Fullstack HR

Subscribe to the FullStack HR newsletter

Yes, you can connect your HR system, ATS, or engagement tool to Claude and ChatGPT. In this episode Johannes Sundlo walks through the exact steps to add an MCP server to both tools, using a safe read-only demo.

In this episode:

• What an MCP server unlocks for your AI, explained simply
• Adding a pre-made connector in Claude
• Adding a custom MCP server in Claude with a URL
• Why you should start with read-only access
• Enabling Developer mode and creating an app in ChatGPT
• Connecting and testing the server in ChatGPT
• Authentication, security, and why to bring IT in early

New to MCP? Start with the written primer that explains connectors and MCP from scratch:https://www.fullstackhr.io/p/connectors-and-mcp-explained-for

Links mentioned:

• MCP and connectors explainer → https://www.fullstackhr.io/p/connectors-and-mcp-explained-for

Subscribe to the FullStack HR newsletter → https://fullstackhr.io

OpenAI just ran an event called Intelligence at Work, and the releases say a lot about where the world of work is going. Here is what dropped and why it matters for anyone leading AI adoption.

In this episode:

• Why the bottleneck has shifted from model intelligence to a company's ability to absorb the tech
• The stat to sit with: 74% of AI's economic value is captured by 20% of companies
• Codex moving into ChatGPT, and what that means if you are on ChatGPT Enterprise
• Six role-specific agent plugins, including a data analytics agent that writes queries, builds charts, and produces a deck
• Annotations: editing one piece of an output instead of regenerating the whole thing, and whether it solves the "last mile" problem
• Sites: turning Codex outputs into shareable, live dashboards and internal apps (Lovable meets Artifacts)
• Why OpenAI keeps framing adoption as a human transformation, not an IT project
• Sam Altman on the next phase: proactive agents that fix things in the background before you ask

Links mentioned:

• OpenAI Intelligence at Work (the event) → https://openai.com/sv-SE/business/intelligence-at-work/

Subscribe to the FullStack HR newsletter

Performance reviews have looked the same since the 1950s, and they are still mostly task-based. AI breaks that. This episode asks what we should reward when the tasks get automated and the only thing left to value is human judgment.In this episode:- Why the task-based performance review is the wrong tool for the AI era- The judgment problem. If an agent does the work, how does an employee prove their value, and does the manager even see it- A real CHRO dilemma. Two standard performers became superstars with AI, two former high performers fell behind, and bonus season arrived- Rewarding the behavior of using AI versus rewarding valid, validated output- Why AI makes the manager's job harder, not easier- The hidden cost of a thumbs-up task that creates a ripple of rework downstream- Should we scrap the performance review, hand growth to AI plus manager, or build something new

Links mentioned:- Nate B. Jones, Nate's Substack

Subscribe to the FullStack HR newsletter → https://fullstackhr.io

Three and a half years into the AI shift, most leaders still know AI matters but have no idea how to use it. This episode is about why that gap exists, and why the answer is getting practical.

In this episode:

• Why leaders keep asking for agentic workflows before they've done the basic Copilot grunt work
• The uncomfortable question every AI educator should answer: are you in the weeds, or only working "strategically"?
• Why you can't set new expectations for your team until you understand the new way of working yourself
• What buyers should demand from anyone selling AI upskilling
• How to do this for free with YouTube and disciplined facilitation if you don't want to hire someone
• Why AI should serve your existing strategy, not become a separate "AI strategy"
• The one-gear Ferrari problem: everyone owns Copilot, almost nobody drives it past first gear

This podcast is hosted by Johannes Sundlo, one of Europe's leading voices on AI adoption in HR and the workplace. He helps organizations move from talking about AI to actually using it, through hands-on workshops, keynotes, and the FullStack HR newsletter.

Subscribe to the FullStack HR newsletter

An impromptu between-episodes take on the AI bubble panic. Johannes cuts through the headlines about rising token costs, Microsoft cutting Claude Code, and Uber's COO questioning AI ROI, and explains what's signal and what's noise.

In this episode:- Why nobody can actually call a bubble, and what the market is and isn't telling us right now- The arguments for a bubble (circular money between NVIDIA and OpenAI) versus against it (Anthropic's first profitable quarter)- Why "rising token costs" is a mislabel. Per-token prices are flat or lower, total spend is what's climbing- How Moore's Law underpins the whole thing: more powerful models for the same or lower price- The full context on Uber blowing its AI budget, and why the COO is far more nuanced than the headlines suggest- Why early subsidize-then-correct cycles are normal, same as railroads, electricity, and the internet- How to do the token economics yourself: if an agent solves a problem cheaper than a human at equal quality, you have your answer- Why the model companies are targeting a ~$58 trillion global labor market, and what that means for their valuations

Links mentioned:- Rapid Response podcast, Andrew Macdonald (Uber COO) interview - Fortune coverage of the Uber AI budget story

The short answer is yes.

But yes deserves an explanation. In this episode I make the case for why HR needs to get technically curious right now, and why that has nothing to do with becoming a programmer.

We cover what's worth understanding (APIs, databases, MCP servers, and why they matter for HR), why playing around with tools like Claude Code, Codex, Lovable and Replit makes you a sharper buyer of HR tech, and how one curious person on a team can shift an entire organization. It might not be the CHRO. It might be the HR assistant fresh out of university.

I share what it felt like walking into Spotify with close to zero backend knowledge, why the threshold for connecting systems has dropped, and why "I'm not the technical type" no longer holds as an excuse.

This is where the world is heading. You can ignore it, or you can skate to where the puck is going.

In this episode, Johannes Sundlo explores a growing issue inside organizations adopting AI: the creation of an “A and B team.”

Some employees get access to AI tools, licenses, education, and experimentation opportunities. Others are left behind. Not because of role or capability, but because of interest, visibility, or who happened to push hardest internally.

Johannes argues that this is becoming one of the most overlooked leadership and employer branding challenges in AI adoption.

He discusses:

• Why selective AI access creates organizational inequality
• The hidden cultural risks of AI “elite groups”
• Why the cost argument often falls apart under scrutiny
• The chicken-and-egg problem of AI adoption
• Shadow AI and the risks of unsupported tools
• Why HR cannot leave this entirely to IT
• How AI access is increasingly becoming a workplace infrastructure question, not an innovation experiment

This is not a conversation about hype or future speculation. It’s about what happens inside organizations right now when some employees are allowed to work with modern tools while others are expected to continue without them.

A practical reflection on leadership, fairness, AI adoption, and the future of work.

Block just fired 4,000 people in a single day. Their CEO Jack Dorsey said it was because of AI. The stock jumped 20%. And the internet lost its mind.

Half the people say this is pure hype. The other half say AI is coming for all of us. I think both sides are wrong, and I have thoughts.

In this video I go through what actually happened at Block, weigh the evidence for and against the AI narrative, talk about why Wall Street loved this so much, and share what I think it means for leaders, employees, and anyone trying to figure out where AI and jobs are actually heading. I also take on the myth that AI skills will protect you from layoffs.

What if an AI manager could do a better job than your current boss? Johannes explores this deliberately provocative question – not to argue that we should replace managers with AI, but to challenge our assumptions about what management is actually for.

From the industrial revolution roots of the manager role to DoorDash's algorithmic task distribution, he examines whether the core function of "making sure work gets done" might be ripe for reinvention.

If AI can handle delegation, and some managers already lack the empathy we claim only humans possess, what should human leaders actually focus on?

A call to start the conversation about leadership's future before it's decided for us.

What happens when a four-week project takes 45 minutes? That's what I experienced over Christmas with Claude Code. In this episode, I share what I built, why this matters beyond coding, and the hard questions organizations need to start asking. Because this isn't slowing down.