The mission was simple in theory: build something that makes our own workday better.
In practice? It was messy, frustrating, and something else I couldn’t name in the moment.
The Brainstorming Buzz
We started with ideas flying everywhere. Together, we narrowed it down to three:
The Engagement Agent – my dream bot. I imagined a company agent that could log into the team’s LinkedIn profiles, mimic their tone, and boost engagement the moment a post went live. No more begging on WhatsApp for likes and comments while the algorithm punished us.
The Onboarding Bot – a friendly guide for new hires that would live in Google Chat. It would welcome them, hand out policy docs, add them to drives, boards, and groups, and basically save everyone from the dreaded “Hey, can I get access?” loop.
The Foodie Friday Bot – a peacekeeper for our weekly lunch wars. It would ask the team what they felt like eating, log all answers in Sheets, figure out the majority choice, and then send out a final summary by noon.
All three felt exciting. All three also looked like more work than we could chew in just two days.
Reality Check
That’s when the first dip of the roller coaster hit.
The LinkedIn bot idea? Too risky. OAuth tokens, privacy hurdles, and the looming threat of account bans. Out.
The Onboarding bot? Way too dependent on getting access and assets from leadership. Out.
So we pinned our hopes on the Foodie Friday bot. It wasn’t just feasible, it was fun. Everyone groaned about how much time lunch ordering wasted each week. Solving that together felt doable. And honestly, I just wanted to build something I could see working right away.
Thrown Into the Deep End
Then came the freefall.
My screen lit up with words like webhook, cron, JSON, and loops. I felt like I’d been dropped in a foreign country without a translator. I clicked, dragged, and connected nodes like I was piecing together IKEA furniture without instructions. This was my first real taste of n8n automation without coding.
At one point, I stared at a workflow and genuinely wondered:
Should all of this sit in one giant workflow?
Do I need separate workflows connected together?
How do I even make Google Sheets talk to this thing?
But here’s the twist: I did figure some things out. I managed to add JSON nodes, connect them to triggers, and even integrate a WhatsApp API that worked. That little victory felt like winning a medal. Until the bot didn’t fire. Or the JSON glitched for no clear reason.
Suddenly, every small triumph spiraled into confusion again. One step forward, two steps back. The jargon soup was real, and I was drowning in it. That’s when my technical teammates jumped in, took on the heavy lifting, and pulled the workflow out of quicksand. While they fought the hard battles, I focused on shaping the actual journey: what the bot should ask first, how menus should flow, how the final order should land in everyone’s chat.
FooFi Comes Alive
By the end of day two, our little chaos child had a name: FooFi. And it actually worked. For me, it was living proof that n8n automation without coding could build something useful in just two days.
- At 10AM sharp, FooFi sent each team member a DM to place their lunch order.
- With one click, you picked healthy or not. FooFi then showed real restaurants and menus.
- Every order got logged into Google Sheets with names and times.
- As noon crept closer, FooFi tallied the majority choice and asked stragglers if they wanted to join in.
- Finally, it shared the order summary with the team and pinged the designated person to place it.
No WhatsApp storms. No scrambling at 11:55. By 12:05, the order was done and dusted.
What Others Built While We Were Wrestling Lunch
Our team wasn’t the only one riding the chaos wave. The other projects were equally ambitious:
HR Khudkaar – an automated hiring system that cut hiring cycles from days to hours. Candidate scoring, future AI job description generation, and even LinkedIn integration were on the table
Muxi – a content automation engine. It grabbed trending topics, spun up blog drafts, added visuals, and generated voiceovers. What used to take hours now took minutes
SciC – an AI voice bot that called candidates, asked screening questions, and logged their answers through Twilio. It was like an HR assistant that never got tired
At Autonomous, we don’t just build bots for hackathons — we build automation that saves teams real time and stress.
Want to see what we can do for your workflows?
The Emotional Whiplash
Frustrating, infuriating, challenging, and gratifying (the something I couldn’t name earlier).
I went from wanting to slam my laptop shut to laughing at how absurd it felt when FooFi finally responded correctly. I realized that being non-technical doesn’t mean being powerless. With n8n automation without coding, I could still design, shape, and push things forward even if I wasn’t the one debugging a stubborn webhook at 2AM.
And yes, the creative spark is still firmly in the hands of us biz folks. The technical team can keep their JSON tangles. But seeing them also mess around with fun things like voice cloning and Canva slides? That was a reminder that hackathons are playgrounds for everyone.
P.S. I told Riz we should get everyone to do a video idea to creation next time. I did love to see the developers jumping hoops through cuts, edits, transitions and text overlays.
Final Ramble
This wasn’t just about lunch or bots or APIs. It was about crossing over into each other’s worlds, fumbling through jargon, and finding wins in unexpected places. FooFi might be a lunch bot, but for me, it became proof that automation isn’t locked away in engineering silos.
With n8n (you can also self-host it), grit, and teammates to catch you when you fall, even a marketer can build something real. And if I can do it, trust me, you can too.
