Ben Fisher✨

Founder @ Rodeo and Fractional CTO to 8-figure DTC brands, sharing learnings as a Full-stack Coder, Designer and Shopify Expert

My friend Paul says I'm a "Human Zapier." That checks out. 

I'm a solid programmer and a decent designer, but they're really a means to an end: I love to build things and solve business problems. 

I'm a product person, a technologist, and an entrepreneur at heart. 

My superpower is at the intersection of tech + design + automation/optimization. And I love dtc ecomm.

I get a DISGUSTING amount of satisfaction by eliminating inefficiency. And I do it through building software tools.

I am obsessed with Systems Thinking and love to experiment with new ways to connect disparate systems.

I won't stop until we've got a deceptively simple solution that's damn fun to use.

**

For the past 10 years, I've built "commerce enablement" tools in the Shopify ecosystem that help DTC brands make more $$$.

With CartHook, we brought post-purchase upsells to the Shopify ecosystem and worked directly w/ Shopify to add it natively. 

CartHook processed > $1B GMV and added $250MM/year in revenue through its post-purchase upsell tech. It was acquired by Pantastic in 2022.

At Rodeo, we released one of the first native Shopify subscription apps (acquired by Recharge) and pivoted to a LTV Growth platform that originated as a small feature. Ultimately, wound it down as Shopify actively started building a solution themselves.

I'm on a mission to build damn good tools that help you make more money and have fun doing it.

Co-founder, Technical CEO

CEOs trying to get their teams using AI — this might be the problem.

Most AI conversations go sideways in the first message.

Not always because people are "bad at prompting" – but because they're thinking out loud. Half-formed ideas, missing context, vague asks. The AI takes that initial direction and runs with it. Now you're course-correcting for the next 10 messages.

For people who steer AI all day, that's fine. You recover and keep going.

But when someone on your team tries AI once, gets garbage, and decides "it doesn't work for me" — that's expensive. And often, it's just a bad first message.

I built Promptó to fix the 𝗳𝗶𝗿𝘀𝘁-𝗺𝗲𝘀𝘀𝗮𝗴𝗲 𝗽𝗿𝗼𝗯𝗹𝗲𝗺.

It's a desktop app – one shortcut from anywhere. Copy your rough thought, press the key, and it turns your half-baked idea into a clear starting point. Then you paste it into ChatGPT, Claude, whatever you use.

What it actually does:

- Asks clarifying questions when you're too vague
- Adds the structure AI needs to run useful
- Saves your context window for actual work, not course-correction

The reason it's a desktop app (not a browser tab): 𝗳𝗿𝗶𝗰𝘁𝗶𝗼𝗻 𝗸𝗶𝗹𝗹𝘀 𝗮𝗱𝗼𝗽𝘁𝗶𝗼𝗻. A shortcut becomes muscle memory. People use it for everything, not just "important" prompts.

The compound effect of starting clean – hundreds of times a week – is real.

One early user told me it writes better prompts for his vertical than his best employees.

If your team has quietly decided AI "isn't for them," it might just be a first-message problem.

𝗜𝘁'𝘀 𝗳𝗿𝗲𝗲 𝗳𝗼𝗿 𝗻𝗼𝘄. Link in the first comment.

Struggling with AI adoption on your team? What's worked (or hasn't)?

--

Link: https://tryprompto.com

Its companion desktop app: https://lnkd.in/ePSw5UHS

Ben Fisher 🏴‍☠️

Playing w/ AI + Automation | Prev @Rodeo, @CartHook (trusted by AG1, Four Sigmatic)

graphical user interface, text, application

Founders / PMs: stop me when this sounds familiar.

You have a great product call. Ideas, constraints, edge cases all get talked through. Everyone hangs up and... nothing really happens.

A week later you're trying to recreate what you already figured out on that call.

Last week I had a 45-minute product brainstorm with my friend Jared Cocken. By the time I closed Claude, I had:

- A first-draft spec the team could actually build from
- User stories mapped to the core problem
- A short list of questions we hadn't answered yet

Total time after the call: ~15 minutes.

In my setup, AI does the boring part:

- Pulling the relevant parts of the conversation out of the call notes
- Turning messy ideas into something an engineer can react to
- Pointing out gaps and edge cases I didn't cover

I still keep the actual thinking:

- Is this even a problem worth solving right now?
- Which requirements are real vs "nice to have"?
- What actually makes it into the doc we share?

Most teams just let transcripts pile up in whatever tool they're using and tell themselves they'll "get to it later."

In practice, later rarely happens.

This little workflow has made it way more likely that good calls turn into actual work instead of just... more recordings.

If you want to see exactly how I run this, I recorded a short "Dear Ben" episode walking through the flow step by step, plus a small Claude skill + GitHub repo you can hand to your team.

(Link in first comment)

--

Show notes & grab the Claude skill: https://lnkd.in/emJ8AbjQ

𝗦𝘁𝗼𝗽 𝘀𝗲𝗻𝗱𝗶𝗻𝗴 (𝗷𝘂𝘀𝘁) 𝘀𝗰𝗵𝗲𝗱𝘂𝗹𝗶𝗻𝗴 𝗹𝗶𝗻𝗸𝘀.

They feel efficient. They're actually kind of rude.

"Here's my calendar. You do the work."

Open link → load page → scroll days → do timezone math → fill the form. All the friction is on the other person.

The most respectful thing you can do instead:

- Propose specific times
- In 𝘁𝗵𝗲𝗶𝗿 timezone
- With a backup link if nothing works

Something like:

> Would any of these times work for a 25 min meeting (ET)?
> • Wed, Jan 7: 10:30am, 10:30am, 1:00pm
> • Thu, Jan 8: 1:00pm, 3:00pm, 3:30pm
> • Fri, Jan 9: 7:00am, 9:00am, 10:30am

> If none of these work, here's my full calendar: [link]

Doing this manually every time sucks:
check calendar → convert timezones → format nicely.

Even premium tools miss this. 

Superhuman has a built-in scheduler, but you still have to *manually* select your available slots. Every. Single. Time. 

I thought AI agents might solve this, but they got timezone conversions wrong or inconsistently replied.

So I did what I usually do with this kind of annoyance:
I built a small tool for myself.

- Raycast extension that pulls from SavvyCal or Cal dot com
- You pick the recipient's timezone
- It batches times around your existing meetings—not random scattered slots & respects your scheduling preferences (buffer times, working hours, etc.)
- Each time is a clickable link to book  

Both are open source.

Same pattern as Promptó for me: build something I genuinely rely on in my own workflow, then see if it quietly becomes indispensable for other people too.

This one just attacks a very specific kind of friction: making other people do the work to meet with you.

👉 (See link in first comment)

Curious: how are you making scheduling less painful for the people you work with?

--

- "Propose Times" — Raycast extension (https://lnkd.in/eVSrzFrp)

- "SavvyCal Booker" — one-click booking form (for Cal.com or SavvyCal) https://lnkd.in/e5P-AVqb

Everyone's hyped about how fast AI can ship features.

Almost no one is talking about the fights it's causing inside engineering teams.

I had two conversations last week with CEOs trying to get their teams to adopt AI. Same pattern both times.

The CEO discovered vibe coding. Built a feature themselves — maybe a proof of concept, maybe a working prototype. Felt the speed. Got excited.

Came back to the team with, "We should all be doing this."

And hit a wall.

Here's what I've realized: the engineers pushing back aren't all pushing back for the same reasons.

I'm seeing two camps:

𝗠𝗮𝗸𝗲𝗿𝘀 — product-oriented. They care about what gets shipped. AI is leverage. Less boilerplate, more interesting problems. They're often already experimenting. (That's me.)

𝗔𝗿𝘁𝗶𝘀𝗮𝗻𝘀 — craft-oriented. They love writing code. The problem-solving, the shaping, the control. They'll happily spend an afternoon refactoring a module no one else will ever see, because the code "wasn't right." AI doesn't just change how they work — it threatens something they've built their whole career around.

The artisan isn't wrong. They're assessing AI through a different lens. And as a leader, you need to recognize: adopting AI might mean giving up parts of the job they genuinely love.

What I've seen work:

𝗙𝗼𝗿 𝗮𝗿𝘁𝗶𝘀𝗮𝗻𝘀 — don't lead with "look how fast it codes." They've seen the output. They're not impressed. Start with workflows that protect the craft: scaffolding, guardrails, and reviews that produce code they'd actually be proud to ship.

𝗙𝗼𝗿 𝗺𝗮𝗸𝗲𝗿𝘀 — they might need the opposite. Coaching on product thinking, not just AI enthusiasm. Sometimes they're over-indexed on the tooling.

The starting point is the same: stop selling AI as a productivity tool. Start by understanding what each person stands to lose — and what they might actually gain.

Which camp are you in? And if you're leading both, what's actually working (or not) on your team?

𝗦𝘁𝗼𝗽 𝘀𝗲𝗻𝗱𝗶𝗻𝗴 𝘀𝗰𝗵𝗲𝗱𝘂𝗹𝗶𝗻𝗴 𝗹𝗶𝗻𝗸𝘀.

They feel efficient. They're actually kind of rude.

"Here's my calendar. You do the work."

Open link → load page → scroll days → do timezone math → fill the form. All the friction is on the other person.

The most respectful thing you can do instead:

- Propose specific times
- In 𝘁𝗵𝗲𝗶𝗿 timezone
- With a backup link if nothing works

Something like:

> Here are a few options that should work on your end:
> • Tues, Jan 14 – 10:00–10:30am your time
> • Wed, Jan 15 – 1:30–2:00pm your time
> • Thurs, Jan 16 – 4:00–4:30pm your time
> If none of these work, here's my full calendar: [link]

Doing this manually every time sucks:
check calendar → convert timezones → format nicely.

𝗘𝘃𝗲𝗻 𝗽𝗿𝗲𝗺𝗶𝘂𝗺 𝘁𝗼𝗼𝗹𝘀 𝗺𝗶𝘀𝘀 𝘁𝗵𝗶𝘀. Superhuman Mail has a built-in scheduler, but you still have to manually pick your available slots every single time. No memory, no defaults. It's just a slightly faster version of the same repetitive work.

I thought AI agents might solve this. Tested three: Howie, Fyxer, and a custom Lindy agent. None of them nailed it. They either got timezone conversions wrong, inconsistently replied, or made the back-and-forth longer instead of shorter.

𝗔𝗻𝗱 𝗲𝘃𝗲𝗻 𝘄𝗵𝗲𝗻 𝘁𝗵𝗲𝘆 𝘄𝗼𝗿𝗸, 𝘁𝗵𝗲𝘆 𝗼𝗻𝗹𝘆 𝗺𝗮𝗸𝗲 𝗶𝘁 𝗲𝗮𝘀𝗶𝗲𝗿 𝗳𝗼𝗿 *𝗺𝗲*. The other person still gets a link. Still has to check their calendar, do timezone math, fill out a form. The friction just moved—it didn't disappear.

So I did what I usually do with this kind of annoyance:
I built a small tool for myself, then opened it up.

So I did what I usually do with this kind of annoyance:
I built a small tool for myself, then opened it up.

- Raycast extension that pulls from SavvyCal or Cal dot com
- You pick the recipient's timezone
- It spits out a preformatted message with a few options
- Each time is a clickable link—there's an optional Vercel app that pre-fills the time slot (neither SavvyCal nor Cal dot com support that natively)

Both are open source:

- Propose Times (Raycast extension)
- SavvyCal-Booker (optional one-click booking form)

Same pattern as Promptó for me: build something I genuinely rely on in my own workflow, then see if it quietly becomes indispensable for other people too.

This one just attacks a very specific kind of friction: making other people do the work to meet with you.

👉 (See link in first comment)

Curious: how are you making scheduling less painful for the people you work with?

Managing Partner at StackedSP

As the Managing Partner at StackedSP, my commitment is to empower tech startups to achieve their audacious goals by stacking their teams with champion talent. This journey requires a close partnership with early-stage founders and the orchestration of world-class recruitment processes. My principal focus is on building high-performing teams from the ground up, and as a strategic partner, I am dedicated to not just meeting, but going above and beyond to ensure the job gets done.

Jared Saks

Strong AI credentials does not equal production-ready for a startup.

After dozens of early-stage AI searches, this keeps proving true.

Many candidates with impressive AI backgrounds have never actually shipped AI into production.

That doesn't make them bad engineers. It just means early-stage startups often need something different.

What I’m actually seeing work at early-stage:

→ Engineers who've shipped v1s they weren't proud of (and iterated)
→ People who've debugged at twilight hours when the model drifted
→ Builders who know when not to use AI

The founders we work with don't ask for perfection. They ask for systems that run, learn, and improve over time.

taplio

The Battle

Ben Fisher✨

Jared Saks

Followers

Posts

Engagements

Likes

Performance ScoreBen Fisher✨'s LinkedIn Performance

Performance ScoreJared Saks's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Ben Fisher✨

Jared Saks

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Ben Fisher✨'s LinkedIn Performance

Performance ScoreJared Saks's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreBen Fisher✨'s LinkedIn Performance