In partnership with

⬇️ Inside this issue:

• Your first AI agent might be easier to build than you think

• Thania shares 5 ways to explain your product so people actually get it

• Skills-first is growing, but 1 in 5 jobs still require a bachelor’s degree

INTERESTING READS

🎯 Smaller employers want skills based hiring too, but they need a model that works for them.

🪪 LERs have plenty of promise. What will actually get people to use them?

🤖 Workers are learning AI faster than employers are training them, and the gap is growing.

🌱 Employees want to grow. Are employers giving them room to?

TECHNOLOGY

How to Build an AI Agent Without Being an AI Expert

“AI agent” is having a moment. It’s also one of those terms that seems to be everywhere while somehow remaining a little fuzzy.

The basic idea is simpler than the terminology makes it sound. An AI agent is something you give a job to, rather than something you have to keep asking for help. Give it a goal, instructions, and access to the information or tools it needs, and it can work through the steps required to get there.

Think about asking ChatGPT, “What happened in LER news this week?” every Friday versus creating something whose job is to find relevant LER news, decide which stories matter, summarize them, and bring the results back.

That second one is the agent.

So where does someone actually build one? Thankfully, building an agent doesn’t necessarily require coding. Platforms now let people build agents using plain-English instructions. OpenAI, for example, offers Workspace Agents for eligible work accounts.

The exact buttons will depend on the platform—and will probably change anyway. What matters more is knowing what job to give an agent and how to teach it to do that job.

Here’s how to start.

Step 1: Pick one job

The easiest place to find a first AI agent probably isn’t in an AI tool. It’s in the to-do list.

Look for something that happens repeatedly and follows roughly the same steps each time. That’s a good place to start because agents are especially useful when they can work toward a clear goal using a defined set of instructions and tools.

That could be researching new developments, reviewing credentials against the same requirements, turning course information into credential descriptions, or translating complicated LER language for employers.

For example, the SkillsScoop editorial team is constantly looking for new stories in the LER world. So an agent could have one specific job: Every week, find new LER and skills-based hiring stories that could be relevant to SkillsScoop readers, explain why each one matters, and flag the three most interesting ones to consider.

That’s much easier to teach than simply telling an agent to “help with LER research.”

Step 2: Give it what it needs

Imagine hiring someone to take over that job. They’d need to understand the organization, its audience, what it cares about, and what good work looks like.

An agent needs that context too.

For example, the SkillsScoop agent could be told that readers work across HR, higher education, workforce development, philanthropy, and credentialing. It could be given examples of past stories the team has covered, topics to prioritize, and publications the team trusts.

A different agent would need different information. One creating credential descriptions might need the organization’s requirements and examples of previously approved credentials.

The goal isn’t to give an agent everything. It’s to give it the right information for its particular job. Sensitive or confidential information should also stay out unless the organization has approved its use with the AI tool.

Step 3: Show it how to do the job

“Prompt engineering” makes this sound more technical than it needs to be. Think of it more like onboarding a very fast new coworker.

For the SkillsScoop agent, instructions might say:

The instructions probably won’t be perfect the first time. If the agent keeps finding stories that aren’t useful, tell it what’s missing and try again. If every explanation amounts to “organizations should continue monitoring this trend,” congratulations—it has discovered AI’s favorite way of saying absolutely nothing.

Step 4: Build it, then test it

Now it’s time to put those pieces into an agent-building platform and see what happens.

Before giving the agent more freedom, run it manually a few times and check how reliably it does the job. Are the sources real? Are the stories actually relevant? Does it understand why they matter? Would someone use what it produced?

Think of the agent like a very smart intern on their first week: fast, helpful, capable.

But, not before another set of eyes reviews it first, before sending it off to your CEO.

Start small and see where it takes you

Once the agent works reliably, some platforms can also have it run automatically. That weekly SkillsScoop research, for example, could eventually happen every Monday without someone remembering to start it.

But there’s no need to build the AI Avengers on the first try. A useful first agent only needs to make one annoying, repetitive job a little easier. Start there, test what works, and build from it.

Because the best first agent may already be hiding in the task that makes everyone think, “Didn’t we just do this?”

COMMUNICATION

How to Communicate Your Product Better

Since day one at Skills Scoop, my goal has been to take something as complicated as LERs and make it simple enough for anyone to understand. That is the hard part, because our product is something you can't physically touch, and most of us trying to explain it work deep inside HR, digital credentialing, or higher ed.

The LER industry, like most disruptive industries, has a communication problem. We know the technology cold and still struggle to tell the story simply. I've talked about this at Badge Summit, and my rule of thumb is this: if you can't explain your thing to your grandma or a five-year-old at the Thanksgiving table, you've got work to do. Fixing it is a product marketing exercise, and it comes down to five rules.

1. Know your audience

Good communication starts before you write a word, with knowing exactly who you're writing for. Close your eyes for five seconds and picture one real person you're trying to reach. Maybe it's the plumber whose credentials live on paper, the student figuring out what comes after graduation, or the veteran translating military service into civilian skills. Ask what feels intimidating to them and what they genuinely care about on a random Tuesday. In product marketing we call this positioning, and it always begins with the buyer. Write for one person, and thousands will feel like you wrote it just for them.

2. Forget yourself

This is the rule we break the most, because it's the hardest. We love to talk about ourselves: our company, what we built, how great the team is. None of it lands, because the person on the other end is thinking about their own life and their own next move, exactly as they should be. It gets worse once you understand your product deeply, because you forget what it felt like to be confused by it. You reach for acronyms and lead with the infrastructure and the interoperability, and you lose your reader in the first sentence. So make your reader the hero and treat your product as the sidekick that helps them win. Once the spotlight is on them, the only question left is which of their priorities to lead with.

3. Lead with the benefits they care about

Every person you talk to is really asking one question: what does this do for me? In our world, the answer usually comes down to four things: Will this help me get hired? Will this save me time? Will this make my skills visible? Is my information safe? This is where most of us slip, because we describe features when people buy benefits. Features are what your product is made of. Benefits are what changes in someone's life because of it, and benefits are what people actually buy. So name the outcome specifically: a job, two hours back in their week, the feeling of being recognized for skills a résumé never captured. Give people the payoff in the first line and make them want the rest.

4. Storytell, don't whitepaper me

A spec sheet informs, but it rarely sticks. People remember how you made them feel and the picture you put in their head, which is why a story will beat a feature list every time. The practical version of storytelling is translation: you take an unfamiliar idea and anchor it to something the person already understands, so the new concept clicks in seconds.

That is why an LER lands better as a digital backpack that carries everything you can do, including the skills a job title leaves out. Skills-based hiring makes more sense as the jump from VHS to Netflix, the same movies with a far better way to find the one you want. Neither explanation leans on a single piece of jargon, and both work because they start from something the listener already trusts.

One caution: respect your reader while you translate. They're smart, just busy and unfamiliar, so make the idea relatable without talking down. And if your draft reads like a white paper, rewrite it until it sounds like something you would actually say out loud.

5. Think outside the box

Disruptive industries win by capturing attention, and building great infrastructure while waiting for the world to notice has never been enough. So don't be afraid to try something new, and look outside your own industry for inspiration. Hand someone a one-page cheat sheet with a QR code instead of a forty-page guide. Give them something physical they can keep on their desk. Creativity and innovation are your real edge here, so take a risk, be different, and refuse to do it the way everybody else does. That is often what makes an invisible, digital thing concrete enough that someone remembers it a week later, when it finally matters.

The simplest things are the hardest

I know that it sounds simple, but sometimes the simplest tasks can be the most challenging. The whole art of good communication and storytelling comes down to simplicity, and simple is rarely easy, especially once you're fluent in the jargon and standing too close to your own product.

Like anything worth being good at, this is a practice. You will not land it every time, and that is okay. Keep these five rules close, keep writing for the one real person on the other end, and keep going.

The more you practice saying it simply, the more people will actually hear you.

BY THE NUMBERS

Employers still value college degrees

Despite the push toward skills-first hiring, 19.3% of U.S. job postings required a bachelor’s degree or higher as of November 2025, up from 16.6% two years earlier. Meanwhile, just over half of postings list no formal education requirement at all.

Source: Indeed Hiring Lab

KNOWLEDGE

What to read, watch, and listen to this weekend

📚 Read: Hard at Work: Job Quality, Wellbeing, and the Global Economy by Francis Green

What actually makes a job a good job? Green looks beyond salary to explore security, autonomy, wellbeing, and how the quality of work is changing around the world. A good one for anyone thinking about the future of work from the worker’s perspective for a change.

📺 Watch: How to Build a Career You Actually Love by Bill Gurley for TED

Career advice can get very “follow these five steps and voilà!” This talk takes a more human look at building a career around what you actually enjoy and value. A lighter watch that might get you thinking about your own career choices along the way.

🎧 Listen: WorkLife with Molly Graham from TED

WorkLife got a refresh this year, with former Meta and Quip executive Molly Graham taking over as host. The new season gets into the wonderfully messy human side of work—ambition, burnout, failure, confidence, leadership, and figuring out what a meaningful career looks like. Consider this your break from another podcast telling you how AI is changing everything.

FOR FUN’SIES

Hire anyone, anywhere — compliant in under 3 days

Found the right person, but they’re in a country where you don’t have an entity? Setting one up can take months and significant cost.

Remote removes that barrier by becoming the legal employer through our own entities — handling compliant contracts, local benefits, tax setup, and onboarding for you. In fact, an employee is onboarded to Remote every 7 minutes.

Once they’re hired, the same in-house teams that support employment locally also run payroll — so you’re not bouncing between disconnected providers. Less setup, less complexity, and less time between finding the right person and getting them started.

See how it works

FEEDBACK LOOP

💌 Have more feedback? Reply to this email and tell us. We read each one.

🗣️ Want to be featured as a contributor? We’re looking for industry people to write editorials. 

✊🏼 Want to sponsor or partner? Connect with us about advertising opportunities.