Super interesting, Sarah! What you’re describing is exactly Amazon’s game — when a product takes off, they produce it themselves and push it to the top, killing the competition.
The only way to fight back is to build a strategic interdependence, making it so they can’t do without you, just as you can’t do without them! I cover this more on @BrainInvest.
Thrilled to have discovered your newsletter — it’s fantastic!
Plug for a piece I've been looking for but haven't found. How should founders think about finding and cornering the "proprietary data or hard to access data" edge so many are talking about. I'd imagine things like depth and breadth matter a ton here.
Hi Sarah, thanks for translating well the provoking ideas! I'm curious if you've also considered in your post the workflows that are already needed and will be more needed as the JTBD shifts.
For instance, I imagine customer support is one such market you had in mind. We’ve seen a shift where companies need less of the traditional and extensible helpdesk features and more around changed existing workflows (eg AI-human collaboration, ticket escalation rules, and AI-driven knowledge bases), new workflows (eg explainability and control of AI and responses, easy back-office integration) and new needs now that they have more time (eg insights to be strategic).
As a practical example, we automate on avg 65% of tickets; our customers still pay more for helpdesk than us but are more and more open to changing their system of record.
In summary, how do you think the LLM companies will attack these verticals that seem obvious but require a lot of workflows?
I can see the plug here and in multiple systems you've described but only see this for the enterprise. I imagined I'm overlooking sth
In addition to what you mentioned, can you imagine other monetization channels making the GPT family of models OpenAI's version of the same cash cow as Meta ad platform + Google ad platform? Maybe through model brand search optimization, eventual SearchGPT ads (traffic in the billions), F500 recurring revenue deals on ChatGPT Enterprise, etc.
I feel like we're in a golden age of LLMs where the novelty and hype are making up for the lack of solid unit economics right now.
At the same time, I think it's rare to see a platform as sticky as ChatGPT & with that kind of data flywheel. It's such an interesting topic to think about how early we are in all of this.
Have you read Chernow's Titan about Rockefeller? It outlines the OG vertical integration strategy for market dominance. 744 is a lot to ask the reader, but it delivers (if you like history).
You can see how this influenced Apple, Amazon and others as they expand up / down their value chains to achieve dominance.
Good point, but I wonder if this will also start a trend of software providers limiting access to their APIs and implementing friction points that make it hard for this kind of 3rd party agents to work.
After all, these software providers will be selling their AI solutions too and need to give them unfair advantages.
Super interesting, Sarah! What you’re describing is exactly Amazon’s game — when a product takes off, they produce it themselves and push it to the top, killing the competition.
The only way to fight back is to build a strategic interdependence, making it so they can’t do without you, just as you can’t do without them! I cover this more on @BrainInvest.
Thrilled to have discovered your newsletter — it’s fantastic!
Plug for a piece I've been looking for but haven't found. How should founders think about finding and cornering the "proprietary data or hard to access data" edge so many are talking about. I'd imagine things like depth and breadth matter a ton here.
Completely agree, would love to chat
Hi Sarah, thanks for translating well the provoking ideas! I'm curious if you've also considered in your post the workflows that are already needed and will be more needed as the JTBD shifts.
For instance, I imagine customer support is one such market you had in mind. We’ve seen a shift where companies need less of the traditional and extensible helpdesk features and more around changed existing workflows (eg AI-human collaboration, ticket escalation rules, and AI-driven knowledge bases), new workflows (eg explainability and control of AI and responses, easy back-office integration) and new needs now that they have more time (eg insights to be strategic).
As a practical example, we automate on avg 65% of tickets; our customers still pay more for helpdesk than us but are more and more open to changing their system of record.
In summary, how do you think the LLM companies will attack these verticals that seem obvious but require a lot of workflows?
I can see the plug here and in multiple systems you've described but only see this for the enterprise. I imagined I'm overlooking sth
Thanks once again.
In addition to what you mentioned, can you imagine other monetization channels making the GPT family of models OpenAI's version of the same cash cow as Meta ad platform + Google ad platform? Maybe through model brand search optimization, eventual SearchGPT ads (traffic in the billions), F500 recurring revenue deals on ChatGPT Enterprise, etc.
I feel like we're in a golden age of LLMs where the novelty and hype are making up for the lack of solid unit economics right now.
At the same time, I think it's rare to see a platform as sticky as ChatGPT & with that kind of data flywheel. It's such an interesting topic to think about how early we are in all of this.
Great article!
Have you read Chernow's Titan about Rockefeller? It outlines the OG vertical integration strategy for market dominance. 744 is a lot to ask the reader, but it delivers (if you like history).
You can see how this influenced Apple, Amazon and others as they expand up / down their value chains to achieve dominance.
Interesting - I haven't. Thank you for the recommendation!
Good point, but I wonder if this will also start a trend of software providers limiting access to their APIs and implementing friction points that make it hard for this kind of 3rd party agents to work.
After all, these software providers will be selling their AI solutions too and need to give them unfair advantages.
time will tell...