How business owners are growing sales with AI agents

The most creative sales agents we’ve seen are coming from paving contractors, package-room operators, and local service firms.

Michael Appell’s paving business is 25 years old. This year, he grew from 6 to 25 franchisee territories and added over $2M in sales. He’s one of many business owners running their entire sales function on Hyperagent.

Sales excellence has always been rationed: The business owner or a top seller gives most of their attention to a handful of big accounts, and everyone else follows a template.

Agents change that math. Every account can now get what you used to save for the biggest customers: deep research, custom assets, and replies within minutes.

You can now treat every account like it’s your only account.

We’ve broken down these sales agents into three core jobs: targeting, outbound, and monitoring. Then we distilled those lessons into an agent template that works for any business.

Prospector Your sales agent for research, personalized outreach, and replies. Get the template

Targeting: find accounts with time-sensitive buying signals

Most buying-signal data comes from a picklist. Industry, headcount, funding round, new hires, a job posting. Everyone filters on the same fields.

But the signals that matter most to a services business are too specific and too local to ever be a column in a database. They live deep in permit filings, construction updates, regional websites and satellite photos.

HelloPackage runs package rooms for multifamily and student housing: 200 buildings, 70,000 residents and more than 10 million packages a year.

The best time for HelloPackage to sell is while a building is still going up. That’s when the developer is deciding how the package room will work. No sales data tool has a filter for “multifamily buildings under construction in this metro, by stage.”

So CEO Paul Favorov did it by hand. The data was spread across CoStar, Salesforce, regional websites and spreadsheets. Prospecting a single region took months, and a custom pitch only happened when Paul had the hours to build it.

Now a research agent does this job. It scrapes large regional databases, maps every building under construction by stage, and builds an individualized pitch and outreach plan for each one. Its first run covered 27 markets and surfaced about 800 opportunities in a single day.

Paul says that the agent did in a day what “required probably months of manual web scraping” previously.

The agent investigates each account the way a human researcher would, using its own browser and computer reading whatever sources exist for that building, and it does this for hundreds of accounts in parallel.

Outbound: generate unique visuals for every account

Michael Appell was paying lead-gen services $1,500 to $4,500 a month to find commercial properties ready to buy. The promised meetings often didn’t even show up.

So he built his own sales agents.

Agents are given a set of ZIP codes where Appell has a franchisee. They use satellite and street view imagery to find commercial properties with large lots that are due for maintenance. They learned to skip corporate accounts where a local email will never reach the buyer, a distinction Michael says he never told them to make.

For each property, an agent pulls the current aerial image of the lot and generates an “after”: the same lot with fresh sealcoat and crisp new striping. It puts both images in a branded email, signed with the local franchisee’s name and phone number, and sends it from that franchisee’s inbox.

The property manager opens the email and sees their own lot, repaved and restriped.

Michael kept a human in the loop at first. He reviewed every email for the first few weeks before letting the agent send on its own. Now the agents handle the research, preparation and routing. Franchisees bring local knowledge and own the customer relationship.

Results came fast. From roughly 40 emails, Michael had two or three hot leads. One prospect replied in about 90 minutes asking for an estimate. Cold outbound reply rates are now over 3%, and the agents generated just over $150K in quotes in two weeks.

“It’s just a mock-up, but it just catches their eye immediately,” Michael says. He describes the owner’s reaction on seeing their own lot: “Okay, hold on a second. Let me, let me look into this a little bit.”

This is the obsessive-salesperson question in practice. Your best rep, given a week, might mock up a before-and-after for one big account. The agent does it for every property in every ZIP code.

No sales-tech vendor would have built this. It took someone who knows what a property manager notices when they pull into their own lot.

From a sales tactic to a franchise pitch

Michael started with one territory. He has now rolled the system out across every franchisee and ZIP code, and Appell has grown from 6 to 25 territories this year.

The agent has become part of how Appell recruits franchisees. A prospective franchise owner can see the pipeline they will inherit on day one: agents already researching their ZIP codes and sending outreach from their inbox.

National accounts used to be Appell’s main differentiator. Over the past six to ten months, Michael says, the agent-built tools have moved to the front of the pitch. He gives candidates a preview on weekly calls and a deeper demonstration when they visit the office. Then he asks them to compare the support other franchisors offer: “Is it just a couple of PDFs, or is it stuff like this?”

Across the national division, compared with last year, Appell has:

  • 30% more work orders
  • 39% more quotes submitted
  • 48% more approved jobs
  • 56% more value from approved jobs
  • 70% more gross profit on approved work

Monitoring: watch the inbox 24/7

Agents provide thoughtful replies to every lead within minutes.

Singular Innovation, an agency that builds agents for businesses, set this up for a language services company that coordinates interpreters for legal and insurance cases. The work is high-volume and deadline-driven.

Requests used to arrive by email, phone and SMS and wait for a full-time coordinator to read them. Now one intake agent handles every request as it lands, day or night. It turns each one into a structured record, flags anything incomplete instead of guessing, and processes the routine requests on its own.

Priority cases, scheduling conflicts and unusual terms go to a person. The rules are explicit: the agent never confirms an appointment to an interpreter without staff approval, never changes agreed rates or terms, and never decides a priority case on its own.

The overnight queue is gone. Only the exceptions reach a human. The coordinator has moved from reading every request to making the calls that matter.

Get started with one account

Targeting finds the right account and the right moment. Outbound turns that research into a visual pitch. Monitoring picks up the conversation when someone replies. Prospector connects the three.

Install the template, then give it a first brief:

Our business is website. We sell to customer type in market. Find one prospect with a reason to buy now. Show me the evidence, create a visual pitch and draft the email. Bring it back for review before sending.

Get that first account right, then set the rules for sending, handling replies and when to bring you in.

Hire Prospector