Said clearly on the earnings call , 150 million left for open reduction this calendar year
1 billion left in the overall plan , to be achieved with AI agents
Summary : more layoffs before end of June and anyone doing manual work to be replaced with automation in the next 12-24 months
@be I have to agree - from our own data policies. Most uses that a bot can do currently would be misuse. Because customers don't consent to bots servicing them and using their data yet. That's why most of our AI integration needs customer redaction.
Now I see some options, but, it's nowhere near as far as OP suggests
1)rather than risk an AI chat bot, automate instead. That way you're not risking $0 perm quotes, or Renewals, and service is uniform. If it's routine do you really need an AI's temperature or unique questions? Probably not.
2)they limit it to internal use persona bots. Something that was already hinted at. People who do specialized work, even if manual would just have something for specialized questions outside their dept. That means more advisory only depts are the ones getting the cuts. We do have dept like that, or have one specialized form that could be formatted consistently. However - those would have to be trained on clean employee data - which means you're telling them to build a replacement and trusting they don't poison/sabotage it.
I'm not going to act like my dept is immune, but if you make licenses all self-service you're going to have to up the amount of Compliance audits. Because there's no real tracker aisde from gen history and customer machine. And then probably need someone in the back or IT to alter counts.
Aside from regional contract requests, SLA issues, etc. OT would have to improve a lot to make that even feasible. And a lot of those older products just lost people that could have helped do that. Both for cloud options or just saving old legacy products customers won't budge from.
That's not hitting every dept equally. 2-3 bad vibecoded releases and customers will start leaving regardless of type. And I can't blame them. 2 very messed up AI driven deals from Sales or Renewals would likely scare OT into mandated, documented double checks.
AI guiding PS wrong means losing projects, I don't know if they're rated/penalized on performance.
I don't think OT is there yet - we're in a 'layoff for payoff' mindset.
@be You are assuming logic and well thought out decisions are part of the equation at OT.
@OP I don't see how manual labor gets replaced with AI at OT if our own AI can't help you find much of anything about a product you support. Internal support is abysmal. So it falls to garbage in=garbage out. AI won't magically fix anything. God help you if you have to use a resource out of your department. That starts the 'who's paying for this' kabuki.
Far as business optimization, we have bigger problems. Let me give you a peek at my world. Of the products I support, there's a decent on-prem one that's modular, scales, and is robust as well as flexible. I'm not privy to ancient tribal knowledge, but we bought that product to replace a cr-ppier one we had, I'll call it the legacy one. I've been doing this for over five years and it's the first I heard of this. And of course, we have a cloud version. This version is easy to use and appears to work well, as long as you stay in the OT ecosystem. It does't play well getting gazintas from somewhere and putting the gazoutas where they need to be. Doesn't integrate easily, is what I'm saying. It has a fraction of the functionality of the on-prem version. The cloud documentation is abysmal, compared to the on-prem. Probably because the on-prem had better docs when it was purchased and they were simply maintained.
There was a discussion in the product management group on how we can migrate the old legacy customers to the cloud. You know the orphaned product? You can't. There's no docs how to do it, no way to do it. Same with the on-prem. You have to migrate, just like deploying new. So in that world, why would you migrate to our cloud and endure the cr-ppy support, when it's just as easy to go to a competitor? Same amount of work, and you'll probably get a discount for doing so.
How do you optimize that? It would require a whole new ethos before you could even start, let along let AI try and do it.
My two cents.