Episode 15 of Multipliers: How to Stand Out in a Stack of 300 Resumes

There is a TikTok trend Spencer describes in this episode: a video of a crowd of college students, all wearing the same mom jeans and white shirts, each one convinced she is unlike anyone else in the frame. The caption is something like “there is no other girl like me.” The point is not cruel. It is just accurate. From the outside, they are indistinguishable.

That is what a stack of 300 resumes looks like to a hiring manager in commercial real estate right now. Good schools, strong grades, some internship experience, maybe a financial modeling course. All of them look the same. And the traditional way through that stack has always been a warm referral: someone on the inside who knows your name and moves your resume to the top.

That is still true. But there is now a second path, one that did not exist eighteen months ago, and it is more accessible than a referral because you can build it yourself on a Saturday afternoon.

In this episode of the Multipliers podcast, Spencer Burton, Michael Belasco, and Sam Carlson work through the PwC survey finding that 91% of financial services executives are raising compensation for AI-skilled employees, what an autonomous agent actually is in plain English, and the exact steps a student or career switcher can take this weekend to build something that makes a hiring manager’s head spin.

• You might also enjoy: The prior episode on the four layers of AI that make it useful for CRE: Episode 14 of Multipliers: Unpacking the Layers of AI That Make It Useful for CRE

• Related: Build the technical foundation this episode references: The A.CRE Accelerator

Episode 15 of Multipliers: How to Stand Out in a Stack of 300 Resumes

All three co-founders are back this week. The episode was recorded on August 5th, which Spencer notes in passing before pivoting to the PwC survey he has been thinking about. Michael is preparing for a trip to Panama City at the end of the month, partly vacation, partly a family gathering with his wife’s relatives coming up from Peru, and partly a half-serious exploration of what a second location might look like.

The conversation opens with a fond detour into shared memories of Panama and Peru, including Sam’s story of standing on a structurally suspect bridge during a flash flood in Nazca while police drove by on loudspeakers telling them to get off, filming what would have been a viral video before it was stolen at gunpoint in Lima that afternoon. From there, Spencer pivots to what is actually on his mind: a survey from PricewaterhouseCoopers, a conversation with a rising junior interning at a great firm, and a question that he thinks more young professionals should be asking themselves right now.

The episode connects back to Episode 14, where Spencer and Michael broke down the four layers of AI. This one is more personal and more practical, a conversation about what to actually do with all of that, and how to make it visible to the people who are hiring.

Listen on: Apple Podcasts | Spotify

Why This Episode, Why Now

The PwC survey is the anchor. A thousand participants, director level and above, across major financial services firms including commercial real estate, all at organizations with at least five hundred million in revenue. These are not startups or early adopters. These are the large, conservative institutions that have historically been the slowest to move on technology. And 91% of them are raising compensation for employees with AI skills.

Spencer has been thinking about this in the context of his son, who is approaching twenty and asking about career paths. He has also been seeing it play out in his enterprise training work, where companies that have adopted Claude are discovering that the hard part is not the license. It is figuring out how to actually get ROI from it. The skill of taking AI and producing a measurable output is not widely distributed. It is in high demand. And right now, the bar for having it at a level that is genuinely differentiating is lower than it is going to be in two years.

The student Spencer spoke with this week is the representative case. Great school, strong real estate financial modeling background, solid internship. He heard about AI, thought he was doing okay with ChatGPT. When Spencer started asking specific questions, it became clear the student did not have the skills the market is going to increasingly demand. Spencer’s message to him: there has never been a moment where it is easier to differentiate yourself, and this window is not going to stay open indefinitely.

Episode Highlights

Here are the themes that stood out.

1. The PwC Survey: 91% Are Raising Comp for AI-Skilled Employees

The headline stat from the PwC survey is the kind of number that cuts through the noise: 91% of executives at major financial services firms are raising compensation for employees who have demonstrated AI skills. Spencer pairs this with the finding from an earlier Claude-assisted research session, where he asked the model to identify the five best careers over the next twenty years by earning power. The number one answer was effectively an AI engineer: someone who can wrangle AI to produce a worthwhile output in a specific domain.

The qualification matters. These firms are not raising compensation for employees who have a ChatGPT account. They are raising it for employees who can take the technology and produce ROI from it. That is a narrower skill set than it sounds, and it is the one that is genuinely scarce right now.

Michael’s observation from the entrepreneur’s perspective is the right complement to the survey data. As a small operator, he now has access to the kind of analytical and operational capability that used to require a team with a significant budget. The AI agents managing his cashierless store inventory, flagging low stock, analyzing pricing across Amazon, Walmart, and Costco, and pre-filling orders: that is a function that previously would have required hiring someone. Now it runs autonomously at a nominal cost. The entry-level jobs most at risk from this shift are also, as Michael notes, the ones where a person who learns to build and manage these systems becomes far more valuable than a person who just executes the tasks.

2. What an Autonomous Agent Actually Is, in Plain English

Sam asked the question that a lot of people in the audience are thinking but not asking: what is an autonomous agent, in plain English, for someone who is still mostly using AI as a better search engine?

Spencer’s answer starts with the default interaction: you open a session with an AI, you give it a prompt, it responds, the session ends. That AI knows nothing before you start and retains nothing after you finish. The harness creates the illusion of continuity, injecting memory chunks and context at the start of each session to make it feel like a relationship, but technically every session is a new agent with no prior knowledge.

An autonomous agent is different in two ways. First, it initiates work without being prompted by you. It has a heartbeat, a recurring schedule that wakes it up, reviews its objective, and figures out what it can do in the current session to move that objective forward. Second, it can be triggered by external events: an incoming email, a Slack message, a change in a connected data source. In the same way that a human employee’s workday is driven by a mix of proactive agenda and reactive response to what lands in their inbox, an autonomous agent operates on both channels simultaneously.

Spencer’s deal sourcing agent for Michael is the live example. Every morning, it scans for new industrial deals that match Michael’s criteria, surfaces them in a Slack channel, adds them to a pipeline report, and can underwrite any deal Michael flags with a simple reply. Michael does not prompt it. It wakes up, does the work, and reports back. That is the difference between using AI and being AI-native.

Sam’s podcast production example is the more accessible version: as soon as an episode finishes recording, an agent could grab the file, move it to the right folder, trigger the editing workflow, and distribute clips to the relevant platforms without anyone touching a keyboard. The task that currently takes a person time becomes something that happens automatically in the background.

3. The Resume That Makes a Hiring Manager’s Head Spin

Spencer’s answer to Michael’s direct question, what does AI-native actually look like on a resume in 2026, is the most actionable section of the episode.

The baseline is completing something like the A.CRE Accelerator and the AI.Edge certificate. Those establish that you have the foundational skills. But a lot of candidates will have those soon, and a certificate on a resume is a credential, not a demonstration.

The demonstration is what makes the difference. Spencer’s prescription is specific: build twenty-five agent skills in Claude. Create five connectors, at least one of which ties to a real data set. Build a data set yourself, which he says takes an afternoon, using publicly available information about the market or asset class you want to work in. Property tax records, lease pendants, foreclosures, utility data, zoning records, public hearing data: most of this is freely accessible and underused. Use Claude Code to connect that data set to an MCP server, and now your AI can query it directly.

Then, in the interview or in the message to your contact at the firm, say exactly what you built. Twenty-five agent skills that take an hour-long task down to ten minutes. A data set covering every industrial property in Dallas, connected to an MCP server your AI can query in real time. The hiring manager does not need to fully understand what an MCP server is. The fact that you built one, and can explain what it does, is enough to separate you from the stack.

Michael’s own story is the proof of concept. When he was coming out of grad school, Adventures in CRE had twenty visitors a month. Nobody knew what it was. He put it on his resume anyway, with a line that said he was an expert financial modeler and that you could verify it by downloading his work at the site. Hines went there, downloaded a complex mixed-use model for two towers in San Francisco, and hired him without a warm introduction because the work spoke for itself. The equivalent in 2026 is not a link to a model. It is a link to a live data set with a working MCP connector, built by you, queryable by AI, demonstrating that you understand both the domain and the tools.

4. The AI Champion Inside Every Firm

Spencer made a point that applies equally to people already inside organizations and to those trying to get in. At large CRE firms, the senior people often do not fully understand what the technology can do, and the junior people who do understand it are not being given the mandate to run with it. That gap is an opportunity in both directions.

If you are already inside a firm, being the person who bridges that gap is one of the highest-leverage moves available to you right now. Spencer was direct: if he were running a large operation and identified someone with real AI fluency, he would pull them off every other responsibility and point them exclusively at building the capabilities layer. That person has a job that did not exist two years ago and will be increasingly critical over the next five.

If you are on the outside trying to get in, demonstrating that you can be that person changes the conversation. You are no longer competing against 300 candidates for a standard analyst role. You are presenting yourself as someone who can give a brokerage firm twice the output from the same headcount, or give a small operator the analytical capability of a team three times its size. That is a different conversation, and it gets a different response.

5. The Cautionary Side: Are We Outsourcing Our Intelligence?

The second half of the episode takes a turn that is worth honoring, because it complicates the enthusiasm in a useful way. Michael raised the question of whether the same dynamic that gave us GPS and calculators and outsourced our ability to navigate and do arithmetic is now playing out at a larger scale with AI.

His example from his own week: he was working through a straightforward math problem about credit card points and almost immediately reached for Claude to solve it. He caught himself, did the math manually, and noted that the impulse to outsource even simple cognitive tasks to AI was already strong enough to have to consciously override.

Sam’s version is more vivid: his sixteen-year-old son did not know how to get to his grandmother’s house fifteen minutes away because he had always been a passenger looking at his phone while someone else navigated. Sam’s own childhood phone number memory, including Spencer’s home number from decades ago, is intact. His son could not reliably navigate without GPS to a house he had visited hundreds of times.

Spencer’s Panama story brings it home: he lived in Panama for years, knew every street, navigated without GPS. He moved to Miami and within months could not name the cross streets near his own house because GPS had made the mental map unnecessary.

The distinction Michael draws is the useful one: there is technology that extends what you can do and technology that replaces what you used to do and gradually atrophies the underlying capability. GPS replaced navigation. The calculator replaced arithmetic. AI is now positioned to replace a wide range of cognitive tasks. The question is which of those replacements are net positive because the capability was a bottleneck, and which ones quietly remove something worth keeping.

Spencer’s answer is that it is genuinely hard to manage. You do not notice the atrophy until it has already happened. The long division analogy he has been using throughout the series is the framework: you learn it by hand first so you understand why the output is what it is, and then you use the calculator. The risk of skipping the by-hand step is not that you cannot use the calculator. It is that you cannot catch the calculator when it is wrong.

---

The Bigger Idea

The 91% stat from the PwC survey is the episode’s anchor, but the insight underneath it is more specific: what those executives are actually paying for is not AI familiarity. It is the ability to take AI and produce a measurable output in a professional context. That is a narrower skill than it sounds, and it is the one that is in short supply.

The resume framework Spencer laid out is the most concrete version of this the podcast has offered. Build agent skills. Build a connector. Build a data set from public sources in the market you want to work in. Connect it to an MCP server. Then go into the interview and show your work. The whole thing takes a weekend. It does not require a computer science degree or a deep technical background. It requires the willingness to sit down and do it, which is exactly what most candidates are not doing.

Michael’s inventory management example and Spencer’s deal sourcing agent are both illustrations of the same underlying shift: work that previously required a human to initiate, execute, and monitor can now run autonomously, triggered by schedules or events, and report back without anyone touching a keyboard. The person who builds and manages that system is more valuable than the person who used to do the work manually, not because the manual work was worthless but because the leverage is an order of magnitude higher.

The cautionary thread is worth carrying too. The tools that extend capability are different from the tools that replace capability. GPS extended your range. It also replaced the mental map. The calculator extended your precision. It also replaced the arithmetic instinct. AI will extend your professional output. The question worth asking, before you outsource a cognitive task to it, is whether the underlying capability is something worth keeping. The long division principle is the answer: do it by hand first, understand why the output is what it is, and then use the calculator. That sequence protects the capability while gaining the leverage.

Sam’s closing is the right one to end on: a year ago, people were scared of AI. Today, it is a critical skill. The shift in framing from threat to opportunity has happened, and the professionals who act on that framing now, rather than watching and waiting, are the ones who will look different from their peers when the next hiring cycle comes around. The A.CRE Accelerator builds the foundation. AI.Edge keeps you current. And platforms like CRE Agents are what it looks like when the capability layer is already built and you just need to connect to it.

---

Frequently Asked Questions about Episode 15 of Multipliers: How to Stand Out in a Stack of 300 Resumes