You’re running a trades business. Your crew is out on the tools, you’re dispatching from the truck, and your phone won’t stop ringing. Half the calls are customers asking when their invoice is ready, whether you service their suburb, or if they can move their appointment to Thursday. The other half are new jobs you can’t answer because you’re elbow-deep in a furnace replacement.

By the time you call them back, three have already booked someone else.

The standard advice is to hire an admin or build a customer portal. But hiring costs $40,000 a year plus super, and most portals just sit there because customers don’t remember their login or can’t be bothered creating an account. The question isn’t whether self-service is a good idea. It’s whether automating it actually pays for itself, and how fast.

Let’s work through the numbers.

What a Customer Portal Actually Does in a Trades Business

A traditional customer portal lets people log in to see their service history, download invoices, and maybe request a callback. In theory, that’s great. In practice, adoption sits around 15 to 20 percent because most customers don’t want another password, and the rest forget they have access.

An AI-powered customer portal is different. It doesn’t wait for someone to log in. It meets the customer where they are, whether that’s a text message, a phone call, or a link in an email. It answers the question immediately, books the appointment, sends the invoice, or escalates to a human if the job is complex.

The ROI case isn’t about replacing a portal. It’s about replacing the 20 hours a week you or your office admin spend answering the same questions over and over.

The Manual Work You’re Doing Today

Walk through a typical week. How many times does someone call to ask:

• Can you come out this week?
• Do you service my area?
• What did you charge me last time?
• Can I get a copy of my invoice?
• Do you do [specific service]?

Each call takes three to five minutes. If you’re getting 40 of these a week, that’s two to three hours of pure admin time. If you’re the one answering, that’s two to three hours you’re not quoting, not on the tools, not closing the big jobs.

Now add the calls you miss. If you’re doing $2 million a year and missing 10 calls a week, you’re losing one to two jobs a week at an average ticket of $800 to $1,500. That’s $40,000 to $150,000 a year walking out the door because no one picked up.

Then there’s follow-up. You send an estimate on Monday. By Friday, you’ve forgotten about it. The customer has moved on. We see trades businesses convert 15 to 25 percent of stale estimates when someone actually follows up. If you’re sending 100 estimates a year and only closing 30, another 15 to 25 are sitting there waiting for a nudge.

This is the work an AI-powered customer portal handles. Not the login screen, the actual work.

What an AI Agent Doing This Work Looks Like

Let’s break it into three pieces.

24/7 Dispatch Voice Agent

This is Omni Voice. It answers every call, even at 9 p.m. on a Sunday. The customer describes the problem. The agent qualifies whether it’s an emergency or a scheduled job, checks your calendar, and books the slot. It sends a confirmation text with the time, the tech’s name, and a link to reschedule if needed.

If the job is outside your service area or requires a part you don’t stock, the agent tells them upfront and offers to put them on a callback list. No wasted trips, no awkward conversations three days later.

One plumbing business in our network was missing 12 calls a week because the owner was on the tools and the admin left at 4 p.m. They installed the voice agent and picked up every single call. Conversion on after-hours calls was 60 percent. That’s seven extra jobs a week at an average ticket of $950. The agent paid for itself in nine days.

Estimate Follow-Up Agent

This is Omni Ops. Every time you send an estimate, the agent logs it. On day two, it sends a text: “Hi, it’s Enterprise DNA following up on the quote Sam sent for your hot water system. Any questions I can answer?” If they don’t reply, it follows up again on day five with a slightly different angle, and once more on day 14 before the estimate goes cold.

The message is tuned to the trade and the job size. A $600 tap repair gets a short, friendly nudge. A $15,000 HVAC replacement gets a more detailed message with a link to book a site visit.

We typically see 18 to 22 percent of stale estimates convert when this agent is running. If you’re sending 150 estimates a year and closing 50, that’s another 18 to 27 jobs without changing your pricing or your sales process.

Review and Reactivation Agent

This one does two things. First, it asks every happy customer for a review the day after the job. Not a week later when they’ve forgotten, not three months later when you remember. The day after. It sends a text with a direct link to Google. One HVAC business went from six reviews a year to 48 in six months. That’s the difference between page three and page one in local search.

Second, it reactivates customers at the right service interval. If you installed a hot water system 18 months ago, the agent sends a maintenance reminder. If you did an air-con service last winter, it reaches out in October. Reactivation typically brings back 8 to 12 percent of your past customer base every year, and those jobs close at twice the rate of cold leads because they already trust you.

If you want a structured way to map out which calls you’re missing and what they’re worth, we’ve built a simple worksheet. Grab the After-Hours Call Recovery Plan for Trades. It’ll take you 15 minutes to fill out, and you’ll have a clear picture of where the leakage is.

The ROI Calculation

Let’s work through a real example. You’re doing $3 million a year. You’re missing 10 calls a week because you’re on the tools or the admin is at lunch. Half of those are quoting opportunities. Your average ticket is $1,200. You close 40 percent of quotes.

Missed calls: 5 quoting opportunities a week × 52 weeks = 260 opportunities. At a 40 percent close rate, that’s 104 jobs. At $1,200 per job, that’s $124,800 in lost revenue.

Now add estimate follow-up. You’re sending 120 estimates a year and closing 50. If the follow-up agent converts another 20 percent of the 70 that went cold, that’s 14 more jobs. At $1,200, that’s $16,800.

Total recovered revenue: $141,600.

The cost of running these three agents is typically $1,200 to $2,400 a month depending on call volume and the complexity of your dispatch system. Let’s say $1,800. That’s $21,600 a year.

ROI: $141,600 in recovered revenue minus $21,600 in cost = $120,000 net gain. That’s a 6.5x return in year one.

And that’s before you count the time you get back. If you’re spending 20 hours a week on dispatch and follow-up, and the agents handle 70 percent of it, you’ve just freed up 14 hours. If your time is worth $150 an hour (and if you’re running a $3 million business, it is), that’s another $109,200 a year.

What This Looks Like in Practice

One electrical contractor we work with was doing $4.5 million a year with two office staff and the owner answering calls between jobs. They were missing 15 calls a week, and follow-up was nonexistent. Estimates went out, and if the customer didn’t call back, the job was dead.

We built them a 24/7 Dispatch Voice Agent and an Estimate Follow-Up Agent. The voice agent picked up every call. The follow-up agent tracked every estimate and nudged at day two, day five, and day fourteen.

In the first 90 days, they booked 47 jobs they would have missed. Average ticket was $1,850. That’s $86,950 in revenue. The follow-up agent converted another 22 stale estimates worth $34,000. Total impact: $120,950 in 90 days.

The owner’s words: “I didn’t realize how much money was walking past us because we were too busy to pick up the phone.”

That’s the ROI case. It’s not about technology. It’s about capturing the work that’s already there.

When It Doesn’t Make Sense

This isn’t for everyone. If you’re doing $500,000 a year and you personally answer every call, you probably don’t need a voice agent yet. You need a better CRM and a follow-up checklist.

If your jobs are highly custom and every quote requires a site visit and three follow-up conversations, automation won’t close the deal for you. It’ll handle the first touch and the reminders, but you’re still doing the heavy lifting.

And if you’re already running at capacity with a six-week backlog, capturing more leads won’t help until you add another crew. The agents are useful for qualification and scheduling, but they won’t solve a capacity problem.

The sweet spot is businesses doing $1 million to $10 million with a mix of scheduled and emergency work, where the owner or a small office team is juggling dispatch, quoting, and follow-up. That’s where the leakage is highest, and that’s where the ROI is clearest.

How We Build This for Trades Businesses

We don’t sell you a portal login screen. We build the agents that do the actual work. The AI audit for trades businesses is a 60-minute working session. We map your dispatch flow, your quoting process, and your follow-up gaps. You walk out with three things: a process map, a prioritized agent backlog, and a 90-day ROI model with your real numbers.

No deck, no discovery phase, no six-month implementation. We’re building the first agent in week two.

If you want to see what that looks like, book a 60-min Omni Audit. Bring your call logs, your estimate spreadsheet, and your calendar. We’ll show you exactly where the leakage is and what it’s costing you.

The Bottom Line

Automating customer self-service isn’t about giving people a login. It’s about handling the repetitive work that’s eating your time and costing you jobs. The ROI case is straightforward: count the calls you’re missing, count the estimates you’re not following up, and multiply by your close rate and average ticket.

For most trades businesses doing $1 million or more, that number is $50,000 to $200,000 a year. The cost of fixing it is $20,000 to $30,000. The payback period is two to four months.

The question isn’t whether it’s worth it. It’s whether you can afford to keep losing that much work while you’re busy on the tools.

We’ve built these systems for plumbers, electricians, HVAC contractors, and roofers across three continents. The pattern is always the same: the work is there, the customers are calling, and the business is too busy to pick up. The agents pick up. The ROI follows.

If you’re ready to see what this looks like with your numbers, book my Omni Audit and we’ll build the model together. Sixty minutes, three outputs, no fluff.

For more on how we’re using AI to solve operational problems in trades businesses, explore the insights and guides we’ve published. And if you want to understand the broader Omni platform, start with Omni and see how voice, ops, and apps work together to handle the work no one else wants to do.