How Local and Regional BSCs Should Adopt AI

Key Takeaways: 

- Use AI to automate routine tasks and improve productivity. 

- Test small AI solutions before scaling them across the business. 

- Start with operational problems, not AI tools.

By Connor O’Brien 

Most AI advice is written for companies with IT departments, innovation budgets, and full-time teams evaluating new software. 

Local and regional building service contractors (BSCs) operate in a different reality. In janitorial services, margins are thin, teams are lean, and technology only earns attention when it has a clear business benefit and a practical path to implementation. 

The good news is that AI can clear that bar when it is applied narrowly. The opportunity is not to “adopt AI” in the abstract. It is to find the parts of the business where work is repetitive, delayed, inconsistent, or trapped in spreadsheets, and then use AI and automation to make those workflows faster and more reliable. 

This article walks through the approach taken at ProEthic Building Services, a janitorial company serving the Phoenix metro area, and lays out practical steps any local BSC can take today. 

Start With the Friction 

Before buying software or building anything new, the first step is understanding where friction already exists in the business. The best AI opportunities usually come from the work people already know is frustrating. 

Before touching any technology at ProEthic, we interviewed 14 managers across sales, service, administration, and operations. The goal was not to ask, “Where can we use AI?” The goal was to ask, “Where is the business slowing down?” 

One of the most useful questions was, “If you could wave a wand and hand off or automate any part of your job, what would it be?” 

That question highlighted the real pain points: forms that delayed the team by a day, spreadsheets that went stale, information that lived in too many places, and processes that depended on a few people remembering the right steps. 

Those conversations became a ranked list of workflows worth improving. 

They also revealed something important: the process written in the standard operating procedure was not always the process people followed. That matters because AI is very good at helping with a process that can be clearly described. It is much less useful when the instructions are vague, inconsistent, or dependent on undocumented judgment. 

This is where technical judgment matters. 

AI has made software dramatically cheaper and faster to build, but customer data, pricing, employee information, and operational workflows should not be placed into tools nobody understands. If that expertise does not exist in-house, this is the one place where outside help can be worth it. 

Many internal software projects that would have required expensive custom development only a few years ago can now be prototyped for a fraction of the cost, especially when the workflow is narrow and well-defined. 

Regardless of who builds the tool, the most important habit is getting it into the hands of the people who will use it as quickly as possible. 

Early feedback shows whether the tool supports the real workflow or simply adds another layer of complexity. The users will say what makes sense, what feels awkward, what is missing, and what they would change if given the option. Then, before trusting the tool, run it in parallel with the existing process. 

Use the same quotes, same data, same accounts, and same inputs. Run the old process and the new process side by side. Compare the outputs. If the tool is working, it should match or improve the current process. If it does not, the parallel test will show where it needs to be fixed before the rollout creates problems. 

That validation step may feel redundant, but it is what separates a useful implementation from a risky experiment. 

What ProEthic Found 

At the end of a two-week discovery process, ProEthic identified 24 potential AI and automation use cases across service, sales, operations, and human resources. 

After prioritizing for business value, implementation difficulty, and clarity of the workflow, the first three major projects were identified. 

The first was lead generation and qualification. The problem here was that sales associates were spending too much time validating whether a lead was actually a good fit for janitorial services. 

Inconsistent office-size data in ZoomInfo meant the team often had to cross-check leads across the CRM, ZoomInfo, and Google Maps. That slowed down outbound activity and made it easier to miss new office openings or new buildings entering the market. 

As a solution, ProEthic created a pipeline that surfaced new Certificates of Occupancy and permit data directly inside the CRM. The system also used permitting data to help populate square footage for current accounts and new prospects where available. 

As a result, the sales team could identify new office openings earlier, reach prospects before they moved in, and spend less time switching between systems to validate whether a company was a good target. 

Instead of reacting after a business had already selected a cleaning provider, the team could become a helpful partner during the stressful period before an office opened. 

The second major project was the ProEthic Quoting App. The problem was ProEthic’s quoting process included multiple different site types including dealerships, schools, offices, and medical facilities. Because those locations required different knowledge and different inputs across all building types, this created challenges for sales and services teams. 

Sales associates and service managers had to work across multiple spreadsheets and forms to build accurate quotes. That created extra data entry and made it harder for every team member to confidently quote more complex accounts. 

ProEthic’s solution was to build a quoting app that allowed sales associates and service managers to measure rooms, capture site information, and generate polished customer quotes from a single workflow. 

The app made it easier for more people on the team to quote complex accounts. It also reduced the number of spreadsheets and forms involved in the quoting process, consolidating the work into one cleaner workflow. That improved speed, consistency, and the team’s ability to cross-sell across service lines. 

The third major project involved back-office operations automation. Like many local BSCs, ProEthic had built many of its operational processes over time. Some workflows started on paper forms. Others lived in Excel trackers. Others depended on systems like QuickBooks, OrangeQC, and Pipedrive. As the company scaled, those workflows became harder to maintain. Account onboarding, equipment management, consumable ordering, and other operational tasks required long checklists, manual updates, and constant coordination across documents and systems. Eventually, a significant amount of team capacity was dedicated to keeping those processes moving. 

To solve this issue, ProEthic built software to replace or streamline many of these workflows. Excel files were converted into reports, updates between systems were automated, and long checklists became structured workflows. 

Importantly, not everything was automated. Some steps still required human confirmation. For example, when a customer account was lost, a person still needed to confirm that equipment, such as a scrubber, had been removed from the site. That kind of human checkpoint is important. The goal of automation is not to remove judgment from the business. The goal is to remove unnecessary manual work while keeping people involved at the moments where their judgment matters. 

The new system reduced the administrative workload enough for ProEthic to operate with a leaner operations team while improving the consistency and quality of its data. The business now has clearer visibility into the current state of accounts, equipment, supplies, and operational work in progress. 

What BSCS Can D0

A full AI or automation program is too much for most local BSCs on day one. It also is not necessary. There are practical ways to begin without hiring anyone, changing systems, or launching a major build. 

One step is to write better prompts before buying new software. A specific, detailed prompt to a general-purpose AI tool can save hours of work today. For example, a BSC can: 

• Paste in their scope of work template and rough walkthrough notes and ask for a formatted scope of work. 
• Paste in a request for proposal and ask for every requirement, deadline, and submission item. 
• Ask for a first draft of a proposal using a past proposal as the style example. 
• Turn inspection notes into a customer-ready follow-up email. 
• Summarize a service issue and draft an internal action plan. 

The pattern is typically similar: Provide real context, give a specific task, and ask for a specific output. The more precise the input, the more useful the output is. 

Most BSCs already have valuable information spread across emails, proposals, pricing sheets, contracts, inspection reports, and shared drives. In ChatGPT these are called “apps” and in Claude they are “connectors.” A simple trip to settings can connect these to a chatbot. 

A chatbot or AI assistant that can search for those materials and answer questions can provide immediate value without changing the underlying process. 

The best starting point is usually not the biggest idea. It is the most annoying task that is easy to describe. Pick one workflow. Give it two weeks. Try to cut the time, steps, or manual effort in half. That workflow might be qualifying new leads, updating CRM records, or tracking onboard tasks for new accounts. 

In any case, measure the result. If the experiment works, move to the next workflow. If it does not, the company has only spent two weeks and has learned where the limits are. 

Getting real value from AI usually starts with something simple. Identify a specific bottleneck, describe the workflow clearly, and test a narrow fix, then repeat the process. For local BSCs, the opportunity is to become a clearer, faster, more efficient organization. That is how AI becomes useful in real janitorial businesses. 

Connor O’Brien is the founder of Colby AI, which helps janitorial companies use AI to find, qualify, bid, and win new business. He previously built data infrastructure at Apollo Global Management and worked as a management consultant for billion-dollar institutions including Lazard Asset Management, Neuberger Berman, and Wellington Management.