The 43 tools that make the modern growth AI & tech stack at Ballpoint

And examples of the 50 workflows created from just two of these tools that power so much of our day-to-day

Sep 02, 2025

The tech stack in growth has gone through many phases since its incubation at Facebook in 2008.

First were the blackhat teenager years. You had to have an engineer on your growth team because there just weren’t the tools created to help run experiments.

Then came the late 2010s, as ad platforms and tech grew in importance. Zapier landed. Shopify took off. Look at the Uber growth teams in the early days and they were essentially ops people: deeply analytical, loving spreadsheets, and piecing automations together. Running Facebook Ads was a case of manually pairing audiences and ads together and optimising bids.

The early 2020s were the creative reset:

Data gaps caused by iOS14.5 and a clampdown in ad tracking.
Over-saturation of advertisers.
TikTok changed how we interact with content. Today’s platforms aren’t really social networks, they’re content platforms.

As a result “creative” became the way to win.

And now as we enter the second half of the 2020s, we enter another new shift. This one is being created by AI and new technology.

Pace is one of the most vital startup characteristics there is. Unless you are fast growth, you are not a startup, you’re just a small business. And so startups live and die based on their pace.

Because of AI we now have an unprecedented level of productivity and pace available to us.

As a result, the late 20s era for growth is starting to look a lot more like the 2000s and 10s. We don’t have a gold standard yet. The Shopify moment hasn’t come. And so those who experiment now will gain advantage.

I’m using to working in tech companies and before that highly experimental social media teams, and so a lot of this feels natural. But I also appreciate if you’re a brand founder, or maybe a growth marketer who has come up during the creative reset years, this will start to feel uncomfortable.

This overview of our tech stack and how we’re using it should help.

What’s on the agenda:

The 43 tools and software we use across functions like performance marketing, data analytics, client servicing, and creative
The 50 other micro-tools built using Zapier, n8n and similar tools that produce small but important workflows
How we’re balancing BAU and experimentation
What I’d do if I was looking for a job in growth today

How this list is structured

Here’s a preview of the list:

We’ve grouped the tools together by the top-level function and where there are specific use cases for repeated tools have listed multiple times.

High-level notes:

The lowest volume of tooling sits in creative strategy: there is still a lot of tooling at hand and the majority of the job often sits directly in the tools mentioned, but it’s not a broad range. We also as an agency believe this is an area where humans currently beat AI.
Highest AI adoption is within creative production: this continues to be an area where there is heavy investment in (A) external tool development and (B) internal experimentation
ChatGPT is as much a part of the modern work toolkit as Google Chrome or Slack is

View the whole stack

The 50 extra micro-tools driven from AI

Amongst those 43 tools, two are the ones that have the power to be representative of the other 41. And it’s not ChatGPT or Claude.

Zapier and n8n are the two which sit as the powerhouses of the business. (And with an honourable shoutout to Notion’s Automations which gets you quite far).

Zapier powers so many minor automations now, that you almost forget its there.

I’ve been a fan of Zapier for the last decade. I ran offsite workshops and lunch and learns called “Automate your life” while I was in my last in-house role 2017-2020. I’ve long been a fan.

But the inclusion of ChatGPT and Claude into Zapier was a game-changer. There’s now a contextual layer to these tools that wasn’t there before. Once you’ve hit Zapier’s limitations, n8n levels it up ever more. These tools really do change limits of what’s possible.

A few use cases we have in place:

Winning ads from all client databases get served into an ad winners channel, which then gets surfaced for our weekly meetings for discussion
Briefs get automated from creative strategy views → client-facing dashboards → creative production processes → back into client-facing dashboards
Trustpilot reviews get scraped → filtered using AI layers → fed-back into creative strategists
Ad ideas → fed into an AI for insights → fed back into a genAI tool for variations
Customer interview transcripts → fed into our trained GPT for JTBD analysis → back into Notion customer framework boards
Weekly facebook data → shared into AI queries → refactoring forecasts

Ultimately, each of these tools now within it powers many more micro-tools.

While there’s 43 tools in general, if you included every automation as a micro-tool, that number would be closer to 80 and that number is growing all the time.

Balancing BAU and experimentation

As operator-CEO of Ballpoint, decisions on level of AI usage ultimately sits with me.

The rate of software delivery at the moment means there is a permanent pace of adaptation required.

If you look back to January of this year – three-quarters of our current AI workflows weren’t possible

75% of our current day-to-day operations has been built and developed within the last six months. That’s incredible and means that our default position now has to be innovation.

This creates both deep excitement for where we can be in six more months – but also some level of anxiety around ‘keeping up.’ That anxiety is something many founders I speak to mention to me.

But all of it comes with risk. I could be spending 100% of my time experimenting with AI. I could also be directive to tell the team to do the same. But how do we end up balancing those things out?

The core team runs BAU

We’re a highly lean agency, as are most businesses I know these days, and so most of the team are at 100% capacity all of the time. That means, there isn’t by default idle time for experimentation.

We did try an ‘experimentation is everyone’s job’ approach earlier this year, and even incentivised the creation of Automations. But ultimately, there’s too much work to do for most operators, that taking a day out to rebuild a process is too much.

Experimentation sit with senior team

Ultimately, experimentation sits with the senior team. Productivity and efficiency are ultimately the KPIs for this team, and so it can only truly sit with them.

Play comes first.

For me personally, experimenting with new tools is pretty much what I spend my evenings or weekends doing for fun now. That’s not the expectation I have on other senior members, who will factor that time into the workweek, but you’ve got to make the time to play.

That could look like a quick 30 minute playaround with something new when it launches, or it could mean a few hours dedicated to testing a workflow.

Triage system for implementation

🟢 A new model or tool that just works launches. Like arcads and it’s an instant ‘let’s get this into people’s hands immediately’

🟠 Needs a bit of training on how to best use it. Senior team member experiments and runs workshop and helps team implement themselves.

🔴 Not yet ready for mass adoption: likely a great tool, but probably needs more development time.

Looking for a job in growth?

If I were on the job hunt for a role in growth today, I’d be investing all the spare time I had to get exceptional at AI.

At any level, whether the junior, mid, or senior, I’d be investing all weekends and evenings at the moment to testing out tools.

Go back-to-basics and ask yourself what you need to do at a task level to do your job. Check something in SEMRush? Why, break that out into micro chunks and think about automating it. Downloading data to analyse in Sheets/Excel? OK, build a process, then learn to use AI to automate it. Had to previously go to a data engineer for a task? Get cracking with Claude Code.

If any hire at any level came to me and said “I’ve been playing around with AI in my spare time for the last 3 months and have built these 5 cool things” they’d shoot to the top of any candidate list.

There is soon going to be a big divide of those who are active in the early-stages of AI and those who join the ranks at the tail-end of it. And I’d imagine, for many that means slowly being out of work or needing to reskill elsewhere.

View the whole stack

How are you doing this stuff?

This is new. I love learning as you can probably tell, so would love to hear what else I can be experimenting with and how else we can think about deployment.

How are you weighing up BAU vs testing?

What tools are in your day-to-day stack?

What automations in n8n and Zapier are saving you hours every month?

Thanks for reading.

Early Stage Growth • Ballpoint

Discussion about this post