Most companies have a values page. Most companies do not have values. The two facts are unrelated, which is the problem.
A values project usually starts with a workshop, runs through three rounds of wordsmithing, ends in a poster, and quietly stops mattering inside a year. The team that wrote them feels good. The team that has to live by them does not notice the difference.
The values that actually shape a company look different.
What stops values from sticking
Three patterns kill values projects. They are predictable enough to design around.
Designed by committee. Every word gets sanded down to the point where nobody could disagree with it. "Excellence." "Integrity." "Innovation." Words so generic they describe every company that has ever existed. Nobody disagrees. Nobody changes anything either.
Costless. A value that does not force a trade-off shapes no decision. "Customer obsession" is costless until it slips a launch date because the customer signal is not yet there. "Long-term thinking" is costless until it costs you a quarter. Pick the cost when you pick the value, or you have not picked anything.
Unwired. The values live on a page and nowhere else. Not in hiring loops, not in promotion criteria, not in the weekly rituals where decisions actually get made. Wiring is the difference between a value and a slogan.
The design pattern that works
Three to five values. Each one carries:
- A statement. One sentence, written in the company's own language. Not borrowed.
- A behaviour. What the value looks like on a normal Tuesday. Specific.
- An anti-pattern. What the value rules out. Equally specific. Without the anti-pattern, the value is too elastic to bite.
- A cost. The trade-off the value forces when it competes with something else. Named, in advance, by the exec team.
That is the unit. Five of those, written by the exec team in the room, argued through, signed off by name. Not by a committee. Not by an external facilitator who hands you a deck.
Wiring the values into the operating system
Once the values exist, the work is operationalising them. Four insertion points:
Hiring loops. One interview in every loop explicitly probes a value with a scenario from real work. Not "tell me about a time you showed integrity." A specific situation where the value would have forced a trade-off. The signal is whether the candidate has ever had to make that trade-off, and what they chose.
Promotion criteria. Each level expectation is rewritten through the values lens. Senior IC at this company means specific things this company values, not generic things every senior IC role values. Promotions reinforce the values or they undermine them. There is no neutral.
Weekly rituals. Demos, retros, decision logs all reference the values explicitly. "We made this trade-off because of [value]." The reference is the practice. Without it, the values fade.
Performance feedback. Specific praise and specific challenge, both grounded in a named value. The same evidence-led shape as the coaching system. "This decision lived our value of X. This other decision did not, and here is why."
If a value is not present in at least three of these four insertion points within ninety days of being adopted, it will not survive the year.
Where AI helps, and where it should not
AI is useful at the edges of values work, not in the middle.
It can pressure-test a draft statement: rephrase it as a forced choice between two attractive options, see if the value still picks one. It can surface patterns in employee experience surveys, exit interviews, and engagement data that map to specific values, evidence the exec team can react to.
It cannot generate values. The exec team has to argue them out in the room, in their own language, with the cost of each one named openly. A value generated by a model is borrowed in the most literal sense, and people can tell.
The test
Six months after launch, ask three questions:
- Can a randomly-selected employee name the values without checking?
- Can they describe a decision in the last quarter that one of the values shaped?
- Can the exec team name a decision the values made harder, and explain why they kept the value anyway?
Three yeses and the values are alive. Two yeses and they are decaying. One or zero, and you have a poster.
What this connects to
Auto-recommended next reads in the People Ops cluster, ranked by shared concepts and headings:
- Leading the AI transformation in People
- Coaching and feedback systems that actually compound
- An AI policy blueprint for People teams
- Automation patterns that pay off
Common questions
- Why do most corporate values projects fail?
- Three reasons. They are designed by committee, so the language gets sanded into harmlessness. They cost nothing to live by, so they shape no decision. And they live in posters and onboarding decks, not in promotion criteria, hiring loops, and weekly rituals. A value that costs nothing changes nothing.
- What does it mean for a value to be costly?
- A value is costly when it forces a real trade-off. "Customer obsession" is costly when it leads you to slip a launch date because the customer signal is not yet there. If your values never cause an uncomfortable decision, they are not values, they are decoration.
- How many values should a company have?
- Three to five. Fewer than three and you cover too little. More than five and nobody can recall them under pressure, which is the only moment they matter. Five is the practical ceiling.
- How do you wire values into operating practice?
- Four insertion points. Hiring loops (one interview explicitly probes a value with a real scenario). Promotion criteria (each level expectation is written through the value lens). Weekly rituals (demos, retros, decision logs reference the values). Performance feedback (specific praise and specific challenge, both grounded in a named value). Without these, values stay theoretical.
- Where can AI help, and where should it not?
- AI can help test whether candidate value statements are specific enough (rephrase as a forced choice, see if it still bites). It can help surface patterns in employee experience that map to specific values. It cannot generate values for you. The exec team has to argue them out, in the room, in their own language.
If this resonated, there's more.
Subscribe to receive new Intelligence pieces as they're published. No noise — just the work.
By subscribing you agree to our Privacy Policy. Unsubscribe any time.
