Docusign is stronger than the “it’s just a wrapper” meme suggests, but it is not a data monopoly like Wolters Kluwer (which I recently wrote about) or RELX. The core of Docusign’s moat lies in workflow integration, compliance, and trust. Not in unique, protected datasets. That makes it a defensible moat, but also vulnerable in an AI world where software will become increasingly cheap and easy to replicate. In my opinion the distinction between a workflow system (software) and a truth system (database) will ultimately determine who preserves, or even expands, their long-term value.
While not a traditional data moat, do you think Docusign's 150 million private consented agreements with 10 million being added every month...is a form of a data moat. If they can actually create better templates and notifications for their users (eg. your agreement is missing these X items used across 10,000 other similar agreements?
I imagine they'll pay for mission critical stuff (security, legal, regulatory)...but experiment around productivity.
Even then, I think the larger principle remains: focus on core stuff. Anthropic and OpenAI are the most AI agent forward companies in the world...and still rely on Slack (instead of building replacement internal comms tool...maybe that changes but hasnt yet)
Those single-day selloffs directly to related AI tools.
Longer term sell-offs a combination of many things, including AI fears which reduce future opportunities as Brad Gerstner eloquently explained: "The profit pool available to the agentic layer is increasing. When that happens, the discount rate [and the] terminal value of those software companies plummets…It could be true that you’re not going to replace [Salesforce], but it can also be true that it’s never going to trade at 30x free cash flow again. It’s going to trade at 17x free cash flow because its available TAM in the future is now dramatically and permanently changed."
FWIW, I was at at a PE panel last week where the CPTO of a large UK company said that he will be replacing many SaaS tools with internally-built tools to reduce costs. AI coding agents have changed the buy-build calculus.
"Probably going to be a lot more site reliability and security engineer roles."
----
I seriously doubt this. Perhaps you missed the below article? It won't be long before AI will be able to run the whole software lifecycle from being given a task to producing the final output, completely tested and debugged.
Exclusive: Anthropic's new model is a pro at finding security flaws
Sam Sabin
5 Feb 2026
Anthropic's latest AI model has found more than 500 previously unknown high-severity security flaws in open-source libraries with little to no prompting, the company shared first with Axios.
Why it matters: The advancement signals an inflection point for how AI tools can help cyber defenders, even as AI is also making attacks more dangerous.
Really well written and argued overview of where SaaS is at and can potentially go based on AI trends.
Thank you for this. It’s an excellent overview and puts words to many things I’d felt or thought about but never really had the right way to explain it.
Right, so I started digging for the exact quote here and there appears to be 2 sources: an Axios interview (but I can't find the transcript) and his essay "The Adolescence of Technology", which has the following:
"Slow diffusion of technology is definitely real—I talk to people from a wide variety of enterprises, and there are places where the adoption of AI will take years. That’s why my prediction for 50% of entry level white collar jobs being disrupted is 1–5 years, even though I suspect we’ll have powerful AI (which would be, technologically speaking, enough to do most or all jobs, not just entry level) in much less than 5 years."
I'm much more skeptical of diffusion speeds than Amodei (I've worked for some very large and heavily regulated industries) but think he may be right on the underlying technology, albeit at the long end of his estimates.
Demis was asked about Dario's prediction at Davos...made similar point. Said that if AI can do 95% of a job, that's still too jagged and won't replace the role.
That was such an awesome weekly recap. Thanks Trung!
Thanks for the read, BJM555!
Docusign is stronger than the “it’s just a wrapper” meme suggests, but it is not a data monopoly like Wolters Kluwer (which I recently wrote about) or RELX. The core of Docusign’s moat lies in workflow integration, compliance, and trust. Not in unique, protected datasets. That makes it a defensible moat, but also vulnerable in an AI world where software will become increasingly cheap and easy to replicate. In my opinion the distinction between a workflow system (software) and a truth system (database) will ultimately determine who preserves, or even expands, their long-term value.
Thanks for the read and note!
While not a traditional data moat, do you think Docusign's 150 million private consented agreements with 10 million being added every month...is a form of a data moat. If they can actually create better templates and notifications for their users (eg. your agreement is missing these X items used across 10,000 other similar agreements?
Great article!
- since the bigger tech companies have the same tools… I wonder what they build?
- will they start to build stand alone AI-native products (like Optimizely did with Opal)
- or do they ramp up M&A?
- does org structure change in R&D?
Thanks for the read Ian!
I imagine they'll pay for mission critical stuff (security, legal, regulatory)...but experiment around productivity.
Even then, I think the larger principle remains: focus on core stuff. Anthropic and OpenAI are the most AI agent forward companies in the world...and still rely on Slack (instead of building replacement internal comms tool...maybe that changes but hasnt yet)
Did the market drop because AI is changing things, or because the big tech companies are overvalued?
Those single-day selloffs directly to related AI tools.
Longer term sell-offs a combination of many things, including AI fears which reduce future opportunities as Brad Gerstner eloquently explained: "The profit pool available to the agentic layer is increasing. When that happens, the discount rate [and the] terminal value of those software companies plummets…It could be true that you’re not going to replace [Salesforce], but it can also be true that it’s never going to trade at 30x free cash flow again. It’s going to trade at 17x free cash flow because its available TAM in the future is now dramatically and permanently changed."
FWIW, I was at at a PE panel last week where the CPTO of a large UK company said that he will be replacing many SaaS tools with internally-built tools to reduce costs. AI coding agents have changed the buy-build calculus.
"Probably going to be a lot more site reliability and security engineer roles."
----
I seriously doubt this. Perhaps you missed the below article? It won't be long before AI will be able to run the whole software lifecycle from being given a task to producing the final output, completely tested and debugged.
Exclusive: Anthropic's new model is a pro at finding security flaws
Sam Sabin
5 Feb 2026
Anthropic's latest AI model has found more than 500 previously unknown high-severity security flaws in open-source libraries with little to no prompting, the company shared first with Axios.
Why it matters: The advancement signals an inflection point for how AI tools can help cyber defenders, even as AI is also making attacks more dangerous.
...
https://www.axios.com/2026/02/05/anthropic-claude-opus-46-software-hunting
Good article on DocuSign! But once human workers are replaced by AI, wil there be any need for electronic signatures and lawyers? I hope not.
I like how you write you lovely apple fanboy.
Really well written and argued overview of where SaaS is at and can potentially go based on AI trends.
Thank you for this. It’s an excellent overview and puts words to many things I’d felt or thought about but never really had the right way to explain it.
SAas collapse .. nice perspective, especially on software development changes..
My take
https://open.substack.com/pub/feisal/p/the-software-collapse?r=9pv9y&utm_medium=ios
I write about Startups and growth strategies and love the idea of partnering and helping each other across the community Please subscribe here. https://open.substack.com/pub/keithnewman/chat?r=3d4ef&utm_medium=ios&utm_source=share
Amodei's view here is misrepresented; he said that 50% of entry level white collar work would be automatable. Still a big claim, but more realistic.
Thanks for the read.
Dario has said "AI could wipe out 1/2 of all white collar entry-level jobs".
Anderson Cooper asked directly about the quote and he said "Yes".
Will add link for extra clarification.
https://www.reddit.com/r/AgentsOfAI/comments/1ozqgej/ai_could_wipe_out_half_of_all_entrylevel/
Right, so I started digging for the exact quote here and there appears to be 2 sources: an Axios interview (but I can't find the transcript) and his essay "The Adolescence of Technology", which has the following:
"Slow diffusion of technology is definitely real—I talk to people from a wide variety of enterprises, and there are places where the adoption of AI will take years. That’s why my prediction for 50% of entry level white collar jobs being disrupted is 1–5 years, even though I suspect we’ll have powerful AI (which would be, technologically speaking, enough to do most or all jobs, not just entry level) in much less than 5 years."
I'm much more skeptical of diffusion speeds than Amodei (I've worked for some very large and heavily regulated industries) but think he may be right on the underlying technology, albeit at the long end of his estimates.
I'm definitely with you on longer timelines.
Demis was asked about Dario's prediction at Davos...made similar point. Said that if AI can do 95% of a job, that's still too jagged and won't replace the role.
Eventually will happen but not 1-5 years in his estimation: https://x.com/bearlyai/status/2013711266922258580?ref_src=twsrc%5Egoogle%7Ctwcamp%5Eserp%7Ctwgr%5Etweet