5. Meet Tom
● Mid-level Salesforce Developer
● Works for a small real estate firm
● He has been asked to write a new utility class
● So as a good SF developer he:
○ Checks out Trailhead for the best practices
○ Spends 1 day writing his class
○ Spends another day writing the test class and manages to get to 76% coverage
○ Whew! Job Done
6. Why Code Generation Matters
1. Code generation significantly boosts developer productivity
2. Code generation improves code quality and consistency
3. Code generation accelerates innovation and time-to-market
4. Code generation democratizes access to development skills
So, let’s have a look at some of the main code generation tools available to us
as developers today.
7. ChatGPT
What is it?
● A powerful large language model (LLM) chatbot developed by
OpenAI
● Can generate text, translate language , write different kinds of creative
content, and answer your questions in an informative way
● Generates creative, human-like code snippets
● Understands natural language instructions
● Can adapt to various coding styles and patterns
● Can provide inaccurate or misleading
code especially for complex Salesforce-
specific scenarios
● Requires careful validation before
deployment
● Still under development, reliability can be
inconsistent
Pros Cons
8. GitHub Copilot
What is it?
● AI-powered code suggestion tool developed by GitHub and
OpenAI
● Integrated directly into your code editor (e.g., VS Code)
● Suggests contextually relevant code in real-time
● "Learns" your coding style over time
● Works well for common, repetitive code
patterns
● Seamless VS Code integration
● May not grasp complex Apex logic or
Salesforce-specific nuances
● Can inadvertently introduce bad coding
habits if not carefully reviewed
● Costs $
Pros Cons
9. Einstein for Developers
What is it?
● Salesforce-native suite of AI-powered development tools
● Includes features like Code Generation, Code Completion, and Code
Recommendations
● Tailored to Salesforce, understands platform-specific
details
● Integrates seamlessly with Salesforce development
workflows (VS Code, Salesforce DX)
● Improves code quality and adherence to best
practices
● Less flexible in creative code generation
compared to ChatGPT
● May require some initial setup or
configuration
Pros Cons
10. Which One to Choose
Key Considerations
● Complexities of Code
○ Simple tasks? Copilot might do
○ Complex logic? Lead towards Einstein
○ Want experimental solutions? ChatGPT
● Salesforce Expertise
○ Deep platform knowledge favours Einstein
● Integration
○ Want it within your workflow? Einstein or Copilot shine
11. The Future of code Generation
● The Big Picture:
○ Code generation is rapidly evolving
○ Expect even more sophisticated and specialized tools to emerge
● Focus on Learning:
○ Don't let tools replace core Salesforce development skills
○ Use them to augment your expertise, not as a crutch
● A Collaborative Future:
○ Imagine tools understanding code intent, refactoring intelligently...
15. Tom Catchup
● He has been asked to write another class
● He has just listened to this talk and is
inspired to try Einstein for Developers
● So, as a good Salesforce AI Developer he:
○ creates a query in Einstein for Developer that create a large piece of the code he requires
○ Fills out the specifics of the code
○ Asks Einstein for Developers to create the test class for him
○ Tweaks the tests to get them to pass
○ Jobs done in 1 day.
18. Brisbane Developer Group
Next Meeting
● Advanced Flow Concepts Every Developer Should Know
○ May 1 - 12pm, virtual with Tim Combridge
Session Topic Suggestions
LinkedIn Group
● https://www.linkedin.com/groups/14345387/