How the Salesforce Architect Role Is Evolving in the Agentic AI Era

There is a common anxiety that as artificial intelligence (AI) becomes more capable, the need for human expertise diminishes. This fear of losing ground to AI is not unique to architects; AI-powered analysis and diagramming tools, design agents, and coding assistants are changing the job descriptions and daily responsibilities of admins, developers, designers, and architects alike. 

As AI becomes more skilled at tactical tasks, this has sparked a much-needed conversation about the skills, oversight and value that humans bring to the table. Many of these conversations have been happening inside the Salesforce ecosystem. Recently admin, developer, and architect evangelists joined a panel at Cactusforce to share their thoughts. The insights and ideas shared there inspired the following exploration of the architect’s evolving role.  

Human skills are an architect’s superpower

A recurring theme was that while the arrival of AI into our workflows and day to day sparks role discussions everywhere, this also creates opportunity. With the ability to accelerate tasks and gain rapid insights with the help of AI, the unique superpowers of the architect are the human skills required to translate those efficiencies toward a meaningful outcome. As technology continues to evolve, the success of an implementation remains dependent on the humans at the helm.

The conviction that human expertise is the secret sauce of a successful implementation is the foundation behind many of our resources. These focus areas, explored in more detail below, are what allow architects to lead the way in this new era:

• Architectural Judgment (Think Like an Architect live series and blog): Sharpening the skills needed to think through complexity, trade-offs, and long-term impact. It’s about the ability to break down massive business challenges into digestible, logical chunks.
• Communication (Talk Like an Architect): Strengthening the communication skills needed to align stakeholders and clarify requirements., In the agentic era, communication is the interface through which we “vibe” the future by providing unambiguous logic for both humans and AI.
• Critical Thinking: Acting as the ultimate filter to know when to challenge a requirement and what AI output to trust. This is the “relentless critical thinking” required to ensure we don’t automate a sub-par process or use generated content as is.

The power of architectural judgement 

Architects have always needed to stay up to date on the latest technology evolutions, but the rate of change has accelerated significantly. As new tools and technologies are released frequently, it is important to focus on the constants that are more stable while at the same time being able to embrace innovation. That means investing in strong decision-making skills, maintaining the depth to assess innovations and sharpening the judgment to decide what to adopt and when.

Staying current can be challenging, but a holistic understanding of the fundamental components underlying business processes helps architects balance high-level strategy with low-level detail. Because new innovations almost always build on core architectural concepts, understanding these “first principles” is the best way to anchor new information. By tying new innovations to well-understood principles such as the Well-Architected Framework, you can navigate rapid change without losing your bearings. Similarly, new innovations are easier to understand when built on the knowledge of the fundamentals of Salesforce architecture. 

Using AI as mirror for the clarity of your requirements

AI assistants highlight the importance of precise requirements. If a prompt is too vague or generic, it immediately becomes apparent in the quality of the output. AI coding assistants like Agentforce Vibes help us evaluate the quality of our requirements and instructions more quickly and easily. If the description of your requirement isn’t good enough for an agent to create a solid first draft, you’ll need to refine it until it is. Important information might be missing, the ambiguity might be too high, or your needs too vague.

Leveraging AI as a “second pair of eyes” to validate your thinking accelerates the design process and improves clarity. This clarity will go a long way in preventing issues during implementation and preventing surprises with your stakeholders.

The need for communication skills is no longer “limited” to talking to stakeholders and translating business needs into technical designs. These communication skills are even more critical as architects must express requirements clearly, precisely, and without ambiguity so AI systems can reliably interpret and execute them. 

Like humans, AI can be confidently mistaken. That’s why it’s important for architects to ensure clarity: the quality of the input is the most important predictor for the quality of the output. And once you have clarity on what you want to achieve, it will also become easier to understand if the output generated is valuable and relevant.

Rubber Duck 2.0: Using AI to iterate on solution design

With a laser-sharp understanding of the requirements, architects can leverage AI to ideate and iterate on solution design options. The Rubber Duck technique is alive and well, only now the duck talks back. Just as developers have been explaining problems step-by-step to rubber ducks for years to help them debug code, architects can leverage AI as a sparring partner for solution logic considerations and trade-offs. Instead of the rubber duck, now our AI agents and tooling can provide instantaneous feedback on our thought processes, transforming the exercise into an active dialogue.

The same steps you take when walking your stakeholders through a process or design are the ones you take when walking your AI assistant through them. Starting from the challenge, you break the problem down into smaller pieces which are easier to understand and solve while making sure the solution keeps making sense as a whole. This requires a good understanding of business needs on the one hand and the platform’s capabilities and best practices on the other.

Thinking that keeping up with technology and continuously learning about platform innovations is a waste of time because AI is faster at looking up this information or knows better would be a mistake. Leveraging AI in your solution design process requires having the knowledge and expertise to evaluate its output. If you’re not able to assess the output, you shouldn’t be delegating this task to AI. Never blindly trust what AI is telling you, but use the right tools for the job.

An architect’s holistic lens on the enterprise

The ability to leverage AI to iterate on a single solution is powerful, but it is only one piece of the puzzle. In the agentic era, there is a risk of building too many right things that don’t talk to each other. And the same goes for agents.

For companies to transform into Agentic Enterprises, it is not enough to build a standalone agent. The increased “autonomization” of agents requires a holistic, architectural lens to prioritize scalability and reusability. Optimization only happens when architects understand the business processes and identify where specialized agents can add value.

This type of process orchestration leveraging agents requires a thorough understanding of technological advancements like MCP and MuleSoft Agent Scanner to create a long-term, scalable strategy. More than ever, architects are the bridge between the what and the how and will need to rely on their communication skills, relationship building skills and an awareness of the latest technology to ensure a cohesive, successful outcome.

This holistic oversight requires deep collaboration between architects and their stakeholders. Architects must become involved earlier in the process to help business stakeholders identify the right problems to solve and ensure they fit into a broader architectural strategy. Simultaneously, business stakeholders are becoming more closely involved in the delivery process as the nuances of agent communication and output are more subtle than traditional “pass or fail” test results.

The future belongs to architects who can adapt

The story of the evolving architect is one of adaptability and adding value by mitigating risk in a high-velocity environment.

As the role continues to evolve, our most valuable assets are our transferable skills: 

• The problem solving skills to break down challenges into digestible chunks
• The communication skills to communicate with both human stakeholders and AI
• The critical thinking skills to know when to challenge and what output to trust 

To lead in the agentic era, architects must double down on the professional standards that automation cannot replicate:

• Requirement Precision: Using AI to pressure-test the thought process ensuring a business need is fully articulated. The architect’s value lies in bringing the critical thinking required to challenge a “requirement” until it is crisp, unambiguous, and technically sound.
• Technical Authority: Depth of knowledge is the only viable way to distinguish good output from nonsense. An architect must possess the technical depth to audit AI-generated solutions and the professional judgment to reject suggestions that fail to meet enterprise standards for security and reusability.
• Holistic Oversight: The architect maintains a holistic lens on the enterprise. This means identifying where in business processes agent optimization is likely to succeed and ensuring that new automations follow a unified strategy rather than becoming a collection of disconnected silos.

Architecting the Agentic Enterprise

As the architect’s role evolves, the systems we build are also changing. Given the innovation and the continually changing scope, sound architecture is more important than ever when covering designs for: 

• Multi-Agent Orchestration: Designing how independent AI agents collaborate, communicate, and share data and context.
• Autonomous Logic & Handoffs: Defining the critical boundaries where an agent acts independently versus where a human must be brought into the loop.
• Strict Systemic Guardrails: Building the safety that ensures autonomous actions remain within corporate policy and ethical bounds.
• A Robust Data Foundation: Ensuring the underlying data is clean, governed, and grounded to prevent hallucinations and ensure agents operate on trusted context.

Adaptable architects are essential to designing the architectures that enable the Agentic Enterprise. 

Start growing the skills you need for the future today

Take the next step to strengthen your skills and grow your impact. 

• Join the Think Like an Architect livestream series to see architectural thinking in action. Learn a repeatable framework you can use to think through complexity, trade-offs, and long-term impact– not just what to build, but why.
• Explore the Talk Like an Architect blog series to sharpen the communication skills that align stakeholders and clarify requirements to drive architectural success.

The most impactful architects will harness AI to move faster without sacrificing clarity. Those who thrive will keep their finger on the pulse of technology and their hearts in problem-solving and stakeholder connection.