AI – Threat or Opportunity for Salesforce Developers?
In with AI, out with developers?
Salesforce CEO Marc Benioff made headlines in December when he announced that Salesforce wouldn’t be hiring any software engineers in 2025. He credited a 30% productivity boost to Agentforce and other AI-driven tools, plus the introduction of an Agentic layer on Salesforce Help as his reasons.
That statement might leave some Salesforce developers wondering about their futures. Is AI going to help us get ahead, or leave us behind?
In this article we’ll look at where AI fits in the Salesforce ecosystem today, how developers can use it to their advantage, and what they can do to future-proof their careers in an AI-driven world.
The Rise of AI in Salesforce
What we today call AI – algorithmic systems built on deep learning, probabilistic modelling, and large-scale pattern recognition – has a more than decade-long history in Salesforce. Here’s a timeline of the highlights.
2014: Marc Benioff declares Salesforce an “AI-first company”
2015: Opportunity Scoring released, using machine learning to evaluate deals likely to close
2016: Launch of Salesforce Einstein, a cross-cloud suite of tools that applies AI techniques to enhance workflows and data
2023: Einstein GPT, Salesforce’s natural language processing tool, comes out on the heels on ChatGPT
2024: Debut of Agentforce, which deploys autonomous bots that use generative AI capabilities to perform domain-specific tasks across the organisation’s Customer 360.
Clearly, Salesforce’s AI offerings have come a long way and aren’t slowing down.
Salesforce AI – Where Do Developers Stand Now?
With all these AI advancements, developers – both programmatic and declarative – might wonder: Where do I stand?
Many of us have marvelled at the speed and quality of code GPTs can generate (more on that below). However, the current generation of AI tools isn’t perfect. Even the best models can suffer from hallucination, making up non-existent syntax and methods, and have to be corrected. That means developers still need to understand the languages and systems in which we work.
AI is also a long way from instigating, testing, and deploying code all on its own. Developers of all stripes still need to have a solid grasp of testing and DevOps principles in order to make sure business logic is reliable and only released when it’s ready.
AI also isn’t yet able to talk to managers, clients and stakeholders for us; we still need to be able to communicate clearly about our projects, plans, and progress. With big claims about AI’s capabilities floating around, being able to sell ourselves and our work is more important than ever.
How Salesforce Developers Can Use AI Today
Though AI can’t replace everything a developer does directly, it still offers some exciting opportunities to help us do our work even better, with appropriate care.
On the programmatic side, a tool like Salesforce’s Agentforce for Developers helps to complete code snippets, analyse classes, and draft tests. Many developers also use other Large Language Models like ChatGPT or Github Copilot to speed up work, but we need to be careful here – no developer wants to be accused of revealing proprietary code or private information. Agentforce uses the Einstein Trust Layer as a way of safeguarding data, but before using any AI tool for development, consider what guardrails and rules are in place for your organisation. If there aren’t any, consider creating some.
New opportunities also now exist for declarative developers. Einstein for Flow, generally available in the Spring ‘25 release, makes it possible to generate a flow from a prompt in a matter of seconds. By learning how to write good prompts, programmatic and declarative developers can both harness AI to get their tasks done faster and more efficiently.
As well as helping with development, AI tools can help us improve our workflows and communication. AI tools for calendar management, email writing, and project planning can free developers from the headache of admin tasks to let us concentrate on the parts of our work we like doing most. AI summary tools for calls and recordings can also help developers get relevant information more easily. Just be mindful of how your data could be used by these services.
Looking Ahead – What’s Next for AI?
That’s all well and good for now, but what can we expect for AI in the future?
Without speculating too wildly, in the short term AI is headed toward:
- Multimodal systems, which process audiovisual, textual, and contextual data together, e.g. AI creating a process diagram based on a call recording with stakeholders
- More sophisticated AI agents, which can automate more of the development process like testing and deployments
- Increased collaboration between humans and AI tools, with deep integration into our workflows and the systems we use (think Her, without the romance)
AI tools are likely only going to become more embedded into our work in the future, presenting fresh opportunities and unforeseen challenges.
Gearing up for the AI Future
So, how can developers prepare ourselves for an AI future?
First, we should familiarise ourselves with what AI can do right now. I’ve covered a few topics in this article, but getting the Salesforce AI Associate and Salesforce AI Specialist certifications is a great way to get a deeper understanding of the technology and its capabilities. In turn, that will help us decide whether the latest AI tool is a genuine advance or merely hype. These certifications are available for free on Trailhead through 2025.
Second, we can prepare by simply continuing to excel at what makes a strong developer strong—developing a deep understanding of programming principles, learning how to communicate clearly, and above all, bringing a consultative approach to the work we do. In a world that increasingly relies on AI, building human relationships will be more important and appreciated than ever.