We are pivoting away from doing enterprise AI transformations ("AI-native Palantir"). For now at least.
I've shared our key learnings below. I'll share a detailed blog post soon.
What's next? We are going to start moving insanely quickly on several other ideas. Stay tuned 😎
I've shared our key learnings below. I'll share a detailed blog post soon.
What's next? We are going to start moving insanely quickly on several other ideas. Stay tuned 😎
Over the last 3.5 months, we've worked very closely with many Global 2000 ($5B+ revenue) companies and deployed several custom AI agents.
Here are the most unexpected learnings:
Here are the most unexpected learnings:
1) It takes *way* longer than anticipated to actually build/deploy custom AI agents for large enterprises.
AI makes the engineering fast. But sales, product, system integration, and implementation are *incredibly* slow.
Customers don't know what they want, getting stakeholders aligned on new initiatives takes time, product requirements change, underlying data is missing / wrong, legal needs to approve everything, internal APIs aren't documented correctly. I feel very confident that AI will not be able to automate a lot of this.
AI makes the engineering fast. But sales, product, system integration, and implementation are *incredibly* slow.
Customers don't know what they want, getting stakeholders aligned on new initiatives takes time, product requirements change, underlying data is missing / wrong, legal needs to approve everything, internal APIs aren't documented correctly. I feel very confident that AI will not be able to automate a lot of this.
2) It takes way more time/effort to maintain the custom AI agents after they are deployed than anticipated.
Enterprise processes have countless edge cases that are incredibly difficult to account for up front. You essentially need full-time product managers to work with each customer post-deployment to account for these and maintain high accuracy.
Enterprise processes have countless edge cases that are incredibly difficult to account for up front. You essentially need full-time product managers to work with each customer post-deployment to account for these and maintain high accuracy.
3) The lack of use case consistency across customers makes product and GTM repeatability really difficult.
We worked on transformations across a variety of industries (insurance, energy, software, professional services) and departments (sales, legal, customer service, industry-specific business units). We were essentially building a product from scratch each time. The lack of repeatability kills scalability.
We worked on transformations across a variety of industries (insurance, energy, software, professional services) and departments (sales, legal, customer service, industry-specific business units). We were essentially building a product from scratch each time. The lack of repeatability kills scalability.
4) Small deals are just as much work as larger deals, but are just way less lucrative.
We thought it would be smart to get our foot in the door with smaller deals then quickly expand to larger deals and new use cases. It turns out the smaller deals are a ton of work too but way less revenue. We should have prioritized bigger deals from day one.
We thought it would be smart to get our foot in the door with smaller deals then quickly expand to larger deals and new use cases. It turns out the smaller deals are a ton of work too but way less revenue. We should have prioritized bigger deals from day one.
We have plenty of other learnings (good and bad). I will share them all in a blog post soon.
It's absolutely still possible to build a big business here, however, given our learnings + how fast the AI industry is moving, we believe that right now it's better to focus our attention elsewhere.
It's absolutely still possible to build a big business here, however, given our learnings + how fast the AI industry is moving, we believe that right now it's better to focus our attention elsewhere.
What's next?
We have a lot of ideas. We are going to start aggressively building/launching them.
What's most important is learning fast. Enterprise AI transformations are slow and so our learnings were slow. We need to 1) get focused and 2) "get into the arena". We believe AI is only just getting started. We have an amazing team and lots of ideas of things we believe should exist. We are going to start moving insanely quickly on them.
We may fail a lot but that's okay because we will be learning faster than everyone else.
We have a lot of ideas. We are going to start aggressively building/launching them.
What's most important is learning fast. Enterprise AI transformations are slow and so our learnings were slow. We need to 1) get focused and 2) "get into the arena". We believe AI is only just getting started. We have an amazing team and lots of ideas of things we believe should exist. We are going to start moving insanely quickly on them.
We may fail a lot but that's okay because we will be learning faster than everyone else.
Finally, building in stealth was a mistake. I personally plan to start building in public more and share my learnings.
Please follow along if you're interested!
Please follow along if you're interested!
• • •
Missing some Tweet in this thread? You can try to
force a refresh
