AlignAI的动态

AlignAI转发了

查看Rehgan Bleile (Avon)的档案,图片

Co-founder & CEO - AlignAI (getalignai.com) | making AI valuable and safe.

I have been having a lot of discussions on how real this major SaaS replacement using AI is for Klarna. And if so, how replicable is this? As mentioned on this last week's episode of All In, Chamath Palihapitiya talks about how the industry historically "never went after the system of record" software, like Salesforce or ERPs. They are too hard to rip and replace. Now we are hearing from Klarna that they have managed to do this with AI. Systems of record have a REALLY low tolerance for being wrong, especially in regulated industries. How much custom work is required to get AI to generate accurate enough results to do this? These systems are notoriously complex. They require A LOT of training for admins to maintain, and even then it is a struggle. Each implementation is different. Consulting firms make TONS of money customizing instances of this software to make it work for companies. Now Salesforce is shifting gears to enable their customers to build agents themselves with Agentforce. What does this mean for their underlying software? Did they see the writing on the wall? Many enterprises have leaned on SaaS to provide a framework to operate, automate, and collect/manage data (you tell us how to do it). These SaaS providers take usage and learnings across companies, build functionality, and deliver it across their customers. And they maintain it! Of course, there are some enterprises who think they know best and build their own. Then companies assign an owner of that system and the processes around it to coordinate efforts and make sure it is providing value (if they do it correctly). They are essentially relying on "what is everyone else doing?" combined with "who knows our business the best internally to make sure it works for us?" A downside (and upside) is that SaaS vendors are incentivized to generalize patterns (especially VC backed companies), so customization is usually always needed at large enterprises. If AI can build and maintain these systems, problem solved on that end of relying on external vendors to do it. If it can also solve the customization problem, even better!! However.... what happens when the knowledge sharing stops? The generalized patterns, although sometimes tough to customize, are also incredibly useful to introduce new frameworks and shifts in the way these enterprises operate. This is why Salesforce has been adopted across so many enterprises. It was a much better operating model than what they had before. This really begs the question of which parts will get open sourced and shared. Will the common patterns and frameworks just become freely available and the customization and maintenance will get automated by AI? "Software was eating the world, now it is eating itself". https://lnkd.in/gZ93bx-S

Klarna Plans to 'Shut Down SaaS Providers' and Replace Them With Internally Built AI. The Tech World Is Pretty Skeptical

Klarna Plans to 'Shut Down SaaS Providers' and Replace Them With Internally Built AI. The Tech World Is Pretty Skeptical

inc.com

要查看或添加评论,请登录