Pinned Tweet
I’ve worked with 50+ high-growth brands doing up to $40M a year.
These 7 steps had the biggest impact.
1. Define Your Problems
Finding operational issues in your company is tricky.
It’s very difficult to find the source of the problem. Sometimes a symptom can look a lot like a source.
The best way to find what's causing friction is to map out each business process, step by step, and make it visual.
Any tool will work, even a piece of paper, but you need to get a visual on every step of every process in your company.
Here’s one we did last week
At OpsKings, we’ve built a proven method to guide teams through this.
2. Fix Operations Before Touching Systems
Once you have the business process map, the next step is to analyze it.
We ask questions like:
- Are we doing this step because we always do it? Or does it actually have value?
- Does this step NEED to involve the founder (or manager’s) time?
- Is this something we can codify in the future? Or will this always need to be a “person” step?
By doing this, you often remove entire sections of the business. (we once removed an entire department for a 20 person company doing this)
You’ll also spot things to tweak, steps to reassign etc.
3. Build a SSOT (Single Source of Truth)
Once you have ironed out your processes, the next step is building out a SSOT.
Here’s a specific example of how we do this:
teapawt.net/Steven_DaSilva_/status…
In general we convert your business into data based objects.
Every piece of information in your business becomes tied to a noun.
Here’s WHY you want a Single Source of Truth
- Reduces repeated questions (people stop asking where things are)
- Centralizes information so everyone refers to the same data
- Makes handoffs between team members cleaner
- Helps avoid version control issues (no more "which doc is correct?")
- Simplifies onboarding by letting new hires learn from the system itself
- Cuts down miscommunications, the truth is the truth
- Enables faster decision-making with accurate, live data
- Allows you to track progress without relying on DM’s
- Makes reporting easier since everything lives in one place
- Encourages accountability, everyone can see who owns what
4. Set Up Visibility
Visibility is the MOST under-rated, misunderstood lever I pull on.
You can capture this concept by asking, “Would you like to take a Taxi or an Uber?”
Even though it’s not intuitive, seeing the right information at the right time has a lot of value.
Once you have a SSOT, you can start showing each person exactly what they need to see.
For example:
Client sees:
- What actions your team is currently taking
- What actions they need to take (and if they’re holding things up)
- The actual results of the work
Team sees
- Everything they need to do their job well (actions, files, context)
- Their own performance metrics (so they don’t have to guess how they’re doing)
Manager sees
- What each team member is working on
- Where things are blocked
- Key performance indicators (KPIs) for every person on their team
Admin sees
- A high-level overview of the entire operation
- Whether marketing should be scaled or paused
- An overview of their unit economics
- Whether the team is at capacity or still has room
- Where their own time is best spent based on what’s happening now
Side note - For the team, you can bake incentives into the system itself. If performance is visible, you can tie compensation to it.
You will notice less friction all around. Less meetings and internal comms.
Here’s more on visibility:
teapawt.net/Steven_DaSilva_/status…
5. Stockpile Your Data
Now that you have a smooth system, you can mine your data.
As time goes on you get more accurate and detailed insights like:
- Where the bottlenecks are
- Who the best employees are
- Which client is likely to churn soon
- Which client ripe for an upsell
- Which employee is over-worked
- Which offers have the best margins
- How fast you can onboard/offboard a client
- Etc.
You are able to make faster decisions with less mental energy.
6. Leverage AI
Now that you have organized data, you can start leaning on AI in many areas.
It becomes very easy to feed context and get good responses because the data is all there.
There are 2 ways I like to use AI
The first is for talking to your data.
You can ask questions and get contextual useful answers.
- “What should I spend by time on improving”
- “Do we have the bandwidth to take on X client”
- “Which client is likely to Churn”
- “Do you see any major risks or opportunities in the business”
The second is drafting and checking.
At this point you can start handing off the first draft of common deliverables to AI.
It won’t be perfect, but it saves a LOT of mental energy for you and your team.
AI does well with client reports, copywriting, technical instructions, and strategy.
The key is always the same: give it contextual data and a prompt, then review and refine.
The checking part is even more valuable.
AI can review deliverables against specific rules and criteria.
I like asking founders how they review things.
What do they look for first? What mistakes do they always spot?
We then bake that logic into prompts, so the AI is thinking like the founder during review.
Done well, this system speeds up output and raises the baseline quality.
It also creates leverage around the founders brain.
7. Build Sales Transparency
A 'sales transparency' view shows prospects and clients exactly what your systems are delivering
This kind of view acts as proof of:
- What makes your service more effective than other options
- Whether the work you're doing is actually delivering results
This builds confidence and will improve your close rate AND retention rate.
Uber and Dominos are successful for the same reason.
They do one thing well.
They give you window into what is happening as they work on your deliverable. 🍕🚗
Visibility is the most important thing you can have when providing a service.
The dominos tracker is a great example of this, you can see (with reasonable detail) what stage your pizza is in.
Visibility into what's happening provides a sense of progress. It stops people from cancelling because they understand it's in the works.
This makes people want to order again (less buyers remorse).
This builds trust.
With all else being equal, the more visibility you provide, the more valuable it becomes.
The business with the most visibility wins.
Every business alive today should have a AI based operating system.
AI is the most useful when it has the BEST context.
You need a system where you can control the information it has access to, and in what way.
AI based operating systems allow you to house persistent context about your clients.
AI can (on demand) grab access to
- core deliverables
- subjective happiness scores
- call transcripts
- slack messages
- emails
- etc
Which allows you to make better decisions much faster.
An Operating system also allows AI to do things for you. Like:
- requesting work from your team
- completing intermediate steps in your process (drafting work)
- capturing missed work
- finding mistakes
- triaging issues
- creating progress reports
etc.
It can only do these things because it has the context and ability via the operating system data.
2
146
SOP's are finally dead. (I never liked them anyway)
For me, they have been happily replaced via AI guided workflows.
The most notable being Claude skills which (unlike old school SOP's) guide you through the work.
They allow for consistent outputs based on whatever principles you give it.
Which is why you should get the most over qualified person at your company to make these.
Then EVERYONE gets outputs like that person.
AND they are quick to update.
You don't like an output? Use claude to help you figure out the missing principle and update it (in 5 seconds).
It's invaluable, has probably saved me 50+hrs in the past month.
1
3
235
AI is eating UI.
Any CRM that does not include the ability to update data conversationally is going to be obsolete within a year.
1
101
Visibility is king.
You have to have great deliverables AND your clients need to be aware of the effect they have.
Between low quality deliverables and bad/absent reporting, most eCom brands have no idea what their SEO agency is doing, or if it's even working.
But not brands that work with us.
We just finished month 1 with this client. To date, here's what they've gotten:
- full tech audit & implementation
- 15 new collection pages + optimized home page
- 10 backlinks (7 link insertions, 4 guest posts + 10x Tier 2 backlinks)
- 4 weekly updates
- 1 monthly KPI report, delivered via video
All deliverables ahead of schedule + frequent reporting = great results and increased trust.
Looking forward to more in month 2!
1
4
295
By far my biggest pet peeve is people who are immune to capitalism.
People who do the bare minimum are the worst.
4
2
274
What if you baked your team's motivation into your processes?
Imagine a system where team performance and compensation are clearly visible.
This is how you create leverage and scale without burning out.
Most founders sink so much time into managing their team. Which can often leave no time for strategic growth.
You're debating hiring new staff, but you're not sure if it will fix the underlying systemic issues.
The real issue isn't people, it's the lack of an effective system that creates motivation.
The Solution: Use Visibility as a Lever.
Start with a Single Source of Truth (SSOT) and then set up dashboards.
Show each team member their performance metrics in real time.
When performance is visible and objective, you can tie compensation directly to it.
Tying compensation to performance aligns your team's "wants" with yours.
You don't have to pay out more money to the team either, but it's the terms of the payment that is key.
The more you incentivize, the better results you'll get.
I just got my hands on the new OpenAI Agent Mode, and it’s pretty wild.
I tasked it with doing QA for a custom app we are working on.
It took my use cases, clicked through the entire system like a human, and was able to find small (and large) issues.
It then summarized and documented all the issues for me.
Automating QA was the last thing I thought AI would be able to do, I am quite impressed.
Couple other area's I'd imagine this would work well for
- Website Design QA: Catching broken buttons and design flaws.
- Form Checking for SEO: Confirming lead forms are still functioning after updates.
- Sales Touchpoint Audits: The agent can act as a lead and click through your entire funnel to find where a system might be broken/confusing.
2
4
252
There are only two types of permissionless leverage:
Code and Media
That’s it.
When customer acquisition isn't an issue, Code is the most reliable lever you can pull.
It’s pure leverage.
- Scales infinitely
- Works while you sleep
- Nearly free (if you avoid bloated no-code tools)
- Zero marginal cost
Code turns your thinking into systems.
And systems don’t get tired, complain, or ask for raises.
If you solve a problem with code, you never have to solve it again.
2
6
285
Huawei's gesture-sensing tech enables touchless car door opening
teapawt.net/i/status/1904588674198…
146
*Logs into LinkedIn*
"Unlock my 50K Notion N8N dashboard that’ll scale your
Dad to MILLION-DOLLAR productivity! 🚀 DM me for the template!"
*Logs out of LinkedIn*
2
6
283
If you'd like to run your business through this process, book in here!
calendly.com/steven-opskings…
1
135
Live look at AI trying to read your data right now
AI is only as useful as your data is organized.
Imagine asking a friend for help with a big decision.
They’ll ask you a bunch of questions:
What are you trying to do?
What are the pros and cons?
What actions have you already taken?
Your friend only knows what you tell them.
If you leave out key details, explain things poorly, or mix up facts, they can’t give you good advice.
AI is no different.
I see a lot of businesses excited to “plug in AI” and hope it will fix their problems.
But they miss the most important part: if your inputs are messy, your outputs will be worse.
This is where having a Single Source of Truth (SSOT) becomes a major lever.
A SSOT means:
- One reliable home for key data
- Alignment on what’s accurate
- No debating between 12 conflicting versions of reality
Without it, you’re setting the AI up to fail.
It’s like asking someone to cook dinner using a fridge full of unlabeled containers.
If you want AI to work for you, you have to give it something solid to work with.
1
198
















