When news broke that Salesforce would be updating the Platform Administrator exam weightings starting December 15, the reaction was… well, loud. And if you’ve spent five minutes scrolling LinkedIn lately, you’ll know it wasn’t all celebratory.
At the center of the discussion is…you guessed it, Agentforce. More specifically, it’s the fact that Agentforce now has its own space in the weighted topics, accounting for 8% of the Salesforce Administrator exam. And that percentage didn’t appear out of thin air – it came at the expense of one of the most core, traditional areas of the exam: Configuration & Setup, which has seen its weighting reduced. Pair that with a higher passing score, and it’s no surprise that anyone currently studying for the exams (or has them already booked) is feeling unsettled. I would be too!
In this post, we’re unpacking why so many are uneasy and whether those concerns are actually justified. Others have taken a broader look at how this change fits into the evolving Salesforce Admin role, and Tim explores that view in more detail here.
It Has Always Been About Foundations
It is common knowledge that the admin exam has always represented the foundation of the platform. For years, it’s been the most reasonable and logical cert to ever take first, as passing the admin exam meant one thing: a clear promise that you understand and can demonstrate the core mechanics of Salesforce at a minimum level.
Topics tackled in the exam range from Configuration & Setup, Object Manager, Sales & Service, Reports/Dashboards, and Process Automation. All of these are core platform fundamentals. And while the exam tests for your theoretical knowledge, having a really good grasp of these topics is the bread and butter of most admin roles, especially at the entry and early-career level. Many spend months (sometimes even years!) building confidence in these areas before eventually branching out into more advanced or specialized products.
This is exactly why the shift in weightings feels jarring to some. Imagine reducing emphasis on foundational concepts like configuration, all while introducing a (relatively) new, premium AI product into the base-level certification?
It raises a fair question: Are we moving too fast?
“I Haven’t Even Used This Yet!”
One of the most repeated sentiments across the community is simple: some haven’t even used Agentforce yet!
To start with, unlike core platform features, Agentforce isn’t something every org has enabled. And while Salesforce has made the effort to make Agentforce more accessible with Foundations, the Agentblazer Statuses, and developer orgs with Agentforce enabled, along with Salesforce renaming Sales and Service Cloud to Agentforce Sales/Service, many admins still work in environments where AI features are still being evaluated, restricted, or simply out of scope.
Even from our latest Admin Survey results, Agentforce ranks very low at only 12.6% of respondents actually using the tool. That makes valuable real-world hands-on experience beyond Trailhead hard to come by.
The Bigger Worry Is: What Gets Deprioritized?
I’m not anti-AI. In fact, I use AI tools daily, both at work and in everyday life. It’s an incredible learning tool, a productivity booster, and when used well, I believe it can also be a genuine quality enhancer.
However, the quality of AI-assisted work still depends too much on the human behind it. Maybe it’s just me, but purely AI-generated work tends to feel hollow. However, in the Salesforce context, this isn’t really about whether AI-generated things feel hollow, but the stakes are higher because running an org smoothly is all about reliability, judgment, and grounding. Being so reliant on AI can potentially surface confident-sounding outputs that aren’t accurate, secure, or aligned with how the org actually works.
The story shifts with “skillful” use of AI – as in mindfully and intentionally. I believe that human work thoughtfully refined with AI can be stronger than either one by itself. An admin who deeply understands security, data quality, configuration, overall org design, or manually building automation will be much better equipped to use tools like Agentforce responsibly and effectively.
That being said, perhaps the most thoughtful critiques aren’t anti-AI at all. They’re about balance.
Many worry that emphasizing Agentforce implies that deep platform knowledge is becoming less important…that maybe, we’re heading into a world where it’s “up to AI” to handle setup and configuration for us. After all, understanding why something works has always been central to good practice. And these concerns are valid because if people continue to develop the kind of mindset that unintentionally starts shifting responsibilities to AI, things start to become dangerous:
Automation without context. Security without fundamentals. Decision-making without understanding. Shortcuts without accountability. I can go on and on…
At the end of the day, AI is still only as effective as the human using it, plus the data model and processes behind it. Once foundational knowledge starts to feel optional, we risk creating more problems than we intend to solve.
Right Direction, Questionable Timing
This gives me a slight feeling of deja vu – echoing the early days of mainstream AI tools, before ChatGPT or Gemini were even considered “everyday” terms. The technology arrived fast, and many people were left wondering whether this was genuine progress or just hype. So there was fear, skepticism, and plenty of debate about whether AI was here to make our lives easier or quietly replace us. In many ways, the reaction to the admin exam changes feels strikingly similar.
A large layer of the frustration is how and when this change is happening.
Historically, major shifts in Salesforce cert exams were signaled well in advance, giving exam-takers enough time to adjust their expectations and study plans accordingly. This update feels different – abrupt, even – in that despite the undeniable significance of AI and Agentforce, admins are being given a limited runway and support to prepare for what’s now being tested.
So if you’re studying for the admin exam right now, you’re not wrong to feel conflicted. The goalposts feel like they’re shifting, and I can imagine how uncomfortable that feels.
But when you really think about it, ask yourself, is there even a perfect time to push this change? After all, there won’t ever be a time when there are no candidates who have exams booked, right? There won’t ever be a time when there are zero students deep into study plans based on the previous weightings. No matter when you apply this change, there will always be people who feel like the adjustments are happening mid-flight.
Combined with the fact that not everyone is fully immersed in Agentforce yet, it’s easy for this to come across as Salesforce chasing momentum around AI rather than actually letting adoption mature organically. And now, especially with an increased passing score, the change can feel a little unfair.
Fair or not, this perception is fueling a lot of the negative reactions, and is also a valid point for people who feel this move is premature.
Where Does This Leave Aspiring Admins?
Salesforce is clearly signaling where it believes the platform is heading. The future is no longer one where AI is a side feature – it’s part of the long-term strategy. Including Agentforce in the admin exam sends a message that future admins are expected to understand AI concepts as a foundation.
The issue isn’t really that people don’t want Agentforce included, but it’s more about whether the admin exam is the right place (and the right time) to introduce it at this weight. As Sasha mentioned here, Salesforce Admins may not yet have all the skills needed to confidently build AI agents – but they’re arguably the best positioned to get there. When admins already understand how their orgs work end to end, the right training for building AI-powered tools becomes a natural next step, not a leap.
Some are even asking a harder question – one that feels uncomfortable to even say out loud: Does this shift risk producing less competent admins in the long run? It’s a hot take, I know. But it does not seem to be coming from a place of gatekeeping.
I see this as a genuine concern, and coming from a desire to protect the quality and credibility of the role. This can also tie into the evolution of the administrator role, where tools like Agentforce can increasingly take on more execution-heavy tasks.
Final Thoughts
Salesforce isn’t wrong to invest heavily in AI, but neither are admins wrong about feeling fatigued by how aggressively it’s being positioned. Rebrands, renamed features, and an ever-growing list of “AI-first” stuff…all these can make it feel exhausting to catch up.
But it’s also worth remembering this: the foundations still matter. Agentforce may be part of the exam now, but it doesn’t replace the skills that make admins effective in the real world. It was never about whether admins should adopt AI, but more about whether the ecosystem is being pushed there faster than those foundational skills are being reinforced.
The open question is whether exam weightings are evolving in a way that rewards admins who “know how to use AI because they understand the platform” rather than admins who “know a little AI and a little configuration, but not enough of either”.
The reaction that this change is getting from the community only makes it clear that people still care deeply about the integrity of the certification, and that can only be a good thing. But whether or not this change proves to be forward-thinking or premature will only be clear with time.
Comments: