Using ChatGPT for estimating

It does need to be fact checked though. I’ve used chat gpt and grok to learn Spanish, code, do my taxes etc. I use it daily instead of google. I’ve noticed it getting stronger and stronger. Grok is better imo.
Grok, can get a hair up it’s but and do some weird crap though, but it’s hilarious

Have you tried to fact, check grouk
With ChatGPT and vice versa?
 
Grok, can get a hair up it’s but and do some weird crap though, but it’s hilarious

Have you tried to fact, check grouk
With ChatGPT and vice versa?
Yes. And you can try copy and pasting into a new instance of either to avoid ai confusion.
 
Yeah I wouldn't necessarily believe that.

I did ask it for examples of cases that went to trial and it listed a few where plaintiffs won. Those are things you can find in Google as well through the state court database. But ChatGPT summarized the case files to hit the high points of the case.


Sent from my iPhone using Tapatalk
 
And why would you not I have been for a year now.
I stoped using McCormick and use chat never go back
What are your thoughts on McCormick? I just transitioned from a Foreman role to now Project Manager/Estimator and I am now getting trained on the software. It seems the system tends to give lower man hours than actually needed?
 
I see that (AI) chatgpt is a mixed bag from the replies here when it comes to estimation. I'm curious, does anyone use or thought to use AI to automate other parts of the electricity business? and if yes, what do you use it for?
 
Job Safety Analysis and Work Permits, 2 week notices, lay offs, HR, Accounting Emailing you want to sit there and baby sit it it will do all that for you.

Don’t babysit it and your HR is going to fire you for not following your own rules you didn’t write, not to mention what the safety guy will do if he catches you pencil whipping safety sheets and someone gets hurt.

AI will be the multiplier of all the things killing our industry right now.

Office Engineers, copy paste prints, college degrees running jobs over time in the field, and no fear of consequences.
 
You've touched on something really important—the accountability piece. AI isn't a replacement for judgment, responsibility, or following proper protocols. It's a tool, and like any tool, it needs to be used correctly. The goal of automation isn't to bypass safety requirements or cut corners on HR compliance—it's actually the opposite. When implemented right, AI can improve consistency and accuracy in these critical areas. Instead of rushed, error-prone manual entries at the end of a long day, automation can ensure documentation is complete, compliant, and timestamped properly. For example, automated work permits can enforce checklists so nothing gets missed. Safety analysis tools can flag incomplete data before it becomes a liability. HR systems can ensure every step is documented correctly, protecting both the company and employees.

And here's where it gets even better: while you're ensuring compliance stays tight, automation handles the time-draining stuff that keeps you from growing. Automated calendaring eliminates the back-and-forth of scheduling—no more phone tag or double-bookings. Customer lead management systems can track inquiries, follow up automatically, and ensure no potential client falls through the cracks. These tools free up hours every week that you can reinvest in the field, in quality control, or in building relationships. This is especially powerful for a small business starting out with only a 1 man show or small businesses of 5-10 people.

The real risk isn't AI itself—it's people using it as a shortcut instead of a safeguard. When businesses approach automation with integrity, asking "how can this make us more compliant and safer while also more efficient?"—that's when it becomes powerful.

Demand for electricians is skyrocketing due to renewable energy installations (solar panels, EV chargers), data center expansion, smart home technology, aging infrastructure needing updates, and general construction growth. The work is growing faster than the workforce. Since electricians are in such high demand, they can command premium wages. But electricians are stretched thin with long wait times for both residential and commercial projects, so why not automate the mundane things like phone calls, scheduling, etc.

You're absolutely right that accountability matters. Technology should enhance our ability to do things right, not give us new ways to do them wrong. But it should also give us time back to focus on what actually matters.
 
I was taught this very young about a JSA, it’s to cover the company your working for, not the safety guy or workers, it’s your jobs to not be complacent when filling it out.

They have the lawyered up rules throwing others under the bus to save their skin.

If there is heavy lifting and you don’t write it down and you lift something heavy and get hurt, why wasn’t it written down your fault worker/foreman for not being safe and safety guy why didn’t you do something?

So till AI is there over your shoulder task by task to include remove heavy rubbish another dumb trade buried your materials with it won’t help.

I’ve worked both on safety driven and production driven jobs.

There are idiots everywhere and your own brain is the only thing to save your skin, using AI will dumb down the work force and make it worse.

Writing it down at least takes some brains, and when you see some of the JSA’s out there, AI hides the idiots that handwriting doesn’t, taking out a lot of red flags operators look for.
Like for instance “Kut the midle pip wit assataleen torch”.
 
You've touched on something really important—the accountability piece. AI isn't a replacement for judgment, responsibility, or following proper protocols. It's a tool, and like any tool, it needs to be used correctly. The goal of automation isn't to bypass safety requirements or cut corners on HR compliance—it's actually the opposite.

One huge problem with AI as 'automation' is the randomness at its core. Large language model generative AI works by generating a statistical model of a training text, using randomness to generate new text with the same statistics, and possibly double checking that the statistics are the same.

Unlike a tool programmed to automate a specific task, where you are trusting the programmer to have correctly implemented an algorithm, and have a chain of expertise that you can follow (for example a voltage drop calculator), AI tools don't have a specific algorithm that you can analyze. You are trusting the random training.

There are still valid ways to use this, but IMHO the burden of verification greatly reduces the benefits of the tool.
 
Top