The Leaflet #128: AI for culture audits, goals instead of roles, a fun emergency plan
AI for culture audits
We’ve recommended a culture audit here in a piece written many LLM models ago. Our core suggestion there was to survey your people and ask them what the company’s values actually are in practice. That kind of audit is still a useful thing to do.
This is also a strong use case for AI. Models haven’t been trained (yet) to deliver rich, idiosyncratic, creative writing. They have been trained to identify patterns, the bigger the data set the better.
Grabbing a chunk of meeting transcripts, Slack threads, and email exchanges and feeding those to Claude with a prompt like the one below can be a useful part of your culture audit. In a more advanced (and more AI-exposed) approach, you have Claude automatically ingest and analyze this data and give you a report every few months.
Culture audit prompt:
“Our team officially has the core values of [value], [value], and [value]. I’m conducting a culture audit of the organization to determine a few things: 1) how well are we embodying our stated values? 2) what are our “shadow values”, i.e., the values we actually practice and demonstrate a revealed preference for? 3) what if any changes should we consider making to our stated values, based on 1) and 2? Can you review the attached materials and give a robust answer to these questions, citing specific pieces of evidence from the materials?”
-eric
Goals instead of roles in the early going and maybe the later going, too
When your venture or project is just getting started, you might lack the skeleton and exoskeleton a mature organization has. You and your cofounders are holding the equivalent of three or four or nine jobs. There’s no HR department. There’s not even a single HR person.
There’s something useful about this period that you can carry forward into the more mature and exo-skeletoned phase of your team’s life. In the early days, you’re pretty much forced to assign goals instead of roles. Things need to happen. Things need to exist by x date that don’t exist now (products, relationships, org charts, policies, a logo, who knows the list is quite long).
It doesn’t make sense in this protozoan chapter to tell one of the four people on the team “You are the Senior Director of Digital Marketing”. The title says too much and too little at the same time. It does make sense to say, “we only survive if we sell lots of this stuff on the internet. Let’s say ‘lots’ means ‘$1M worth by the end of the year’. Can you own that for us?”
I’d suggest this can be a useful approach after the founding frenzy, too. When you realize you need more people on the team, start with the short list of things the new person will make true by the end of the year. Then back-map a title onto that set of goals.
Starting with the title can confuse you and the candidates you recruit, because titles are little dynamite sticks of cognitive bias. They are three-word behavioral economics poems. It takes a lot of effort to explain and understand what they mean here, now.
-eric
A fun emergency plan
Oh my word the Dowager Countess has croaked and wouldn’t you know, she secretly bequeathed your organization a handsome sum, I mean a 10x-your-annual-budget kind of sum, one fine day while the sun smooched the crocuses and the ill-mannered Pomeranian sat upon her noble lap!
What are you going to do with all dat money???
Run the exercise. Get yourself a robust answer to the question. Adjust the details however you like (I know some people are partial to long-haired Dachshunds, eg, other people prefer their windfalls from Lithuanian gardeners made good, eg).
Two benefits I hope come out of this for you:
Object level: you have a plan for what to do in the case of a happy emergency. Ideally this promises a higher return on luck.
Meta level: the exercise reveals something that might have been tacit or hidden about your ambitions. You have a clearer sense of what really maxing this thing out would look like. Maybe you also have a sense that you could get there without a wild bequest. The crazy success may be more tractable than you realized.
Also interesting in this thought experiment is what you wouldn’t change at all when you get the fat windfall.
-eric
COMPELLING QUOTATIONS
Critic Henry Oliver on reading Henry James and walking the dog:
Although some of my efforts to read James had been successful … this was an entirely new sensation, and I no longer read in that plodding, linear manner I had become accustomed to with James’s other books. Now I read like an enthusiastic spaniel, who, when let off the lead on a walk keeps running ahead and coming back, making loops, working over a piece of ground, moving on, coming back, making comparisons of obscurely doggy matters, so that in the end, the dog has not only had a walk of three or four times the length of the amused humans, but has experienced everything in a wild, rotating reality, quite different from the pleasant linear walk that takes place beside him.
AI observer Ethan Mollick on the questions he’s now trying to answer:
Two years ago, the question I was trying to answer was how to think alongside a new kind of intelligence. Now the questions are weirder: When should you refuse AI’s help, even when it is offering? When should you hand over the keys entirely? And what do you do when the AI is no longer just your assistant, but your reader, your critic, and the gatekeeper standing between your work and its audience?
Poet David Whyte on witness:
The ultimate touchstone is witness: the privilege of having been seen by someone and the equal privilege of being granted the sight of the essence of another, to have walked with them and to have believed in them, and sometimes just to have accompanied them for however brief a span, on a journey impossible to accomplish alone.