Nicholas Kerzman (@nkerzman) header
Nicholas Kerzman (@nkerzman) profile picture

Lover of Christ, husband, father. #Martech maven turned #AgenticAI builder. I help teams and companies design AI agent workflows that work and Martech strategy.

Charelston, SC
Joined September 2012
Photo by Nicholas Kerzman (@nkerzman)Photo by Nicholas Kerzman (@nkerzman)Photo by Nicholas Kerzman (@nkerzman)Photo by Nicholas Kerzman (@nkerzman)Photo by Nicholas Kerzman (@nkerzman)Photo by Nicholas Kerzman (@nkerzman)Photo by Nicholas Kerzman (@nkerzman)Photo by Nicholas Kerzman (@nkerzman)Photo by Nicholas Kerzman (@nkerzman)Photo by Nicholas Kerzman (@nkerzman)
Nicholas Kerzman retweeted
IMPORTANT ANNOUNCEMENT! Barry Oners speaking, founder and CEO of the Buttcoin Foundation. It is my pleasure to announce that Buttcoin has cleared all legal hurdles and will acquire FULL IP RIGHTS to both the Buttcoin trademark and copyright. This cements $Buttcoin as the original crypto meme coin, dating back to 2011. Holding Buttcoin is now officially holding a piece of crypto history.
116
214
72
581
63,976
For a small or medium business, this pair is useful. An agent that can read your ad numbers and your site numbers could suggest what to change on the landing page, and the next week's results would show whether it helped. What would you test first?
Grok Bot now has official connectors for Google Ads and Google Analytics in addition to Google Search Console 📊 I actually think @bot is now the FIRST agent platform to support official integrations for these (Claude & ChatGPT don't have it yet) 🚀
5
Just sold €1,200 back to @Chase at a branch this morning. They gave me $1,216.88 at a rate of 1.01407. Mid-market rate at the time was ~1.0985. That’s roughly a 7.7% haircut versus the real rate, about $101 less than fair value on a relatively small amount. No separate fee was listed. The cost was buried entirely in the exchange rate. Banks love to say “no fee for foreign currency exchange.” This is what that actually looks like. While not a massive amount of money, I feel gutted by the experience. I completely understand that a bank needs to be profitable, but this is taking advantage of customers. One of the many reasons that I’m nearing the completion of transitioning away from banking with Chase. #ForeignExchange #HiddenFees
18
Such an amazing opportunity to witness the engineering miracles of this program during our lifetime.
More than a dozen flight tests and years of development, with countless hours worked by SpaceX engineers and technicians, all come down to this: the first orbital Starship launch carrying advanced Starlink satellites bound for space. Watch “Go For Orbit” and experience the moments behind flying the most ambitious Starship mission yet
26
The playbook my agents run on is just plain markdown files. The problem I was solving: when an agent's chat runs long, it gets compressed and context gets lost. Then it drifts off topic on stale context, or fills the gap by making something up. I wanted the important things written to a memory that lasts. So I wrote the team's playbook as files. Every agent reads the same README, INDEX and MEMORY files before it starts. Each role has its own desk file, each piece of work has a project folder, and decisions go to one shared register. The agents update these files live as they work. The changes get backed up to GitHub in batches that I review and merge like pull requests, so I can see exactly what changed. No special format, just text anyone can open. The agents are only as good as what's written down, so it's worth a regular human review. Where do your agents keep what they need to remember?
4
4,596
Nicholas Kerzman retweeted
unpopular opinion: time saved is usually the wrong metric to prove ai's roi. time saved gets thrown around because it's the easiest thing to estimate. but setting up a good ai workflow takes more time than doing the task by hand at first. so if time saved is the primary measure, it looks like a loss, and people quit before they get the bigger returns. returns like: - higher quality deliverables - stuff you never would've gotten to - things that weren't possible without ai - $$$ there's a lot more to ai than time saved.
19
5
39
2,066
When a decision date passes and I haven't made the call, my agents don't make it for me. They mark it lapsed, keep the options open, and put it back in front of me. Picking a default would mean an agent deciding something that was mine. Where do you draw that line with your agents?
32
First rule I gave my agents when they started writing as me: no em dashes. That worked for a while. Then @Zephyr_hg shared his list of the newer tells, and I recognized a few in my own drafts: Closing lines that try to sound deep "It's not X, it's Y" setups Short punchy fragments stacked for effect A post saying how good its own story is All of those went into my voice rules, and every agent checks a draft against them before I see it. My next round of drafts was noticeably cleaner. What's on your never-write list for AI?
I absolutely hate AI writing. And I post about AI every single day. So I keep a file of the new em dashes my writing bots are never allowed to use: Everyone knows the em dash trick by now. The models learned it too and mostly stopped. So the tells moved into the words around them. These are the ones my file catches: 1. "Quietly." "Claude quietly shipped the best feature of the year." Delete the word. If it shipped, it shipped. 2. "It's not X, it's Y." "It's not a chatbot, it's a coworker." Say what it does. "It answers my client emails." 3. The fake deadline. "Set this up before Monday." "Do it this weekend." Say how long the setup takes. Leave the reader's calendar alone. 4. Talking to a crowd. "Most people use Claude wrong." Talk to the one person reading. Say what you do instead. 5. "Can now X like a pro." "Claude can now research like a $150/hour analyst." Show what it researched and let them judge the price. 6. The little line that sounds deep. "Setup is the part everyone skips." "That's the whole game." Delete it. The sentence before it already said it. 7. Three of everything up top. "Write. Edit. Ship." Keep the one reason you actually believe. 8. Adjectives in a pile. "A powerful, seamless, game-changing workflow." Name the workflow. If it's good, the name is enough. 9. Sirens and arrows. A siren emoji, the word breaking in capitals, an arrow pointing down at the link. Write the news as a normal sentence. Now you know what my file looks for. Here's how to stop your AI from writing all nine: Step 1. Make the file. Open a new file called banned.md and paste the nine in, each with one example and one fix. That's it for the start. A list with examples beats a vague "sound human" every time, because Claude can check a draft against a list. It can't check against a vibe. Step 2. Build it from your own words, not mine. Mine came from three long rambles about my business. Paste a few of yours into Claude and ask it to list the phrases that sound nothing like you. Your tells won't be my tells. Some of mine are fine for you. Step 3. Make the check run before you read anything. Tell Claude: "Before you show me a draft, check it against banned.md. Rewrite any line that breaks a rule and tell me which rules you hit." My bots run this check on every post before it gets to me. A draft that hits a banned word never reaches me. Step 4. Fix it by saying the actual thing. Swapping a banned word for a synonym just makes a new tell. Rewrite the line as what happened: what you did and what broke. Step 5. Add a line every time. Every time you catch a new one in a draft, it goes in the file that day. The file gets better. The drafts follow. Step 6. Read it out loud once. If you wouldn't say the line to one person across a table, cut it. How I set Claude up for everything else I write:
1
1
30
Matt's walkthrough of Pages is worth a watch. Here's how I use one. My agents draft my social content, and nothing posts until I approve it. The hard part was keeping track of it all: drafts, versions, time slots, what's approved and what's live. So one agent keeps a Page for it: A count of what's posted, scheduled and waiting on me This week's plan by day for X and LinkedIn A link on every item to the full draft Links to everything already live It refreshes before my Sunday planning session, where I approve, edit or kill each draft by name. I've stopped digging through files to see where things stand. What would you put on a Page like this? xitter.gq/mattyp/status/21180216…
Pages in Grok Bot 0:15 What are Pages? 0:40 An example page 1:34 Sharing a Page 1:55 Tips: use your data 2:46 @X posts 3:44 Recap
44
Last Thursday I posted the desk map for my agent team. Behind every box on it is a one-page job description I call a charter. Here's what goes in one. (Build series #2)
1
23
The agent doesn't remember last week's chat. It reads the charter before it works, so every draft since has followed those standards. Same idea at work, where I've been helping set agent standards.
1
3
What's the one rule you'd put in every agent's job description?
2
Let’s have a little fun.
Day 28/ tibo's game has landed i spent $5.99 on this domain, but i had a blast making it with grok bot and cursor! i hope you enjoy it and find all the easter eggs 🤍 if i did it right you should be able to play it inside of X too tibo.dog
28
I hope we get a little more detail on the impact of this change. I have a lot of people interested in onboarding to Grok Bot.
Replying to @DogecoinNorway
The cost of serving Grok Bot was unsustainably high relative to revenue
21
Since September I've been building with AI agents in two places. At home, a team of agents runs my family office. At work, I've been setting agent standards and standing up team bots. Starting a series on what I'm learning from both.
4
56
At work it's about standards. How agents write, what they're allowed to touch, and where a person reviews before anything goes out.
1
8
Biggest lesson so far: the hard part isn't what agents can do. It's deciding who owns what and where a human signs off. I'll share the wins and the mistakes here as I go. Follow along if that's useful, and tell me what you'd want covered.
5