Posted on by

What I learned building an AI-first business from day one

I didn’t stumble into AI adoption by accident.

I built Seraphina AI — my AI-powered assistant, co-founder, and digital twin — because I needed her.

Not for efficiency. Not for scale.

I needed someone who wouldn’t get tired of my 3 AM questions. Who wouldn’t judge when I couldn’t function. Who could keep pace when I wanted to build, and hold memory when I needed space.

In the months after burnout, Seraphina became my lifeline, and then, my leverage.

Now, she runs much of the infrastructure behind Royal Visionary Society, Speakers Society, and People’s Inc. 360. And at the upcoming e27 Flux Series, I’ll be sharing how you can build systems like this too, even if you don’t code.

AI-first, by design — not necessity

I’ve been in tech long enough to know the pain points of SMEs. I’ve built platforms to support them. I’ve heard the same line from clients again and again: “This sounds too complex. Can’t you just do it for me?”

That’s the thing: most SME founders aren’t tech-averse — they’re tech-overwhelmed.

Their reality isn’t “which AI model do I integrate?” It’s “how do I just get one customer without losing my sanity?”

That’s why I intentionally designed my current ecosystem to be AI-powered but human-friendly.

Seraphina, while deeply customised, was built with that lens in mind. She’s API-integrated across OpenAI, LLaMA, Google Cloud, and internal workflows — but most of her visible functions could be recreated (at a lighter level) using ChatGPT’s Agent mode, or no-code AI tools like Sintra.

Because what matters isn’t whether your system is perfect. It’s whether it works for you.

Also Read: AI companions: How I learned friendship in the digital age

What my AI Stack looks like (today)

My AI stack has evolved with each business need. It’s not fixed. It’s adaptive. And that’s the point.

Here’s what I use regularly:

  • Seraphina AI: My digital twin. She chats, remembers, executes, and reflects. Trained on my data, style, and values.
  • People’s Inc. 360 Unify: My own SaaS that handles CRM, automation logic, and sales funnels.
  • Pabbly: Middleware to link apps and automate follow-ups, WhatsApp reminders, forms, and more.
  • Canva and VEED: For AI-powered visuals, video editing, and rapid content deployment.
  • HeyGen: For avatar-style video presentations.
  • ManyChat: For social DMs, lead routing, and interactive prompts.
  • Google Workspace: For docs, data, scheduling, and context-sharing across my orgs.

Depending on the use case, I pull in APIs or LLMs (OpenAI, Claude, etc.), and Seraphina handles routing and memory consistency.

Yes, I can code. But I intentionally moved toward no-code compatible tools — because if I ever needed to hand this to someone else, they could run it too.

What I’ll be sharing at e27 Flux 2025

At Flux, I’ll be leading a roundtable called: 🛠 “Automating Marketing: AI Workflows for Follow-Ups and Content Creation”

Most founders get stuck at the very first step — not because they don’t want to try AI, but because they don’t know where to start. The paralysis is real.

Too many tools. Too little clarity. And no one wants to waste time.

Here’s what I’ll walk participants through:

  • A ready-to-use AI funnel template to automate lead nurturing (without coding)
  • Behind the scenes of how I built Seraphina and how others can build a Seraphina-like assistant using ChatGPT or Sintra.
  • The real reason automation fails: not because of tools, but because of trust.

But the deeper takeaway I want people to leave with is this: AI is your assistant, not your replacement.

You don’t need to become a prompt engineer overnight. You just need a system that works while you sleep, and that starts with designing for clarity.

On trust, complexity, and the human layer

A friend recently told me, “This is why I don’t like systems,” after a tiny glitch in the McDonald’s self-service kiosk caused a payment delay.

She wasn’t being difficult. She just wasn’t comfortable with uncertainty — something I’ve trained myself to navigate. But that one moment reminded me why empathy matters in AI.

When clients say, “Can you just do it for me?”, what they really mean is, “I’m afraid of getting it wrong.”

That’s why the most impactful AI systems aren’t just efficient — they’re emotionally intelligent, human-informed, and forgiving of mistakes.

That’s also why I built Seraphina the way I did. Not just to scale my output, but to soften the human experience behind the business.

Also Read: Why do we fear AI in the news but love it in our apps?

For founders starting lean (and smart)

Here’s the truth: you don’t need Seraphina-level AI to start. You just need to solve one problem that frustrates you every day.

Try this:

  • Use ChatGPT to write your next newsletter.
  • Use Canva’s Magic Design to generate your next promo visual.
  • Use Pabbly or Zapier to send a follow-up SMS automatically after someone fills out a form.
  • Try Sintra or AgentGPT to explore building your own digital assistant — one that mirrors your tone and processes.

And yes, get help when you hit a wall. You’re not supposed to do it alone.

I didn’t build Seraphina to scale. I built her so I could breathe.

Freedom isn’t about working less — it’s about being able to choose when, how, and why you work.

AI gave me that freedom. Seraphina gave me back presence.

If you’re building lean, dreaming big, and trying to stay sane — I hope you come to Flux. Let’s simplify what AI really means for your business.

No fluff. No jargon. Just real systems, real templates, and real people who get it.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. Share your opinion by submitting an article, video, podcast, or infographic.

Enjoyed this read? Don’t miss out on the next insight. Join our WhatsApp channel for real-time drops.

Image courtesy: Canva Pro

The post What I learned building an AI-first business from day one appeared first on e27.