Best Electrician Jobs
Scale AI runs a data partnership program that pays companies for their internal records. Its list of industries of interest includes “Field Services (HVAC, Plumbing, Electrical).” That puts electrical contractors on the same list as software companies, insurers and law firms.
Most contractors have never thought of their job history as something an AI lab would want. But the buyers are specific about what they are after, and a lot of it sits in systems a mid-size electrical shop already runs every day. This guide covers what those records are, how big a company buyers say they want, what you should protect, and what to do before you talk to anyone.
Disclosure: I run howmuchismycompanydataworth.org, which introduces companies to AI data buyers and earns a referral fee from the buyer when a deal closes.
What the buyers say they want
Scale’s page is the clearest example. It asks for chat history from Slack, Microsoft Teams and Google Chat. It asks for email from Gmail and Outlook. It wants files from Google Drive and SharePoint, work tracking from tools like Jira and Asana, and records from business systems such as Salesforce, HubSpot and Zendesk. Scale describes those last ones as the CRM, HR, ERP and support records teams use every day.
The point is not any single document. Buyers want the trail of how real work gets done: a request comes in, someone asks questions, someone prices it, someone schedules it, a problem comes up, and someone fixes it. Buyers like micro1 say they use this kind of history to train AI agents that do work, not just answer questions.
An electrical contractor produces that trail all day.
The records your shop already has
You may not use Jira or Zendesk. You probably use something that does the same job. Here is how a typical electrical contractor’s records line up with what buyers list.
| What you have | Where it usually lives | Closest buyer category |
|---|---|---|
| Service calls and work orders | Field service software, dispatch board | Support and ticket records |
| Estimates, bids and proposals | Estimating software, spreadsheets, Drive | Files and documents |
| Dispatch notes and tech notes | Field service app, text threads | Work tracking |
| Change orders and RFIs | Email, project folders | Email and files |
| Customer emails and follow-ups | Gmail or Outlook | |
| Office and field chat | Teams, Slack, Google Chat | Chat |
| Service agreements and customer history | CRM or field service software | CRM records |
| Hiring, onboarding and payroll records | HR and payroll systems | HR records |
Years matter. A shop with eight years of work orders and estimates in one system has more to offer than a shop that switched software last spring and left the old records in a box.
Clean structure matters too. A work order with a clear problem, a diagnosis, the parts used and the final result tells a story. A ticket that says “done” does not.
The size buyers publish
This is where many electrical contractors will drop out, and it is better to know that now.
- Scale AI says 40 or more employees is its “preferred starting point,” though it says it looks at each partnership case by case.
- micro1 says on its data partnerships page that it wants established companies with 30 or more full-time employees, with data mostly in English. It is putting US companies first.
- Replay starts lower. Its published tiers begin at companies with 20 to 50 employees.
Count everyone on payroll: electricians, apprentices, service techs, estimators, dispatchers and office staff. A commercial contractor with three crews, a service department and an office team can clear 30 faster than the owner expects. A four-person residential shop will not, and no amount of record keeping changes that.
Headcount is a filter, not the price. Buyers also look at how many years of records you have, how complete they are, and how much of your work is in English.
What buyers publish about money
Nobody can tell you what your records are worth before a buyer looks at them. What we can do is report what buyers say on their own pages.
Scale lists partnership value of $10K to $1M+, which it calls illustrative and says scales with how often data is delivered. micro1 lists $100K to $500K for qualified partnerships, $500K to $2M for large datasets or ongoing participation across teams, and $2M+ for highly unique data. Replay lists $10K to $100K for companies with 20 to 50 employees and higher tiers for larger companies.
The closest thing to a trades benchmark comes from Polyshares, a company that licenses operational data and sells it to AI labs. Its track record page lists a $450,000 license with a roofing company, $300,000 with a general contractor and $275,000 with a solar company. Those are Polyshares’ own reported numbers. No electrical contractor appears on the list, and none of these figures predicts what your shop would get.
Treat every one of these as a buyer’s published range, not an offer.
What to protect before anyone sees a file
Your records are full of other people’s information. That is the part that can hurt you.
Customer addresses and access details. Residential service records hold home addresses, phone numbers and notes like “dog in backyard” or “key under the mat.” Some hold gate codes and alarm codes. Commercial records can include building access instructions and after-hours contacts.
Photos. Techs photograph panels, meters, nameplates and the inside of people’s homes. Many photos show house numbers, faces or license plates.
Employee data. Payroll, license numbers, wage rates, injury reports, drug test results and disciplinary notes. Your crew did not sign up to have any of that leave the building.
Confidential project information. General contractors, facility owners and data center clients often put confidentiality terms in their contracts. Drawings, security layouts and site details for critical facilities may be off limits no matter what a buyer promises.
What you told customers. Read your own privacy policy and service terms. If they say customer information is used only to provide service, that matters. The FTC warned companies in February 2024 that quietly changing terms of service or privacy policies to use data for AI training could be unfair or deceptive.
Buyers say they handle much of this. Scale’s page says it anonymizes data inside your environment before anything is extracted, removes personal information, and requires approval for each data package. Those are the right things to promise. You still need to check how they work in your systems and get the promises in writing.
Questions to ask any buyer
Before you sign anything, get clear answers to these:
- Which systems do you need access to, and which can I leave out?
- Where does the anonymization run, and who on my side reviews the output before it leaves?
- Can I exclude specific customers, projects or contract types?
- Is the deal exclusive? Some buyers are exclusive by default and some are not.
- Is payment upfront, ongoing, or both?
- What happens to the data if you sell your company or end the program?
- Who on your team will see raw records, if anyone?
If a buyer cannot answer these plainly, that is your answer.
A practical first step
Do not start by calling a buyer. Start with a one-page inventory you can finish in an afternoon.
Write down each system your company uses: field service software, estimating, accounting, email, chat, shared drives and HR. Next to each, note how many years of history it holds and roughly how complete it is. Then list your current headcount and any customer or GC contracts with confidentiality terms.
That page tells you three things fast. It shows whether you meet the size bars buyers publish. It shows which records are worth anything because they go back far enough. And it shows which records you should never offer.
Once you have it, you can run the answers through a short estimator that matches companies to the buyers that fit them. It asks nine questions about size, age, industry, where your records live and any limits on sharing. If you want the wider picture first, there is also a plain rundown of the different kinds of companies that buy business data, from data brokers to AI developers.
The short version
AI buyers are naming electrical contractors as a target. They want the records you already create: work orders, estimates, dispatch notes, customer email and team chat. The published size bars start around 20 to 40 employees, depending on the buyer. Protect customer addresses, access codes, photos, employee files and anything a client contract says is confidential. Make your inventory before you make a call.