Thread regarding Dell Inc. layoffs

To all the workers out there, the cruel irony is that every bit of data you feed to the AI machine is a further step toward your own career extinction. You’re helping build the beast. It will be your undoing.

It is horrible. Good data is going in, but saleschat can’t find it with multiple prompts.

What is sales chat based on? Is it a Microsoft or Salesforce thing?

HAHAHAHAH we got another nastygram from our boss today about low SalesChat adoption.
"delete"

@bm a decline in CSAT can thus be attributed to that bot behaviour, you're not to blame

Saleschat was probably backed by someone very senior at Dell. That's they they're pushing it so hard. If it was only about replacing all of us they could have done that with Dellchat. It's a much better tool!

@OP when sales chat gives me a wrong answer I give it a thumbs up!

I find the AI-adoption at Dell truly hilarious. First they introduce the DellChat, make people take courses on "prompt engineering" and telling everyone that they need to learn and adopt to stay relevant. Then a short while after, they sunset the whole thing and replace it with SalesChat which is objectively worse in any sense of the matter. Last but not least they introduce some arbitrary KPI and try to force people to use the inferior tool.

Perhaps AI is short for 'All In' rather than anything related to intelligence.

If any leadership is looking, the answer for low adoption has nothing to do with me fearing it replacing me, it has to do with being an objectively terrible AI model. I've asked it fairly basic questions; Who is the PLM for x-line, what is a SKU for this particular type of transceiver, what servers are compatible with x and y parts in tandem, etc. and without fail the answer is always wrong. The PLM one was particularly egregious since it tried to say that there was no information available on PLMs which was objectively wrong.

I don't even ask it questions to meet the KPIs, I want them to make a stink about me not using it so I can say to their faces that I'm not going to use time wasting an inferior tool just to make them happy about their sh---y investment.

Until whatever is sp-t out of the AI is at 100%, you still need to check it. If you don't know if it is right or wrong, how can you send it? Then you'll need to research it, verify it etc etc.
People are conflicted because they know eventually it will be the catalyst to get rid of people.

Output of the model needs to be checked so that takes extra time. So a deterrent if you are being measured on customer satisfaction and your job depends on it.

I'd add the KPI to measure adoption is pretty unrealistic. They require us sales to launch at least 25 queries per week, and this is stupid since it often happens I don't need to interrogate a bot so often.
To keep my numbers green I just end up chatting randomly with the bot.

that's because it's really bad. period.