Walk into any admin community right now, and you’ll hear the same nervous undertone. Everyone’s asking some version of the same question: what happens to my role when AI can build the flow, generate the report, and clean the data faster than I can?
In the Admin Keynote at Dreamforce ’26, Gordon Lee, managing director of technology at Co-op Careers and a Salesforce admin of nearly two decades, gave the clearest answer. He didn’t tell the room that AI is nothing to worry about. He told them where they should focus their attention. And the line he built his whole segment around is worth writing on a sticky note and pasting on your monitor: delegate the work, but don’t delegate the judgment.
That reframes the entire AI conversation for admins, so let’s get into what he actually meant, and what to do with it on Monday morning.
Table of Contents
The Real line between Good Admins and Great Admins
Lee has watched this ecosystem for twenty years, through workflow rules, Process Builder, Flow, and now agents. When the host asked him which admin skills have mattered most across all that change, his answer wasn’t technical at all.
He said that two things always distinguish a good administrator from a great one: people skills and knowing when to delegate tasks.
That’s a little uncomfortable to hear if you’ve spent years grinding on certifications and Trailhead badges. But think about the admins you actually respect. It’s rarely the ones who memorized the most obscure governor limit. It’s the ones who could sit with a frustrated sales manager, figure out what was really bothering them, and translate that mess into something the platform could solve. The tech was table stakes. The judgment was the job.
The 90/10 flip every Admin eventually makes
This is the part I wish someone had told me early on.
When you start, Lee said, you spend roughly 90% of your energy on the technology and 10% on the people – and that’s correct. This is a genuinely technical platform. You need to get your confidence and your “sea legs” underneath you before you can do anything else. Nobody trusts an admin who can’t build the thing.
But once you have that foundation, you flip the script. “90% people, 10% technology”.
Why? Because AI is a tool, and tools are used by humans. The more deeply you understand the people you serve, the more useful that tool becomes in your hands.
| Career stage | Where your energy goes |
|---|---|
| Early admin (getting your sea legs) | ~90% technology / ~10% people |
| Experienced admin (post-confidence) | ~90% people / ~10% technology |
If you’re a few years in and still living almost entirely in Setup, this is your nudge. The technical fluency you built was the entry fee, not the destination.
Delegate the work, but don’t delegate the judgment
This is the heart of it.
Lee has a perspective on AI. He views it as an assistant that is truly expert at technical tasks, so leave those tasks to it. Have it draft workflows, suggest formulas, and summarize case histories. After all, these are precisely the kinds of tasks you want to offload.
What you don’t hand over is the judgment. Admins carry experience and context that AI doesn’t have: knowing the “quick fix” the VP wants will break three downstream automations, or that a field the team swears they need will sit empty within a month because you’ve seen that exact request fail twice before.
So the split is clean. Delegate the work. Keep the judgment.
A practical approach: view the AI’s output as the first draft from a fast, tireless junior teammate who knows nothing about your organization’s history. This gives you something to react to immediately. Your job is to determine whether or not it is actually right for your users.
You don’t have to master every new tool
One of the more reassuring things Lee said was aimed at anyone feeling behind: you do not need to become an expert in every technology that ships.
And things ship constantly now. Trying to be fluent in all of it is a fast route to burnout. What you need instead is a mindset – curiosity and ownership.
He borrowed a line from Ted Lasso here that landed with the room: be curious, not judgmental. His observation was pointed. Everyone is googling their AI strategy. Everyone is frantically trying to figure it out. But very few people are the ones actually leading their organization through that uncertainty. The gap between the frantic googlers and the calm leaders isn’t knowledge – it’s perspective.
Curiosity scales. Trying to memorize every release note does not.
Start Small, Iterate, Expand – the Admin’s leadership playbook
So how do you go from overwhelmed to leading? Lee gave a sequence that’s almost embarrassingly simple, which is why it works.
- Solve a problem for yourself. Find one annoyance in your own daily workflow and fix it. Learn the tool on something low-stakes where you’re the only person affected.
- Solve that same problem for a teammate. Take what you learned and apply it to someone next to you. Now you’re helping one other human.
- Solve it for the colleagues who’ve been complaining for months. Scale it to the people who’ve been quietly suffering through that broken process the whole time.
His summary of the pattern: if you learn how to empower yourself, you’ll learn how to empower others.
I like this because it removes the pressure to have a grand AI transformation plan on day one. You don’t need a roadmap. You need one small win, then the next, then the next. That’s how great admins have always operated – start small, iterate, expand. The tooling changed. The method didn’t.
The ownership advantage (and what everyone can steal from it)
Lee made a point about his nonprofit background that applies far beyond nonprofits.
For Nonprofit admins, he said, they often work with tighter budgets. Because resources are thin, they end up owning the whole process, from step A to step Z. Yet, there is a hidden advantage: when you handle the whole process yourself, you gain visibility into what works and what doesn’t at every stage. The admin who only touches one siloed piece never gets that visibility
That end-to-end view lets you connect the dots, and connecting the dots is what lets you connect with the people around you and become more useful than your job description suggests.
You don’t have to work at a nonprofit to get this. If your role has you boxed into one narrow slice of the org, go find out what happens upstream and downstream of you. Sit with the team that receives the records you create. Ask what breaks. The more of the full picture you hold, the better your judgment gets – and judgment, remember, is the thing you’re never delegating.
Mistakes to Avoid
A few traps that follow directly from Lee’s advice:
- Staying in 90% technical mode forever. Getting your sea legs is step one, not the whole journey. If you never flip to the people side, you cap your own value.
- Delegating the judgment along with the work. Accepting AI output without checking it against your org’s real context is how you end up cleaning up a mess later.
- Trying to master everything. Chasing every new feature is a burnout machine. Cultivate curiosity and ownership instead of encyclopedic recall.
- Waiting until you feel “ready.” Nobody hands you readiness. You build it by solving one small problem, then a slightly bigger one.
- Staying in your silo. If you only see your slice, your judgment is working with half the information.
Your next step
Pick one problem this week something small and annoying in your own workflow and solve it with an AI-assisted approach, then review the output with your own judgment before you ship it. That’s the whole playbook in miniature: delegate the work, keep the judgment, and let the wins compound from there.
Frequently Asked Questions
No, AI is very good at technical execution, building, drafting, and summarizing, but it lacks the experience, context, and judgment an admin brings. As Gordon Lee put it at Dreamforce 2026, delegate the work but not the judgment.
People skills and knowing when to delegate. Technical ability gets you in the door; understanding the humans you serve is what makes you effective as AI takes over routine execution.
Early in your career, spend about 90% of your energy on technology and 10% on people to build confidence. Once you have that footing, flip it – 90% people, 10% technology.
Treat AI as a skilled assistant for technical work and a first draft you always review. Keep decision-making judgment firmly in human hands.
Priya Rastogi
Priya is a 3x Salesforce Certified who believes in the power of continuous learning and collaboration. She’s passionate about exploring how Salesforce can simplify work, boost productivity, and create better user experiences. When she’s not experimenting with new features or automating processes, Priya enjoys connecting with fellow Trailblazers and sharing insights to help others grow in their Salesforce journey.
- Priya Rastogi
- Priya Rastogi
- Priya Rastogi
- Priya Rastogi
