GM @ https://xitter.cf/t.co/lEewtDBZei
France
Joined April 2009
- Tweets11.7K
- Following223
- Followers140
- Likes1.3K
Good documentation is like flossing. Nobody wants to do it, and the consequences only show up six months later when everything starts hurting.
What nobody tells you when you work in customer service: people have a thousand and one ways of asking for help (not just different channels).
A lot of companies have one person doing both CM and CS, and it's quietly breaking both functions.
Community management is about engagement, vibes, channel growth. Customer support is about resolution, accuracy, speed.
Different jobs, different metrics, different brain.
I'd rather have an AI tell me it can't help me and suggest I speak to a human than have it try to pretend to be helpful
Managing AI is like raising a child: you're proud when it goes to the bathroom on its own, but you're still the one who has to wipe
"If we write docs, we'll get fewer tickets."
Docs aren't a way to avoid support, they're the foundation every support layer depends on.
Without good docs:
- AI hallucinates
- agents improvise, so users get inconsistent answers
- new hires take 3x longer to ramp
Most SMBs founders i talk to are still personally answering support tickets at midnight. They think it's a badge of honor.
It's a bottleneck disguised as work ethic, because when the founder is the support, you cap your growth at the founder's bandwidth.
Most "AI support" pitches are wrong about the same thing: they treat the AI as the support. It's not, it's the first 60% of the support.
What actually decides whether AI helps or hurts you is the handoff. Its most important skill is knowing when not to answer.
First response time is the most overrated metric in customer support. Teams optimize it, brag about it, and still lose users.
The metrics i actually see correlate with retention across our customer base ⬇️
Repeat contact rate -> same user, same issue, twice = product or docs gap
Time to resolution, not time to first response -> nobody cares how fast you said "looking into it"