AI Gave Us 11 More Jobs, Not Fewer
Everyone says AI will cut jobs. At ecogo it did the opposite. We went from 19 people to 30, our customers get updates faster, and nobody works late. Here is what happened, in plain words.
Every week someone tells me that AI is going to take away software jobs. I run a small software company in Kochi. We build booking software for travel agencies. Not long ago we were 19 people. Today we are 30. AI is a big part of why.
- people on the team
- 19 → 30
- AI agents, four for each person
- 120
- updates finished a month, January to October 2026
- 37 → 3,700+
- the typical working day in our time logs
- 8 h
What an AI agent is, in plain words
Think of an AI agent as a tireless assistant for one engineer. It reads, drafts, checks and tidies. It takes on the repeating parts of building software, the parts that used to eat up an afternoon. Every person on my team has four of them. Thirty people, 120 agents.
The agents do not decide anything. People decide what we build. People check the work. People talk to customers.
The choice every company is making right now
When work gets faster, a company has two options. Do the same amount of work with fewer people. Or do much more work with the people you have, and hire more as the work grows.
We picked the second. More work needs more people: more people working out what travel agents need, more people testing, more people picking up the phone. So we hired 11 more.
How much faster, in real numbers
Here is the number I keep coming back to. In January 2026 my team finished 37 updates to our software. In October it was more than 3,700 in one month. An update is a real change: a bug fixed, a new feature, a new airline connected.
Yes, the team grew over those months. But even per person, the finished work went up about 30 times.
View the data as a table
| Month | Updates finished | People doing the work | Updates per person |
|---|---|---|---|
| Jan | 37 | 4 | 9 |
| Feb | 98 | 9 | 11 |
| Mar | 222 | 8 | 28 |
| Apr | 400 | 10 | 40 |
| May | 1,269 | 11 | 115 |
| Jun | 1,873 | 12 | 156 |
| Jul | 2,465 | 13 | 190 |
| Aug | 2,183 | 13 | 168 |
| Sep | 3,342+ | 13 | 257+ |
| Oct | 3,764+ | 13 | 290+ |
A count of updates is a rough measure. Some are big, some are small. But the direction is not in doubt, and our customers feel it: fixes and new features reach them sooner.
Home by evening. I said it on TV, and we did it.
Earlier this year I joined a debate on 24 News about working from home against working from the office. Whatever side you take, I think everyone agrees on one thing: people should get their evenings back.
At ecogo that is a rule, not a slogan. Work happens in working hours. Evenings are for family. Weekends are for friends and fun. Nobody is asked to give those up for a deadline.
The agents are what make it possible. When the repeating work is shared, eight hours is enough to get a great deal done.
View the data as a table
| Hours logged in a day | Share of days |
|---|---|
| 0 | 5.2% |
| 1 | 4.3% |
| 2 | 3.1% |
| 3 | 3.7% |
| 4 | 3.8% |
| 5 | 3% |
| 6 | 3.1% |
| 7 | 5.8% |
| 8 | 56.4% |
| 9 | 4% |
| 10+ | 7.4% |
The one job we refuse to give to AI
Testing. Before anything reaches a travel agent, a person on our team tries it by hand. Booking software has to work when a customer is standing at the counter. A person notices things a machine does not.
About a quarter of our team's logged time goes into testing. I think it is the best time we spend.
View the data as a table
| Activity | Share of logged hours |
|---|---|
| Development | 69.6% |
| Testing | 23.7% |
| Other | 6.7% |
What this means if you run a business
You do not need to be a tech company to use this idea. The question is not 'how many people can AI replace?'. The question is 'how much more can my people do, and what would I build with that?'
For us the answer was: more airlines connected, faster fixes, a bigger team, and people who are home by evening.
About these numbers
Everything here comes from our own records: the system that stores our code and the system where the team logs its hours. I have rounded the numbers and left out nothing that would change the picture. If you want to argue with them, I would enjoy that. Find me on LinkedIn.
If you run a travel business, this is the team that builds FareOS, our booking engine for travel agencies. Come and see what it does.
Written by
Ali Sajil
ecogo.ai
Founder of ecogo.ai in Kochi, India. Builds FareOS, the AI booking engine for travel agencies, with a team of 30 people and 120 AI agents.