reused.  

2. We're on a journey: considerations and upcoming improvements

As mentioned above, Salesforce has made significant progress with its AI offerings. However, as with any  innovation, there are trade-offs and upcoming improvements. 

The AI NOW Tour provided valuable implementation guidance and signposted key upcoming enhancements. Some of the main items which caught my attention were: 

  • Processing efficiencies. Salesforce is continuing to refine its generative AI performance. For example, reducing round trips to Large Language Models (LLMs). 
  • A single Copilot per org. Currently, an org can have only one copilot. This can be challenging if you need to deploy in different contexts (e.g. internal and public).  
  • Care when testing. Traditional software testing checks outputs are as expected for certain inputs. However, Generative AI can be variable, both for inputs and outputs. Care must be taken with testing any model. 
  • The sensitivity of prompts. Being able to use Natural Language to generate a prompt is powerful. However, during the workshop, there was an example which showed how changing the capitalisation of a word impacted the output (i.e. ensuring capitalisation is applied when referring to the name of something). 
With all of these items, potential feature enhancements will help in future. Nevertheless, it still pays to learn and explore the features now. By doing so, it will help embed the approaches and maximise the opportunities, both now and in future.  

3. The Importance of Descriptions!

Anyone who has worked on Salesforce long enough will know the importance of descriptions. They are now even more important with Generative AI. The reason for this is that, when copilot receives a prompt, it must interpret this and create a plan from the various actions available. To do so, the descriptions are key; as this is how the ‘Planner Service’ will determine which actions are needed. 

Screenshot taken from Trailhead outlining Einstein Copilot's Reasoning Engine (Planner Service).

Screenshot taken from Trailhead outlining Einstein Copilot's 'Reasoning Engine', referred to as the 'Planner Service'.

Bonus: Code Builder

The workshop includes various exercises where you will need to deploy code. For this, you will need to use Code Builder. Up until now, I have used Code Builder for testing when reviewing the Release Notes. The AI NOW Tour provided a great opportunity to use it more! 

Despite technical difficulties on the day (not related to Code Builder!), it was quick and easy to setup. That said, there were moments where Code Builder was slow to load/respond. This may have been due to internet connectivity. However, I believe there may be scenarios where a local installation of Visual Studio Code (VSC) may have advantages. 

Would I recommend the AI NOW Tour?

If you are a Salesforce professional looking to get hands-on with Salesforce’s latest AI and Data Cloud offerings, I would recommend attending. However, you may wish to check out some of the other resources ahead of the session. This will help familiarise you with the content and provide context for the options selected when working through examples. 

Where can I learn more?

The session covered plenty of content. If you want to learn more, I’d recommend checking out the following:

Summary

Salesforce is continuing to invest and innovate with its AI offerings. The AI NOW Tour provides a great opportunity to get hands-on with Data Cloud, Einstein Copilot and Prompt Builder. If you wish to learn more, be sure to check out the resources above.

Until next time, thanks for reading! 🐧

Bonus Penguin Fact 🐧

We often see penguins swimming in water. However, even penguins must learn how to swim! At London Zoo, eleven Humboldt penguin chicks that hatched in April recently had their first ‘swimming lesson‘. After hatching, it typically takes 6-12 weeks for their original downy feathers to be replaced by two layers of waterproof feathers. These feathers help to keep the penguin warm whilst swimming and diving into cooler waters.  

Learn more about London Zoo’s Humboldt penguin colony (including penguins named after Games of Thrones characters!) here

Want More?

Found this article useful? Why not share!

Facebook
X
LinkedIn
Email
WhatsApp
Pocket
Reddit

Or check out some of our other articles:

Leave a Comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.