First, to reassure anyone who clicked with concern: no actual dog food was eaten in the making of this article.
"Eating your own dog food" is a tech industry saying. It means you use the things you make, every day, in your own business. The phrase is usually traced back to Microsoft in the late 1980s. The idea is simple: if it's good enough to sell, it should be good enough for you.
At 2Gen we take that literally. Well, not literally. You know what I mean.
What's in our bowl
Everything we offer clients, we built for ourselves first. Here's what's on the menu at 2Gen right now.
- Our own knowledge base, connected to Claude. How we run things lives in one place. Staff ask questions in plain English and get answers from our approved documents, with the source page linked.
- AI triage for new support tickets. Every new ticket is read, categorised and routed before a technician picks it up.
- Our own connectors. Claude reaches our ticketing, email and documentation systems through connectors we built. We built our own email connector because the off-the-shelf one couldn't handle attachments or properly formatted emails, which is a bit of a problem for email.
- Monthly IT reports in plain English. Each client's monthly health report gets turned into a clear email with a prioritised list of what to fix.
- Monitoring on every AI action. Every time Claude reads a ticket, searches email or drafts a reply through our connectors, it's logged: who, which tool, what it touched.
- This website. Yes, we built it using AI too.
Why we bother
It would be easier to sell tools we'd only read about. Plenty of people do. Here's why we don't.
We find the sharp edges first. Every tool has a moment where it does something weird. We'd rather that happen to us on a Tuesday afternoon than to you on the day of a big deadline.
We know what it costs to run. Building something is the easy part. Living with it for a year is the real test. We know which of our tools needed constant babysitting and which ones just quietly work.
Our feedback is fast and unfiltered. When something we built is annoying, our own team tells us about it. Quickly. Often before morning tea. There's no better quality control than a technician who has to use your tool forty times a day.
We only recommend what we'd use. If we wouldn't run it in our own business, we won't suggest it for yours.
Some of it tasted terrible
Not everything we cooked up was a winner. Some things were pretty sweet. Some weren't that flash. A few were genuinely crappy.
That's the point of eating it yourself. You find out fast. We learned three things along the way:
- It doesn't matter how clever the technology is if your staff don't use it.
- If it's complicated to use, you've probably made something complicated that didn't need to be.
- Build security, permissions and auditing in from the start. They're not things you want to worry about after everything is running.
The bad ones taught us more than the good ones.
What it means for you
When we recommend something, someone at 2Gen has already used it in real work, on a real day, when it mattered. We know where it shines and where it falls over. We know what it'll cost you to keep running. We also know what your staff will grumble about, because ours already did.
If you'd like to see one of these in action, have a look at how we keep business knowledge out of people's heads. It's the same setup we use ourselves. Same bowl, same food.