The Inside-Out Employee Experience Principle
- drkimberlydunwoody
- Jul 25
- 5 min read
When people reach out to talk about my dissertation, they rarely want to talk about the whole thing. Fair enough. They usually want to talk about one line: your product or service experience is often just an outward expression of your employee experience.
I have never formally named that idea, mostly because I did not want to brand it unless I had a name good enough to deserve it. But for now, this will do: The Inside-Out Employee Experience Principle.
The idea is simple. If the systems your employees use are broken, your customers will eventually feel it. If the workflows are chaotic, the tools are outdated, the incentives are warped, or leadership keeps confusing cost control with operational health, that damage does not stay politely contained inside the building. It leaks. It shows up as bad service, bad follow-through, bad judgment, and a thousand little moments customers experience as incompetence but employees experience as Tuesday.
So the next time you get bafflingly bad service, try asking a different question than most people ask. Not just what did that person do wrong? Ask what they were working around. Ask what system failed them before they ever failed you.
This is also why, if you really want to understand a company's customer experience, you should spend a little time reading how its employees describe working there. Go read Glassdoor. Read the complaints about broken tooling, contradictory leadership, impossible workloads, missing follow-through, and endless process churn. Very often you will find the upstream version of the exact frustration customers feel downstream. The bad experience was not random. It was structured.
What is changing now is that this pattern is getting harder to hide, and it is showing up earlier. It shows up before someone is even hired. A site like Did They Ghost You? 👻 should not need to exist, but of course it does. It keeps a public record of companies that go silent after interviews, and yes, the pre-hire experience counts too. When a company cannot manage basic clarity, closure, and respect before you are even on payroll, that is rarely an isolated recruiting glitch. It is usually an honest preview of the conditions waiting behind the curtain.
The fantasy meets the invoice
Here is where the same dysfunction runs into a much louder story executives have been telling themselves.
Earlier this year, senior leaders at major companies were talking as if AI would finally let them replace knowledge workers at scale and free themselves from the inconvenient cost of human complexity. It was a tidy dream. Fewer people, fewer problems, better margins, and a very confident slide deck.
A few months later, the dream is meeting the invoice. Companies are throttling AI usage, capping tools, and discovering that large-scale adoption gets messy once the token bills start arriving. There is even a name for it now — the "Tokenpocalypse" — and the punchline is that it is often the non-engineers, quietly turning PDFs into slide decks, who are running up the meter. The Economist has documented executives openly worried that agentic AI costs could spiral out of control as every piece of software starts shipping its own bots.
None of this should surprise anyone. The same leaders who ignored the inside-out reality of employee experience are now colliding with the inside-out reality of AI economics. Indiscriminate AI adoption is not a strategy. It is a spend pattern. You cannot brute-force customer experience, candidate experience, or productivity gains from the outside if the human system underneath is weak.
The part I do not usually write
Now I have to say something I would not normally say this directly.
This is not only a story about service design, hiring habits, or AI budgeting. It is a story about drift — at the leadership level, and increasingly at the societal level too.
Over the last twelve months, too many people in power have run the same play again and again: dismiss the warning signs, wave away the obvious playbook, act genuinely shocked when it unfolds exactly as described, then overcorrect in the ugliest direction available. We saw it with Project 2025, a document plenty of executives insisted was not really a plan right up until it started being implemented as one. We are seeing versions of it in AI. And we are seeing it in the quiet refusal to admit that the United States is no longer offering the stability many companies still casually pencil into their forecasts.
I have written elsewhere about what I call the Dunning-Kruger Dark Age — the condition where confidence badly outruns self-awareness, and people with enormous power keep mistaking familiarity, status, or proximity to a trend for actual understanding. That pattern is expensive inside a single company. In a destabilized political and economic environment, it becomes reckless.
Stability is not a soft issue
Here is the uncomfortable part, and I mean it as more than a talking point: stability is a precondition for ethical AI.
We need institutional stability, workplace stability, and social stability to make AI ethical in any meaningful sense. A system built on fear, precarity, denial, and executive fantasy does not become responsible because you bolted a model onto it. AI can help human beings work better inside a stable and accountable environment. It cannot manufacture ethics in the absence of one.
So AI is not going to replace the need for functional institutions, working employee systems, honest leadership, or a livable future. It can help people do better work inside those conditions. It cannot compensate for their collapse. And no, despite what the more excitable corners of the media keep implying, we are not all moving to Mars to escape the bill.
The wake-up call
If you are in the C-suite, this is not the moment to keep your head buried in the comforting belief that workers are interchangeable, that the public will absorb any amount of instability without noticing, or that AI will smooth over the structural dysfunction you refused to fix when your own people were begging you to look at it. The reset is already underway. The only real question is whether you will learn from it before your employees, your candidates, and your customers decide they are done waiting.
A serious response looks different from the one most companies are running. It treats AI as a way to augment human capability, not erase it. It invests in the systems employees actually use, rather than filing internal friction under invisible overhead. It treats candidate experience as an early trust signal instead of an administrative afterthought. And it thinks much harder about what resilience now requires — including making it genuinely easier for knowledge workers to work remotely, even from outside the United States, when that is the safer, saner, more sustainable choice.
I have a book coming soon on knowledge graphs and human-centered AI design, and part of why I wrote it is that I refuse to accept the choice we keep being offered: anti-human nostalgia on one side, techno-utopian nonsense on the other. We can build AI that respects context, preserves human judgment, and makes systems more accountable rather than less. But only if we are honest about what the underlying problem actually is.
The underlying problem is not that humans are still here. It is that too many leaders still treat the human layer as the inconvenience, when it is the only layer that makes any of the rest worth building.
