https://fortune.com/article/why-do-thousands-of-ceos-believe-ai-not-having-impact-productivity-employment-study/

The article is interesting, and mentions that productivity actually decreased at the beginning of the age of computers, as measured by economic variables.

But, I think that the hidden thing here is that leveraging AI requires investment into changing processes, which takes both time in working hours and time in calendar days.

From my own experience, my job has gotten much more efficient due to my AI tools and triggers. What have I done with that extra time? Invested in yet more AI tools and triggers.

Eventually, I am going to run out of low hanging fruit, and start taking on additional responsibilities with the free time I have. Until that happens, any increase in efficiency isn't going to be able to be measured in our profit margins.

On another note, Claude for Excel dropped on Saturday. It became much more efficient for me to invest in AI. A lot less work to build new tools that would have taken me longer to do on Friday.

My productivity is very much up (as is my law firm's across the board). It's hitting our bottom line nicely. I'm not looking forward to Claude being priced for actual value in the future.

I would bet money it was and is costing my former company money.

Customers that use the AI assistant were measured less likely to call and talk to an engineer. When customers spoke to an engineer they were more likely to ask for a sales contact and then place an order.

A little difficult to measure what is actually going on but the AI tool answers while complete, were not as likely to show key insights the human engineer would explain. And since that company has the most expensive products it also likely means they used the AI tool to get an answer then run to a lower cost competitor and ask for the product. So maybe a lose lose proposition.

James Forsyth said:

---

My productivity is very much up (as is my law firm's across the board). It's hitting our bottom line nicely. I'm not looking forward to Claude being priced for actual value in the future.

---

This won't happen.

Claude isn't the only game in town, and things move quickly.

BusterAg said:

---

https://fortune.com/article/why-do-thousands-of-ceos-believe-ai-not-having-impact-productivity-employment-study/

The article is interesting, and mentions that productivity actually decreased at the beginning of the age of computers, as measured by economic variables.

But, I think that the hidden thing here is that leveraging AI requires investment into changing processes, which takes both time in working hours and time in calendar days.

From my own experience, my job has gotten much more efficient due to my AI tools and triggers. What have I done with that extra time? Invested in yet more AI tools and triggers.

Eventually, I am going to run out of low hanging fruit, and start taking on additional responsibilities with the free time I have. Until that happens, any increase in efficiency isn't going to be able to be measured in our profit margins.

On another note, Claude for Excel dropped on Saturday. It became much more efficient for me to invest in AI. A lot less work to build new tools that would have taken me longer to do on Friday.

---

Interesting. Excel is arguably one of the powerfully underutilized applications that exists and it's been ubiquitous for decades. Will be interesting to see how people elevate Excel to do things they didn't think were possible.

I'm going to check this out just for my own tinkering

Quote:

---

But, I think that the hidden thing here is that leveraging AI requires investment into changing processes, which takes both time in working hours and time in calendar days.

---

The bigger issue imo is that companies are slow rolling AI adoption because of data exposure fears: company data being stored/used to train future models that are available to the world, data leaked through prompt injection, bugs in the models, etc, or data exposed thru "shadow AI", where employees bypass IT controls and upload data to unauthorized tools. At some point, there will be a waterfall effect here, where that issue disappears.

YouBet said:

---

BusterAg said:

---

https://fortune.com/article/why-do-thousands-of-ceos-believe-ai-not-having-impact-productivity-employment-study/

The article is interesting, and mentions that productivity actually decreased at the beginning of the age of computers, as measured by economic variables.

But, I think that the hidden thing here is that leveraging AI requires investment into changing processes, which takes both time in working hours and time in calendar days.

From my own experience, my job has gotten much more efficient due to my AI tools and triggers. What have I done with that extra time? Invested in yet more AI tools and triggers.

Eventually, I am going to run out of low hanging fruit, and start taking on additional responsibilities with the free time I have. Until that happens, any increase in efficiency isn't going to be able to be measured in our profit margins.

On another note, Claude for Excel dropped on Saturday. It became much more efficient for me to invest in AI. A lot less work to build new tools that would have taken me longer to do on Friday.

---

Interesting. Excel is arguably one of the powerfully underutilized applications that exists and it's been ubiquitous for decades. Will be interesting to see how people elevate Excel to do things they didn't think were possible.

I'm going to check this out just for my own tinkering

---

In my experience, finance professionals in small to mid-sized businesses use Excel to do things that very much should be part of a relational database.

For largish companies, this is going to help with financial reporting, accounting, forecasting, budgeting, etc. Things that would normally take a good excel guru a month to do are likely going to be achievable in hours.

BusterAg said:

---

James Forsyth said:

---

My productivity is very much up (as is my law firm's across the board). It's hitting our bottom line nicely. I'm not looking forward to Claude being priced for actual value in the future.

---

This won't happen.

Claude isn't the only game in town, and things move quickly.

---

I hope you're correct.

BusterAg said:

---

YouBet said:

---

BusterAg said:

---

https://fortune.com/article/why-do-thousands-of-ceos-believe-ai-not-having-impact-productivity-employment-study/

The article is interesting, and mentions that productivity actually decreased at the beginning of the age of computers, as measured by economic variables.

But, I think that the hidden thing here is that leveraging AI requires investment into changing processes, which takes both time in working hours and time in calendar days.

From my own experience, my job has gotten much more efficient due to my AI tools and triggers. What have I done with that extra time? Invested in yet more AI tools and triggers.

Eventually, I am going to run out of low hanging fruit, and start taking on additional responsibilities with the free time I have. Until that happens, any increase in efficiency isn't going to be able to be measured in our profit margins.

On another note, Claude for Excel dropped on Saturday. It became much more efficient for me to invest in AI. A lot less work to build new tools that would have taken me longer to do on Friday.

---

Interesting. Excel is arguably one of the powerfully underutilized applications that exists and it's been ubiquitous for decades. Will be interesting to see how people elevate Excel to do things they didn't think were possible.

I'm going to check this out just for my own tinkering

---

In my experience, finance professionals in small to mid-sized businesses use Excel to do things that very much should be part of a relational database.

For largish companies, this is going to help with financial reporting, accounting, forecasting, budgeting, etc. Things that would normally take a good excel guru a month to do are likely going to be achievable in hours.

---

Absolutely agree with that. Same experience. Even in large companies.

AI for excel has been great for me. Since it has a set of rules to follow its a no brainer to use and trust the results,

It has helped me make several shortcuts for our team when I knew something was possible, but just didnt know the exact formula to get it done.

Does anyone else see Microsoft being the big winner here? I don't know about you all, but for my wife and I, both of our companies (mine is a financial/investment company and my wife's is a medical device company, just for point of reference) have completely shut off any other AI tool outside of Microsoft's Co-Pilot. I have a personal Claude account that we use across the family (my teens and wife and I) and I absolutely love using it for many different tasks, but I can't use it with my work computer at all. Given the fact that Microsoft has such a stranglehold on business machines (the only company I worked for that didn't use Microsoft was when I worked as an HRIS consultant a few years ago for a small, boutique consultancy that used Google), it seems like they have a head start on everyone else. Thoughts?

Shumba said:

---

Does anyone else see Microsoft being the big winner here? I don't know about you all, but for my wife and I, both of our companies (mine is a financial/investment company and my wife's is a medical device company, just for point of reference) have completely shut off any other AI tool outside of Microsoft's Co-Pilot. I have a personal Claude account that we use across the family (my teens and wife and I) and I absolutely love using it for many different tasks, but I can't use it with my work computer at all. Given the fact that Microsoft has such a stranglehold on business machines (the only company I worked for that didn't use Microsoft was when I worked as an HRIS consultant a few years ago for a small, boutique consultancy that used Google), it seems like they have a head start on everyone else. Thoughts?

---

I can't use AI for most of my work, but I've tried my company's two Copilot instances and they both are so far behind Claude and ChatGPT I don't see Microsoft catching up.

All of the different AI models have their strengths. I'm not sure what Copilot's strength is other than being approved for work.

Shumba said:

---

Does anyone else see Microsoft being the big winner here? I don't know about you all, but for my wife and I, both of our companies (mine is a financial/investment company and my wife's is a medical device company, just for point of reference) have completely shut off any other AI tool outside of Microsoft's Co-Pilot. I have a personal Claude account that we use across the family (my teens and wife and I) and I absolutely love using it for many different tasks, but I can't use it with my work computer at all. Given the fact that Microsoft has such a stranglehold on business machines (the only company I worked for that didn't use Microsoft was when I worked as an HRIS consultant a few years ago for a small, boutique consultancy that used Google), it seems like they have a head start on everyone else. Thoughts?

---

That is exactly what's happening at my company. Thus far, they have refused to allow any AI use other than Copilot because MS has rolled it out in Teams and VS. Copilot is one of the worst AI environments right now, based on OpenAI crap. My company is ******ed because they are so risk averse. MS doesn't have a head start, they have a large installed base. They bake Copilot into everything like they do with every new tech that comes out. MS wins by default a lot of times despite being a barnacle.

Elon claims by the end of this year his AI will be on par with all other AI in the world and that within three years it will be light years ahead of anyone else. That's drastic and rapid change that's going to happen.

LOL @ copilot users.

if copilot was a smartphone, it would be a windows phone.

I get that, and I agree, I think Claude runs circles around Co-Pilot. When I say that Microsoft has a head start, I mean that they have already become the "approved" AI for most IT departments, so what's to stop MS from going to Claude and saying, "hey, let's partner up, I got all of the end users and I will just roll your AI into Co-Pilot". I know that OpenAI is the current backbone for Co-Pilot, but I could see Microsoft making deals with many AI companies and rolling them into Co-Pilot. Maybe I'm wrong, but it seems like that is what they do. I mean, the MS Suite of tools (excel, powerpoint, etc) all had competitors back in the day, and some of them were way better than what Microsoft was using, but Microsoft seemed to have won based on user acceptance (and IT departments playing it safe and easy, i.e. not having to work with a different spreadsheet software because excel is already approved).

Claude is available for co pilot. Happened about a month ago

BadMoonRisin said:

---

Claude is available for co pilot. Happened about a month ago

---

Not inside my org.

Their strength is the connection to the data being built in. Can connect to your Sharepoint data and pull in results without having to search for the documents.

That being said, copilot is junk.

YouBet said:

---

BusterAg said:

---

https://fortune.com/article/why-do-thousands-of-ceos-believe-ai-not-having-impact-productivity-employment-study/

The article is interesting, and mentions that productivity actually decreased at the beginning of the age of computers, as measured by economic variables.

But, I think that the hidden thing here is that leveraging AI requires investment into changing processes, which takes both time in working hours and time in calendar days.

From my own experience, my job has gotten much more efficient due to my AI tools and triggers. What have I done with that extra time? Invested in yet more AI tools and triggers.

Eventually, I am going to run out of low hanging fruit, and start taking on additional responsibilities with the free time I have. Until that happens, any increase in efficiency isn't going to be able to be measured in our profit margins.

On another note, Claude for Excel dropped on Saturday. It became much more efficient for me to invest in AI. A lot less work to build new tools that would have taken me longer to do on Friday.

---

Interesting. Excel is arguably one of the powerfully underutilized applications that exists and it's been ubiquitous for decades. Will be interesting to see how people elevate Excel to do things they didn't think were possible.

I'm going to check this out just for my own tinkering

---

Excel is grossly underutilized at work, used mostly to track comment submissions.

During one of the woke doxxing phases I used Excel to scrape and archive all my TexAgs posts, do batch edits and even restore my original posts.

Mega Lops said:

---

LOL @ copilot users.

if copilot was a smartphone, it would be a windows phone.

---

I miss my Windows OS phone. That was a great phone there for a bit before they threw in the towel. So much potential but they couldn't get devs on board.

It has increased my efficiency at work, for which the time saved I just use for more personal time. It's great.

LoL

James Forsyth said:

---

My productivity is very much up (as is my law firm's across the board). It's hitting our bottom line nicely. I'm not looking forward to Claude being priced for actual value in the future.

---

Can you share how you use AI?

Shumba said:

---

I get that, and I agree, I think Claude runs circles around Co-Pilot. When I say that Microsoft has a head start, I mean that they have already become the "approved" AI for most IT departments, so what's to stop MS from going to Claude and saying, "hey, let's partner up, I got all of the end users and I will just roll your AI into Co-Pilot". I know that OpenAI is the current backbone for Co-Pilot, but I could see Microsoft making deals with many AI companies and rolling them into Co-Pilot. Maybe I'm wrong, but it seems like that is what they do. I mean, the MS Suite of tools (excel, powerpoint, etc) all had competitors back in the day, and some of them were way better than what Microsoft was using, but Microsoft seemed to have won based on user acceptance (and IT departments playing it safe and easy, i.e. not having to work with a different spreadsheet software because excel is already approved).

---

Good idea.
My company has Copilot but I would like to use Claude, so maybe they will market what you say above.

infinity ag said:

---

James Forsyth said:

---

My productivity is very much up (as is my law firm's across the board). It's hitting our bottom line nicely. I'm not looking forward to Claude being priced for actual value in the future.

---

Can you share how you use AI?

---

Other people have done that and you didn't respond to them. If you are having trouble, use Claude to look up use cases.

Logos Stick said:

---

infinity ag said:

---

James Forsyth said:

---

My productivity is very much up (as is my law firm's across the board). It's hitting our bottom line nicely. I'm not looking forward to Claude being priced for actual value in the future.

---

Can you share how you use AI?

---

Other people have done that and you didn't respond to them. If you are having trouble, use Claude to look up use cases.

---

Some people just gave vague replies. I am looking for details.

I am looking for how real people use it. "check out Claude" or "Google it" does not count.

infinity ag said:

---

Logos Stick said:

---

infinity ag said:

---

James Forsyth said:

---

My productivity is very much up (as is my law firm's across the board). It's hitting our bottom line nicely. I'm not looking forward to Claude being priced for actual value in the future.

---

Can you share how you use AI?

---

Other people have done that and you didn't respond to them. If you are having trouble, use Claude to look up use cases.

---

Some people just gave vague replies. I am looking for details.

I am looking for how real people use it. "check out Claude" or "Google it" does not count.

---

Others - real people - gave specifics and you disappeared. But if you can't figure out use cases in your own industry, then perhaps AI is not within your capabilities.

infinity ag said:

---

Logos Stick said:

---

infinity ag said:

---

James Forsyth said:

---

My productivity is very much up (as is my law firm's across the board). It's hitting our bottom line nicely. I'm not looking forward to Claude being priced for actual value in the future.

---

Can you share how you use AI?

---

Other people have done that and you didn't respond to them. If you are having trouble, use Claude to look up use cases.

---

Some people just gave vague replies. I am looking for details.

I am looking for how real people use it. "check out Claude" or "Google it" does not count.

---

For me, I am using both Claude Opus 4.6 and Claude Sonnet 4.5. I write an agent. If I need deep reasoning for something like an architectural analysis on a piece of software, I tell it to work with Opus. When I need to have it convert old code to newer code (COBOL to Java), I tell it to use Sonnet. If I need it to simply read a set of files, I don't need the power that either of those models can bring to bear, so I tell it to use a "cheaper" model, say GPT-4.1 or similar. I am doing all of this in Visual Studio Code using the GitHub Copilot chat session and the CLI.

YouBet said:

---

BusterAg said:

---

https://fortune.com/article/why-do-thousands-of-ceos-believe-ai-not-having-impact-productivity-employment-study/

The article is interesting, and mentions that productivity actually decreased at the beginning of the age of computers, as measured by economic variables.

But, I think that the hidden thing here is that leveraging AI requires investment into changing processes, which takes both time in working hours and time in calendar days.

From my own experience, my job has gotten much more efficient due to my AI tools and triggers. What have I done with that extra time? Invested in yet more AI tools and triggers.

Eventually, I am going to run out of low hanging fruit, and start taking on additional responsibilities with the free time I have. Until that happens, any increase in efficiency isn't going to be able to be measured in our profit margins.

On another note, Claude for Excel dropped on Saturday. It became much more efficient for me to invest in AI. A lot less work to build new tools that would have taken me longer to do on Friday.

---

Interesting. Excel is arguably one of the powerfully underutilized applications that exists and it's been ubiquitous for decades. Will be interesting to see how people elevate Excel to do things they didn't think were possible.

I'm going to check this out just for my own tinkering

---

I've tried 4 AI's, none did a truly great job with complex formulas and macros. Grok and Gemini pretty much defaulted to use python, which is fine for me, but not for folks that have become experts on spreadsheets. Claude does seem the way right now, followed by ChatGPT. This was before the latest drops of Claude and ChatGPT. I did think working with Claude did take more of a learning curve for new, non-techie users.

Troy91 said:

---

BadMoonRisin said:

---

Claude is available for co pilot. Happened about a month ago

---

Not inside my org.

---

I should qualify that, it's available to use on the Researcher Agent.

Cinco Ranch Aggie said:

---

infinity ag said:

---

Logos Stick said:

---

infinity ag said:

---

James Forsyth said:

---

My productivity is very much up (as is my law firm's across the board). It's hitting our bottom line nicely. I'm not looking forward to Claude being priced for actual value in the future.

---

Can you share how you use AI?

---

Other people have done that and you didn't respond to them. If you are having trouble, use Claude to look up use cases.

---

Some people just gave vague replies. I am looking for details.

I am looking for how real people use it. "check out Claude" or "Google it" does not count.

---

For me, I am using both Claude Opus 4.6 and Claude Sonnet 4.5. I write an agent. If I need deep reasoning for something like an architectural analysis on a piece of software, I tell it to work with Opus. When I need to have it convert old code to newer code (COBOL to Java), I tell it to use Sonnet. If I need it to simply read a set of files, I don't need the power that either of those models can bring to bear, so I tell it to use a "cheaper" model, say GPT-4.1 or similar. I am doing all of this in Visual Studio Code using the GitHub Copilot chat session and the CLI.

---

I'm sure you know this, but if not... they raised the subscriber rate limit ceilings to compensate for the higher token burn rate in Opus 4.7 to match the 4.6 equivalent. Boris posted that late last week on X. In case you are not using 4.7 because of the higher burn rate.

We're using it for medical notes, probably saves me 45 mins a day. Supposedly getting something to help with xray interpretation soon.

Logos Stick said:

---

Cinco Ranch Aggie said:

---

infinity ag said:

---

Logos Stick said:

---

infinity ag said:

---

James Forsyth said:

---

My productivity is very much up (as is my law firm's across the board). It's hitting our bottom line nicely. I'm not looking forward to Claude being priced for actual value in the future.

---

Can you share how you use AI?

---

Other people have done that and you didn't respond to them. If you are having trouble, use Claude to look up use cases.

---

Some people just gave vague replies. I am looking for details.

I am looking for how real people use it. "check out Claude" or "Google it" does not count.

---

For me, I am using both Claude Opus 4.6 and Claude Sonnet 4.5. I write an agent. If I need deep reasoning for something like an architectural analysis on a piece of software, I tell it to work with Opus. When I need to have it convert old code to newer code (COBOL to Java), I tell it to use Sonnet. If I need it to simply read a set of files, I don't need the power that either of those models can bring to bear, so I tell it to use a "cheaper" model, say GPT-4.1 or similar. I am doing all of this in Visual Studio Code using the GitHub Copilot chat session and the CLI.

---

I'm sure you know this, but if not... they raised the subscriber rate limit ceilings to compensate for the higher token burn rate in Opus 4.7 to match the 4.6 equivalent. Boris posted that late last week on X. In case you are not using 4.7 because of the higher burn rate.

---

My client is currently evaluating 4.7, so I am not currently using it. My team and I have discussed how we will use that one, if at all, once it is made available to us.