How AI is helping me build a company
We’re well into the early throes of the AI revolution, and every business is trying to work out what that means for them. The one question I keep hearing is, “how do we become an AI-first organisation?”. It’s clear that there are huge benefits in terms of cost and productivity from taking a tech-enabled approach, but making the shift is challenging and can be confusing.
When we launched Versapia, we had a unique opportunity. As an AI company focused on high performance culture, starting from scratch in May 2025, I realised that we had to build technology-first. Not only would it afford us significant advantages in speed and cost of execution, but I felt we owed it to our future customers to learn, first hand, how companies should be forged in today’s world.
We’re still very early in that journey, but we’re at a point where it makes sense to share what we’ve done.
I’ll start with some definitions. When people talk about AI-first, they almost always mean technology-first. Yes, there are core AI tools that make sense to leverage, but the value of automation is just as great. Tech-first is the most accurate definition of what we’re doing at Versapia. And tech-first doesn’t mean no people, but fewer people with much greater leverage and impact. Making sure those few are the right people is more important than ever.
Before diving into the specific tools we’ve used to build the business, there’s an overriding observation that’s important to understand. Just using technology won’t help you. You’ll end up spending a fortune on a vast set of apps that confuse everyone involved. It’s crucial to have absolute clarity on your objectives, your strategy and your tactics. Look for tools that can help you only once you’ve determined those three things.
We’re nowhere near done yet, but in less than three months we’ve made a huge amount of progress. Here are the things we’ve focused on, and the exact tools we’ve used:
Initial idea generation: I spent a lot of time going back and forth with ChatGPT when the concept of Versapia originally came up. Unfortunately, this was at the very heart of the LLM’s “sycophant” era, so there was a lot of “this isn’t just a SaaS business - it’s a world-changing, category-defining pivot worthy of the greats”, but over the course of several days it helped me to refine the initial concept and plan our approach to discovery.
Branding & website: After putting the initial brand kit (logo, colour palette and font set) together in Canva, which took a few hours, I used Squarespace to build the Versapia website, with videos hosted in Vimeo for a more premium feel. We’ve since used Canva extensively for decks, banners, social media content and other materials - it’s an absolute must-have tool for us.
Business planning: Back to ChatGPT here, but with a very specific focus. I created an “Analyst” GPT with expertise in business, market and competitor analysis, then set it to work scoping the TAM, SAM and SOM for our segment, as well as in-depth mapping of the competitive landscape. It’s worth noting that what was returned needed to be checked, and its conclusions weren’t always on the mark. Even GPT-5 is a PhD-level engine, whereas these activities benefit from commercial savvy and experience. Still, this cut my workload significantly and gave me a much clearer picture of how to position Versapia in the market.
One of my most highly-recommended hacks is creating a “board of GPTs” to challenge your thinking and highlight assumptions, blind spots and alternative ways of approaching things. I set up a Chair GPT as an experienced, successful B2B SaaS entrepreneur and investor, a CTO with expertise in B2B software and AI, particularly prompt engineering and LLM development, a Chief Marketing Officer with deep sector knowledge and a CFO, all of which I put to work analysing and improving our business plan.
The two key tricks here are making sure you’re specific about the references they draw from (identify specific domain experts and name them specifically if you can), and don’t feel you need to take the advice they give you. I ignored maybe 80% of the feedback provided, but there were some real nuggets that sharpened our approach.
Developing a prototype: This is where AI really came into its own. We used our CTO GPT to create an initial spec for the Versapia prototype, then jumped onto Replit to vibe-code our way through it. By hooking a Flask app up to an OpenAI Assistant, Typeform and SendGrid, we built a working prototype that’s going out for initial beta testing this week (if you want to try it out, you can apply here). We didn’t know how to do any of this when we started out, but the CTO GPT guided us through it. The development cost of the prototype was less than £150.
Marketing: There are so many marketing tools available today, it’s hard to know where to start. Aside from our CMO GPT, we’ve used Midjourney for high-quality image generation, Munch for chopping long-form video up into short social media reels with subtitles and Mailchimp to send out our weekly Versapiens blog.
Sales: We’re not doing much in sales yet as we’re pre-launch, but we are proactive in building an audience. We chose Salesforce as our core CRM - although not the fanciest, it’s got great scalability - and are starting to experiment with Lemlist for automated yet personalised outreach and lead generation.
Alongside these tools, we’re using the usual productivity and communication apps. The next phase of infrastructure for Versapia is focused on building agents to drive our marketing and Go-To-Market activity, and bringing Veriti (our core AI culture expert) to life as a virtual influencer.
There are plenty of companies out there far more advanced than us in their use of technology, but having gone from a standing start to a fully established brand, a prototype ready for beta testing and a robust, validated business plan in less than 90 days with just two people is something I’m very pleased with - and I’ve managed to take a two week holiday in that time.
Crucially, this would not have been possible even twelve months ago. Just imagine what we’ll be able to do in 2026…
(P.S. If you know someone who needs to read this today, send it to them and encourage them to subscribe to the Versapiens blog. If you haven’t subscribed yet, come join us on our journey through the intersection between culture, technology and business.)