Managed AI for Columbus Contractors and Field Service Companies

Is this just another login nobody in the office will use?

That is the right question to ask first. Most offices have already bought software that was going to fix the paperwork and ended up as one more tab open on the estimator’s second monitor. AI can fail the same way, and it fails faster, because a draft somebody has to rewrite from scratch costs more time than writing it once.

So the test we hold work to is whether it removes a step from a job the office already does. A workflow that adds a form, a field, or a standing meeting has failed that test and gets pulled. If a daily log template does not get the log written faster than typing it into the form, it gets rebuilt or dropped, and we would rather hear that in the first month than a year in.

The office already knows which parts of the day are wasted. What they have not had is a way to try a fix on a Tuesday without a budget conversation and a vendor call. In a year, the best AI ideas in your company should come from your people, not from us.

Where does this actually save time during bid week?

Most of bid week is reading. Somebody has to get through the invitation to bid, the spec sections that touch your scope, the front-end conditions, and the addendum that landed Tuesday afternoon, and then write a scope letter that states exactly what you are carrying and what you are excluding. That reading happens after hours because the phone stops then.

A workspace template takes those documents and returns a first-pass inclusions and exclusions list in your own scope letter format, with the spec paragraph cited beside each line. Your estimator reviews a draft against the documents instead of building one from a blank page. Pricing, judgment, and what you choose to exclude stay with the estimator, because that is where your margin lives.

The same mechanism handles addenda. An addendum that changes a few sentences across a long spec set comes back as those sentences with the pages named, so the check is a short read and not a second pass through the whole set. None of that decides the number you submit. It shortens the distance between opening the documents and being able to think about them.

What should we keep AI out of?

Some of this work should not go near an AI tool, and saying so up front is part of the setup. Contract terms, Ohio lien deadlines and notice provisions get read by your attorney. Stamped engineering stays with the engineer. A submittal you review and pass along carries your certification and your name, so that judgment stays with your project manager. Safety incident facts get written by the person who was standing there.

The larger exposure is confidentiality. Owner-provided documents often arrive under an NDA, geotechnical reports and model files included, and bid confidentiality clauses restrict what leaves your office. Your labor productivity assumptions and your pricing history are the two things a competitor would most want to read. Nothing stops an estimator from pasting that material into a chatbot account opened with a personal email address, and nobody in the company would see it happen.

The workspace answers that at two levels. Company AI use sits in one place with an owner and a record, so you can see what is being used and by whom instead of guessing across private accounts. And anything built on your documents, your pricing or your specific process stays inside your workspace and is never shared with another client.

ScaleSight AI runs on a SOC 2 Type II audited platform. The other half of that question, who can reach the file server and what leaves it, sits with TTS Cyber, the managed IT and cybersecurity company in the same Columbus office and the same Koine Cyber group.

Most of our payroll never opens a laptop. What are we paying for?

Seats are per person, and you buy them only for the people who would open the workspace. In a contracting business that is usually the office: estimating, project management, service coordination, safety and admin. If a superintendent would use it daily, that is a seat. If not, the information goes to a coordinator who has one, which is how that paperwork already moves, so your seat count follows the office and not the payroll.

The managed service is priced per seat plus usage. Every seat comes with a monthly usage allowance, and nothing is spent past the ceiling you put on the account, so a heavy bid month cannot produce a bill you did not plan for. Published budget ranges are monthly totals by seat count: up to 10 seats runs $250 to $500, 10 to 25 seats runs $500 to $1,175, and 25 to 100 seats runs $1,175 to $3,800. A $250 monthly minimum applies, and no per-seat rate is published.

Budget is allocated by team and department, never person by person. Estimating gets an allowance sized to bid volume, service coordination gets one sized to call volume, and it moves when the work moves. We also report back what kinds of work AI is being used for across the company, so a quarter in you can see whether it is going into scope letters and service follow-ups or somewhere you would want to steer it. Those are the two controls that let you leave this running: you decide what it can cost, and you can see what it is used for.

Why not just put ChatGPT or Copilot on the office computers?

You can, and that is where most offices start. Two things happen afterward. The first is that whatever your estimator builds sits on one AI company’s model and one AI company’s pricing. When that price moves or that model is retired, the work gets rebuilt on your time. Inside the workspace we re-point the same template at a different model and it keeps running, so what you paid to build stays built.

The second is that no single company is best at every task. Reading a long spec set, pulling live pricing off a manufacturer’s site, marking up an image, and drafting a run of short service follow-ups are four different jobs, and the strongest model for each is a different one. Each job is routed to whichever model handles it best, and the user never has to know which one they are using.

Then there is what you inherit. Work that is identical from one business to the next is built once and given to every client, so a company that starts in March gets everything built before March, and new templates, agents and workflows ship into your workspace every month out of that shared library. Anything built on your own documents, your own data or your own specific process stays in your workspace and is never shared.

The part that compounds fastest is internal. When one project manager works out a way to summarize a submittal log that holds up on a call with the architect, that method becomes a template every project manager in the company can run. When you need something specific enough that no template covers it, an AI Build Project is scoped work with a start date and an end date that builds that one workflow, template or agent and then ends.