When I first started using ChatGPT, I did what I think a lot of business owners did.
I asked it questions, and I asked a lot of them. I used it for market research, product research, and exploring ideas that would normally have sent me down a Google rabbit hole for hours. I thought it was incredible.
I was also the person double-checking the sources. I have an MBA, and years of having “cite your sources” drilled into me were not going to disappear because a chatbot gave me an answer in seconds.
Still, I knew pretty quickly that something had changed. Information that used to take hours to collect and organize could suddenly be synthesized in minutes.
Naturally, my next thought was: What else can I make this do?
First, I got really good at prompts.
Once I understood how much context changed the quality of the output, my prompts became more sophisticated. I started feeding AI actual marketing strategy, sales psychology, customer insights, and frameworks I had spent years using professionally. The work got faster.
That was probably my first major shift with AI: I stopped seeing it as an interesting research tool and started using it to execute parts of my work.
Custom GPTs changed what I thought was possible.
When custom GPTs arrived, something clicked. Instead of rebuilding the context every time, I could take the strategies and frameworks I used repeatedly and create specialized GPTs around specific jobs. That was a game changer.
I started building them into my own workflows and eventually into client work. I also measured everything.
In my own business, I tracked roughly 400% revenue growth as I began integrating these AI-powered workflows into how I marketed and operated the business.
One client experienced approximately 1,900% revenue growth from her previous baseline during the period after we implemented AI into her marketing workflows.
Those results were not something I was willing to chalk up to “AI magic.” Real marketing strategies, offers, customer insights, and sales psychology sat behind the systems we built. The technology was not the strategy; it made a good strategy easier to execute.
That distinction became increasingly important to me because, at the time, I thought this was what advanced AI use looked like. It was not.
~400%
Revenue growth tracked in my business
~1,900%
Client revenue growth from previous baseline following implementation
The real shift happened when I stopped thinking in prompts.
Eventually, I realized I did not want to keep telling AI what to do every time I opened a chat. I wanted the context to already exist.
The strategy should already be there. The frameworks should not need to be copied and pasted again. If something changed in the business, I wanted the rest of the system to know.
That was when I started thinking less about individual prompts and more about connected workflows.
Read this diagram as text
Four steps arranged vertically. First, a prompt is typed fresh every time. Second, a custom GPT saves time on one task. Third, a workflow connects jobs in order. Fourth, an operating system holds the business information and standards that your human and AI team leverages to work autonomously. Paired bars show system context increasing as context supplied by the user decreases. A final comparison contrasts repeatedly briefing AI with asking the operating system to plan next week's social content.
Claude made me curious, but I still wanted more.
When Claude became a bigger part of the AI conversation, I experimented with it too. I will say something that might be unpopular: at first, I was not obsessed.
I tested Projects and played with different ways of organizing information. Later, Skills made much more sense to me because I liked the idea of creating reusable capabilities instead of rebuilding the same instructions in every conversation.
But I still was not interested in switching tools simply because the internet had a new favorite. I cared about what AI could actually do inside my business, and that question eventually led me somewhere much more interesting.
Then I moved outside the chat window.
I connected Claude to Visual Studio Code, and that blew my mind.
Instead of giving AI a little bit of context inside one conversation, I could put it inside an actual working environment with access to the information that makes my business my business.
That environment could hold knowledge files, business information, playbooks and frameworks, and founder standards. Eventually, I could connect more of the technology my business was already using.
This was not another chatbot I had to brief every morning. I was starting to build an operating environment powered by AI, and that changed how I thought about the opportunity entirely.
Faster tasks were never the biggest opportunity.
Most conversations about AI for business still start with productivity: write this faster, summarize that, analyze this spreadsheet, or give me 30 content ideas. Those uses are absolutely valuable, but eventually, you hit a ceiling.
Someone still has to provide the context, make the decision, review the output and figure out what happens next.
Meanwhile, the brand guide exists somewhere. The strategy is sitting in a document. SOPs were written six months ago. Important decisions happened during a meeting that only three people attended.
The information exists. The business just is not particularly good at using all of it together. In many founder-led companies, one person becomes responsible for connecting those dots: the founder. That was when I realized I had been thinking about AI too narrowly.
Your business already has intelligence. The problem is where it lives.
Think about all the judgment you develop from actually running a company. You can look at a campaign and sense when the message is slightly off. You understand which customers matter, where the business is going, and why a perfectly reasonable idea still might not be the right move. That is business intelligence.
But so much of it gets scattered across conversations, documents, meeting notes, Slack messages and the founder’s memory.
Hiring more people does not automatically transfer it, and adding another piece of software does not either. AI certainly cannot use business context it does not have. That became the problem I was interested in solving.
My own marketing became the testing ground.
Social media was one of the easiest places for me to see the difference.
Today, I can get roughly a week’s worth of social media planning, filming, editing and scheduling completed in about one hour.
The time savings are great, but what interests me more is why it is faster. I am not sitting down every week rebuilding the strategy from scratch.
The system already has the context behind the work: audience, positioning, messaging, brand standards, content strategy and the frameworks used to make decisions.
So much of the repetitive thinking simply does not need to happen again.
I think we’re asking the wrong AI question.
Business owners constantly ask, “What AI tools should I use?” I think there is a better place to start.
Where is my business repeatedly finding, explaining, interpreting, approving or recreating information it already knows?
That is where AI gets interesting. You probably do not need it everywhere, but there are places inside your business where the same decisions are being reconstructed repeatedly. Those are the places I would look first.
Because I don’t think the future of AI in business belongs to whoever collects the most tools or memorizes the newest prompting technique.
It belongs to businesses that figure out how to turn what they already know into intelligence their people and technology can actually use.
I started with questions, then prompts, then custom GPTs and workflows. Eventually, I realized I could build something much bigger.
And once I experienced that shift inside my own business, it became very difficult to go back.
Want to see what this could look like inside your business?
If you’re already using AI but still find yourself connecting everything, reviewing everything or repeatedly explaining what happens next, the next opportunity may be bigger than another tool.
See how Archeva turns business knowledge, strategy and founder judgment into systems powered by AI.