Spencer wrote up the whole story. Approval workflows, prompt caching, error handling, and tool-loop control all went from workarounds to first-class APIs. Plus how we moved 2.5M+ messages to the new schema in four phases.
https://t.co/eMtevMlFBp
It’s been fun watching paradigms @spencerskovy created for our agents become core parts of RubyLLM 2.0.
AI agents are revolutionizing how human travel advisors do their jobs via Tern and it’s all running on RubyLLM
Our AI stack runs on RubyLLM, and we'd built up several workarounds the past few months.
We just upgraded to 2.0. A few of our use cases were promoted into first-class APIs, including our multi-phase migration pattern that was upstreamed!
https://t.co/jedqZFac2N
your biggest opportunity right now is finding industries that still operate extremely inefficiently because they were built when demand was huge and technology was limited
then ask: could this be run 10x more efficiently with technology that exists today?
there’s still a lot of time and a lot of markets left
technology is moving insanely fast. something that made zero sense yesterday can make perfect sense today
start with the boring industries: heavy equipment, construction, mining, etc.
Being a travel advisor is a growing and lucrative career for the people who love selling.
Lots of epic host agencies out there. Happy to make intros to folks interested in learning more.
I was curious to learn why some of my friends have recently quit their tech jobs to become travel advisors.
Apparently in 2025, "Travel Advisors" were the fifth fastest-growing job on LinkedIn
I looked into it - the top travel advisors on platforms like Fora make $800,000/yr planning trips for people. Kinda wild.
If you do it casually side, you could probably make over $10,000/yr easily (book 8–10 trips/yr).
It's probably one of the lowest-risk businesses you can start right now.
If you have great taste, you can basically monetize the things people already ask you for.
AI is the best thing that's ever happened to the travel advisor.
80-90% of their day is admin that doesn't use their expertise. Take that off their plate and give them back 10 hours a week, and a $127B industry becomes a $300B one.
Talked to @PhocusWire about it.
https://t.co/Tsk3uIbRMA
1000% this. Totally believe the biggest AI winners won’t be “sexy AI infra” but the tools that embed AI to transform entire industries.
At Tern, we’re building the AI layer that reimagines travel agencies/advisors. Not by replacement but by improving efficiency. Low margins and massively fragmented SMB/IC user base.
The point on adoption is spot on. The top owners get it right away and use it, but there is a long tail who don’t want to learn a new thing even if that thing makes life easier, so the play is to embed it and automate it and just show them the work has been done while they were away and is ready for review.
There are lots of good reasons to want in person work but “I can’t personally get anything done when my kids are in the house so I don’t think anyone else can either” is the weakest of arguments.
I talked to a founder today whose customers are all really nice people, because their initial market happens to be one that attracts such people. The founders themselves are nice for the same reason. They're enjoying starting a startup as much as any founders I've known.
46 of 64 buyers walked from the world's largest corporate travel company, citing AI.
Long Lake paid $6.3B anyway.
Their bet: AI makes a people-powered business more valuable, not less.
It's the exact bet we're making at @ternsoftware. For the independent travel advisor.
https://t.co/Le7mq2UZji
@aslater Agree.. lots of data that AI is already being used as a research tool for travel. But per Expedia, only 8% of users trust AI to book travel for them.
Trust is going to be the challenge here. Will be fun to see how it plays out.
AI will help human travel advisors immensely. Yesterday we launched Tern's AI Agent for travel advisors.
It does a lot of what you see these AI planning startups do, but focused on the channel that's already selling $2+ billion in travel a year on Tern.
The latest AI travel planning app (Odessia) is making the rounds. I tried it. It's slick. It won't displace human advisors though. Let me tell you why (thread)...
Humans are uniquely good at selling travel. Travel is emotional. Scary. Exciting. It's one of the largest discretionary purchases someone makes.
Trust matters more so than others. In high trust purchases, from enterprise software to travel, humans thrive.