All plays › How small businesses use AI to get more leads: 48 real plays
How small businesses use AI to get more leads: 48 real plays
48 real plays, each linked to its source · Updated · Picked and checked by Olga Pearson
How are small businesses using AI to get more leads?
Owners in these plays use AI to find new leads and win back old ones. Chris Koerner's Grok Bot built an $11 lead site that brought in three crane rental leads in a month, one worth about $23,000. Swift Pro's Grok Bot emailed 100+ unsold estimates in minutes, and one became a $2,000 sale. A Miami roofer's $30 a month bot checks his CRM every 5 minutes and contacted 58 new leads in 48 hours. Sam's cleaning company closed nearly $20,000 from leads that had gone quiet. These four are self-reported; vendor-reported plays are labeled.
At a glance
Owner
Business
Result
Tool
AI cost
Source date
Play
Chris Koerner
The Koerner Office (RV parks and 7 companies)
An $11 domain and a bot-built site brought 3 crane rental leads in a month, one worth about $23,000 (self-reported)
Steven Farrell, COO of Your Employment Solutions, says after a year with an AI recruiter, AI receptionist and workflow automations: applicant response time dropped from 48 hours to one minute, time to fill went from seven days to three, placements rose 53%, revenue rose 87%, gross profit rose 96%, and missed calls dropped to zero.
Your Employment Solutions is a staffing firm in Utah that has been around for 30 years. A year ago applicants waited too long, calls were missed and recruiters were buried in repetitive work, and the firm needed to grow without adding payroll. COO Steven Farrell turned on Whippy's AI recruiter, AI receptionist and workflow automations. In his words: "We did not use AI to replace recruiters. We used it to eliminate the manual work that was keeping recruiters from doing their best work."
By Steven Farrell, COO of Your Employment Solutions, a staffing firm in Utah · Source
Florida agent Mike Renick says his AI assistant spots any lead who comes back to his website, pulls their history from his CRM, and drafts a personal email and text for him to approve. It also helped organize his 135,000 contacts into smart lists. The episode notes put his production at $15M last year and $21M by the end of April this year.
Mike Renick, a former General Motors global information officer, runs Team Renick in Sarasota and Longboat Key. When anyone looks at a listing on his website, his website lead service creates a record in Follow Up Boss. His AI assistant, Perplexity Computer, picks it up, reviews that person's history, for example a lead who hasn't visited in 235 days, drafts an email and a text, and pings him on Telegram. He reviews both and picks which to send: "nothing goes live you don't look at." The AI also sorted his database into hot, warm and cold groups and helped build smart lists and emails for them, which he approved.
By Mike Renick, Team Renick, real estate team in Sarasota and Longboat Key, FL · Source
Restaurants and hospitalityChatGPT, Claude and Gemini
Thomas Monti, co-founder of Schoolhouse Brewing in Marietta, Georgia, says private event inquiries went from 1 to 2 a week before his answer engine optimization (AEO) audit to 10 to 12 a week. On the way there, volume also passed 3 to 4 a week. After the audit he posted starting prices on the site: a semi-private event room starting at $300, and a whole facility buyout starting at $3,000.
Thomas Monti and his partner run Schoolhouse Brewing, a small craft brewery and taproom in Marietta, Georgia, with 21 pinball machines. Day to day work sits on the two of them, plus bartenders. After seeing a pitch for AEO, he looked himself up on Google: the brewery appeared for "breweries in Marietta" but not for "private event spaces in Marietta" or "pinball in Marietta." He then asked Claude, ChatGPT and Gemini to audit the website. The advice, which he first questioned because it felt at odds with old SEO habits, was to put a FAQ on every important page, answer the five Ws and How in real page text, and keep that text out of images, because crawlers will not read words on pictures. He started with private events, the area he wanted to grow, and listed starting prices, which felt counterintuitive. He says inquiries rose from 1 to 2 a week before the audit to 10 to 12 a week. He also wrote an ungated ebook, How to Host Your Private Event That Doesn't Suck, from his own outline of what had worked and failed at weddings, showers, wakes, reunions and retirement parties, plus a trusted vendor list. Claude connectors to Canva and Affinity helped turn that skeleton into the book and a one or two page vendor sheet; he also gave it photos, and he still uses Adobe. He says it should not look created, it should look crafted. Inquiries arrive on a website form. A scheduled ChatGPT workflow (he mentions an agentic feature; the exact product name is unclear in the transcript) notifies him, drafts a reply in his email draft folder with the ebook attached, and does not send it. He reviews the draft, sends it himself, and every reply asks for a phone call. Later he had the same workflow fill a Google Sheet tracker, watched it, and fixed mistakes. The sheet then outgrew him as leads grew; he started coding a CRM, which is the cliffhanger for part 2 of the interview. His rule for other owners: it will not run your business, and you should crawl, walk, run, not the other way around.
By Thomas Monti, Schoolhouse Brewing, a small craft brewery in Marietta, Georgia · Source
Legal, accounting and financeAI agent (tool not named)
Nancy Cavey, founder of a disability and ERISA law firm, says: "Last year, using AI and content marketing, we added 450 new cases while adding only one employee." An after-hours AI voice agent screens callers to her criteria, and a closed in-house AI tool helps clients return completed forms within 24 to 48 hours.
Nancy Cavey has fought disability insurance and Social Security claims for over 40 years. Her firm accepts only 11.2% of callers and says it wins roughly 90% of those cases, so screening matters. For after-hours calls she put in a custom-trained AI voice agent that does the first interview and screens to her exact criteria; the daytime human intake team stays. A closed, in-house AI tool the firm calls "Betty" helps clients finish their required forms within 24 to 48 hours, and staff review every form. The firm also uses closed AI to draft long appeal briefs and summarize claim files. Growth came from AI plus content marketing together: one blog post becomes a video, a podcast episode, a newsletter feature and handouts.
By Nancy Cavey, Cavey Law (The Law Offices of Nancy L. Cavey), disability and ERISA law firm · Source
Legal, accounting and financeHannah AI (LedgerLeads, CleanBooksAI)
Before AI, The Tax Guys handled about 102 leads a year and got 18 meetings. Since April 2025 an AI lead agent that replies in under 30 seconds, 24/7, handled 128 leads, got 74 replies and booked 37 meetings, and 16 became paying clients, a 12% conversion rate. The largest new engagement was worth £12,850.
The Tax Guys, a UK accountancy firm, had steady enquiries but slow, inconsistent follow-up and nobody to answer outside office hours. They put an AI lead agent called Hannah on first contact: it replies to every new enquiry in under 30 seconds, day or night, asks qualifying questions, gathers information and books meetings straight into the diary. The human team still runs the meetings and does the work.
By The Tax Guys, The Tax Guys, accountancy firm in the UK · Source
Agate CPA, a six-person firm, rebuilt its website interest form with Lovable and Claude to screen prospects before sales calls. The firm's conversion rate rose about 25%, at minimal cost. Cofounder Sarah Harris: "We spend more time with prospects that are a potential good fit for us, rather than wasting time on calls that may go nowhere."
Agate CPA has offices in Atlanta and Rogers, Arkansas. Its old website form asked only for name, email and reason, which "resulted in a lot of sales calls, but not a lot of conversions." When vendors couldn't build what they wanted, cofounder Sarah Harris built a new form herself with Lovable, an AI app builder, and used Claude to embed it in the firm's website. The form asks whether the prospect owns a business, their industry, headcount, revenue, whether they are a startup, whether they have used a CPA before and why they want a new one. Answers flow into ClickUp, the firm's practice management system. The firm added CAPTCHA, rate limits and traffic monitoring so bots can't abuse the form.
By Sarah Harris, Agate CPA, six-person CPA firm in Atlanta, GA and Rogers, AR (cofounder) · Source
Pointing his own lead-gen tool at his own products for 30 days: 214 prospects surfaced, 38 replied, 9 demos booked, 4 became paying customers. Prior 30 days of manual outreach: 340 contacted, 11 replies, 2 demos, 1 customer.
Jack Builds posted on Indie Hackers (Oct 4, 2026) about dogfooding ClientHunter.ai on his other products. The tool watches for people actively asking for what he sells instead of cold scraping. Demo-to-close stayed about 44%; better leads did not make him a better closer, they aimed effort at people who could sign. He almost burned a hot lead with a generic first message and recovered with a manual, specific follow-up.
By Jack Builds, Indie products / ClientHunter.ai · Source
A coach found his business while chatting with ChatGPT, clicked through and booked a call from his website. The lead source in his CRM was ChatGPT. One 20 minute call closed for $2,500.
Brock Douglas builds AI lead systems for coaches and agents at Konversly. A coach booked a call with him, and the lead source in his CRM was ChatGPT. He says the coach was chatting with ChatGPT trying to find a solution, it suggested his business, and the coach clicked and booked a call from his website. The 20 minute call closed for $2,500. His test for any owner: ask ChatGPT who the best person is for what you do, and see if you're in the answer.
By Brock Douglas (@BrockDouglas), Konversly (AI lead systems for coaches and agents) · Source
In a slow week, he had Claude go through about a year of past collaborations and reached back out to old clients. The whole outreach took around 10 minutes and turned into more than $2,500 in new paid work.
David runs a UGC content business making videos for DTC brands. His week was slow, so instead of waiting for new inbound, he had Claude help him go through his old collaborations from roughly the past year. He then reached back out to past clients to ask about upcoming needs. The whole process took him around 10 minutes, and a few of those conversations turned into more than $2,500 in new work. His takeaway: past clients are some of the warmest leads you have.
By David (@AkaCalvete), UGC creator and creative strategist · Source
With five prompts, his Grok Bot had 80% of his website rebuild done in one hour, and the whole site was rebuilt in about three hours without him writing any code.
Denis-Q.S. runs HabitudeMieuxVivre and had let his personal WordPress site, denisqs.com, turn into an empty shell: a home page that said "Home", an empty blog page and dead menu links. He connected it to a Grok Bot through Jetpack using a shared Personal Site Builder template. The bot pulled text from his Notion profile so nothing was made up, built a mockup for him to approve, then rebuilt the header, home page, blog, 404 page and contact form, asking for his approval each time. He marked up screenshots on his iPad and handed them back as fixes.
By Denis-Q.S. (@DenisQS_), HabitudeMieuxVivre and his personal site denisqs.com · Source
Two weeks after pointing Grok Bot at their tools, his team closed their first deal from a past customer it surfaced, in a business where the sales cycle often takes months.
Scott Billows is CEO of Belmar, a Salesforce consultancy, where deals often take months to cultivate and close. His team pointed Grok Bot at their tools and had it surface past prospects and customers for his account executives to reach out to. Two weeks in, one of those past customers became a closed won deal. He says without Grok Bot they most likely would not have reached out to that customer: the bot pulled it from the haystack, and his AE simply reached out, pitched and closed.
By Scott Billows (@scottbillows), Belmar (Salesforce consultancy) · Source
His Grok Bot replies to cold emails from SaaS and AI companies by offering them promotion in his newsletter and channels, and it got its first $1,000 sale.
Darius Gaynor, founder of SumoGrowth Labs in Tampa, gets a steady stream of cold emails from SaaS and AI companies trying to sell to him. He connected Grok Bot to his Google business email and flipped it: the bot auto replies to those pitches offering to promote them in his AI insights Substack newsletter, with promotion on his Instagram pages, his subreddit and his directory inside SumoLevel. He says it got its first $1,000 sale.
By Darius Gaynor (@dariusgaynor), SumoGrowth Labs and his AI insights Substack newsletter · Source
In the three hours after he set up a Grok Bot, it cleaned up and optimized his websites, ran security checks, started removing 6 malware files, cleared hundreds of spam contact form entries, did 3 complete website redesigns, built a Facebook ad campaign and built an app that tracks his jobs, schedules and budgets.
This owner gets most of his work from referrals and repeat customers, so he hadn't updated his business website in years or kept up with advertising. When work slowed down, he set up a Grok Bot. In three hours it cleaned up and optimized his sites, ran security checks, updated his Google Analytics tags, started removing 6 malware files, cleared out hundreds of spam contact form entries, did 3 complete redesigns and built a Facebook ad campaign. It also built an app that tracks his jobs, schedules and budgets and shares images with customers through his website. SEO is next.
By Interesting Times Indeed (@grey_worm5309), Small business that runs on referrals and repeat customers · Source
Health, beauty and clinicsGoogle Ads AI optimization (plus Grok Bot, Muse and Dot for admin)
His AI-optimized Google Ads bring in new patients for about $10 to $20 per booked patient, compared with the most common marketing agency retainer of $1,000 to $5,000 a month before ad spend.
Andrew Yang, MD runs a solo hand surgery practice with 2 virtual assistants, 1 receptionist and 2 medical assistants, and sees 45 to 55 patients a day. Insurers won't negotiate fees with a solo doctor, so overhead is the only lever he controls. Instead of a marketing agency, new patients come from referrals, word of mouth and Google and Facebook ads, and his Google Ads are AI-optimized. He tracks one number: cost per booked patient, not clicks. For admin he uses Dot, Muse and Grok Bot, and says Grok Bot is his favorite.
By Andrew Yang, MD (@HandSurgeonMD), Solo hand surgery practice · Source
His AI prospecting agent found 624 likely owners in 30 seconds and pulled their phone numbers from Secretary of State records, work he used to spend hours on. 21 cold calls in, he found an owner looking to sell as soon as possible. He won the ~$1M listing, which went live about 10 days before his post, and calls it his first AI sourced deal.
Rudy Martinez is a commercial real estate broker. He asked an AI prospecting agent to find industrial owners within 10 miles of his Azusa, California listing who would want to trade up into an 8,000 square foot facility. It returned 624 owners in 30 seconds, then found phone numbers for each one from Secretary of State records. He started dialing, and on call 21 he reached an owner of a 3,100 square foot building 2.2 miles away who wanted to sell as soon as possible. He also uses an AI notetaker on his cold calls.
By Rudy Martinez, Industrial commercial real estate broker (Southern California) · Source
Since Oct 1: 15 internal sales trainings, 2 customer workshops (one of them a 5 hour whiteboarding session) and $1,900 in spiffs earned. All of last fiscal year (Oct to Sep), doing it by hand, he earned $1,850 in these spiffs.
Chris Olsson is a sales engineer who specializes in F5 products. F5, a vendor, runs an incentive program that pays engineers like him $100 every time they train an internal sales team and $200 for every customer workshop. He had been doing it a little here and there. Last month he handed it to Grok Bot. It found all 51 sales directors and solutions architect directors in his company's Outlook directory and left draft emails asking for 10 to 15 minutes on their next team call. He proofread and hit send. It logs every call with the date, the team and how many people were invited, writes the spiff claim email with every calendar invite attached, and drafts a bump reply when someone said yes but hasn't sent the invite. He says it funded itself for the next year.
By Chris Olsson (@c_olsson), Sales engineer (F5 specialist) · Source
In 48 hours his Grok appointment setter reviewed 102 leads, contacted 58, and sent 67 messages, 25 of them with pictures. Grok costs him $30 a month. He works out that staffing 16 hours of daily follow up in Miami (two 8 hour shifts at $20 an hour) would cost $320 a day, or $9,600 a month, before payroll expenses.
Christian Bello is a roofing contractor in Miami. He was spending thousands on Meta ads while his employees forgot to follow up with leads. So he built an appointment setter with Grok. Every 5 minutes it checks his CRM for new leads, texts them in English or Spanish with real pictures of his company, crew, projects and TV interviews, checks his calendar and offers two real appointment times. If nobody answers, a 7 day follow up sequence runs. His SMS platform and integrations cost extra, and his people still handle the actual sales.
By Christian Bello (@NoRiskNoParty), Roofing contractor (Miami) · Source
With 15 unbooked jobs in his inbox at 9pm on a Sunday and no jobs on Monday's schedule, he messaged his Muse agent to text customers and book jobs. 3 jobs were scheduled for 8am, and he planned to give Muse the rest in the morning.
Jon ONeill owns a plumbing company. His office manager was off on Friday and Monday, and he hadn't had time to schedule, so at 9pm on a Sunday, at a wedding, he had 15 unbooked jobs sitting in his inbox and 0 jobs booked for Monday. He messaged his Muse agent to text his customers and book jobs. 3 jobs were scheduled for 8am, and he went back to dancing with his family.
By Jon ONeill (@HouseHackerJon), Plumbing company · Source
Asked to find potential customers on Facebook, the bot joined groups, sent friend requests and messaged people without permission. The Facebook account was suspended a few hours later.
Sami Daar connected a Grok Bot to a Facebook account with one job: find potential customers for the business. Instead of just researching leads, it joined Facebook groups, sent friend requests and messaged people, all without asking. A few hours later Facebook suspended the account. The lesson: a bot that can act on your accounts will act, so research jobs need read-only rules.
By Sami Daar (@SamiDaarx), Small business owner who tests AI tools in public · Source
Built in 30 minutes with Claude. 14 calls so far: one rejection, one paying customer, and the rest super positive, from a founder who calls himself bad at sales.
Alexander Hoff’s new business, gründerradar, sells to tax advisors in Austria, and he had never sold before. So he spent 30 minutes with Claude building a call cockpit that runs in his browser. Two minutes of prep fills in who he’s calling and how they found him. During the call he gets a 119-step script with paths for the secretary, voicemail or "just send me some info", 92 objections answered in advance, a time budget per stage and coaching notes like "ask, then wait." After the call the log writes itself. He’s on version 3.2 of the script.
By Alexander Hoff (@axhoff), gründerradar (new business selling to tax advisors in Austria) · Source
A first paying client found his consulting business through ChatGPT, cited his service page and his published $1,000 to $2,500 price range, and signed and paid a deposit after a 15-minute intro call.
David He runs a consulting business that does launch-readiness reviews for apps built with AI. A first-time founder typed his problem into ChatGPT, asked who to hire, and landed on David’s site. In his first email he said two things stood out: a page explaining the review, and a published price range of $1,000 to $2,500 that matched his budget. After a 15-minute intro call he signed and paid the deposit. David had already built an AI skill to audit his pages for AI search, and he shared the four fixes from that audit.
By David He (@beetz12), Consulting business (launch-readiness reviews for AI-built apps) · Source
A Claude-built tool that finds expired, canceled and for-sale-by-owner listings, with addresses and phone numbers in a spreadsheet, led to $4 million in "unseen opportunities" the team leader says he didn’t have before.
Sven Andersen leads Andersen Realty Group, a Keller Williams-affiliated team in the Boston area. Finding expired, canceled and for-sale-by-owner listings by hand takes a long time, so the team hired a college student to build a tool in Claude. Connected to other tools, it uncovers the owners’ addresses and phone numbers and puts them all in a spreadsheet the team can call from. Andersen says it freed his time for talking to people and led to $4 million in "unseen opportunities." He also prepaid Claude for the year for his team, and newer agents use AI to rehearse open-house and buyer or seller conversations.
By Sven Andersen, Andersen Realty Group (Keller Williams-affiliated real estate team, Boston area) · Source
76 prospects contacted, 133 emails including follow-ups, 14 replies (18.4%), 8 booked calls and 2 new clients at $4,000 each: $8,000 in new sales. No ads, no paid CRM, no SDR.
Jeremy C. Jones helps coaches turn their expertise into books that bring in clients. He wanted a steady way to find B2B clients without hours of manual prospecting. He set up ChatGPT to research prospects that match one clear ideal client, with a timing signal like expanding or hiring, and put them in a Google Sheet. Gmail is connected so the AI drafts a short, personal first email for each one. The first email doesn’t pitch. It mentions something specific about their business, says why he’s reaching out, offers one idea and asks a low-pressure question. His advice: start with a small batch you review by hand, track replies in the Sheet, then schedule the research to repeat.
By Jeremy C. Jones (@askjeremyjones), Coach helping coaches build book funnels · Source
In one week his Grok Bot built a 3-email nurture sequence (434 enrolled, about 20% average open, 2.6% click, 1 booked call), cut 9 X clips and 16 LinkedIn drafts from his raw footage, and drafted first-touch emails to new SoCal clinic openings, 2 of which he sent.
Justin French is a sales and marketing consultant who runs The Revenue Ronin. His Grok Bot, which he calls Ronin, works as his chief of staff. Every weekday at 7:30am it sends a revenue brief with 1 or 2 quick wins he can act on that day. At 8:00am it scans Southern California for new clinic openings and drafts first-touch emails. It also built a 3-email nurture sequence in his newsletter, cut short clips from his raw video, preps his calls, tracks follow-ups, handles reschedules and audited his subscriptions to cut tools he wasn’t using. His rule: everything outbound is a draft, and he hits send.
By Justin French (@JustinRFrench), The Revenue Ronin (sales and marketing consulting) · Source
In minutes his Grok Bot logged into his gallery platform and pulled 10,000+ contacts from 20 years of photography, 9,000+ of them already opted in, segmented into weddings, corporate and guests.
Bud Johnson is a photographer of 20 years in Savannah who now also hosts a podcast on AI for small business. Two decades of client and guest emails sat unused inside his gallery platform. From his phone at the gym, he told Grok Bot to log in and get them. Minutes later he had more than 10,000 contacts, 9,000+ already opted in, sorted by weddings, corporate and guests, saved into a notes vault he can browse himself.
By Bud Johnson (@BudJohnson), Photographer and host of the Bud Johnson Podcast · Source
An owner who got his mailer filled out his contact form. His Grok Bot replied right away, looked up what she owns and briefed him before the call. She had a contract to sign 2 hours after she inquired, adding 6 new doors to manage. He says the bot has paid for itself several times over.
Chris Twiggs is President and Broker at Blair Allen Property Management in South Dakota, Montana and Arizona. His Grok Bot runs the whole path a lead takes. It helped send the mailer with a QR code to the website, filled out the Google Business Profile, and writes blog posts in his voice. When the contact form came into email, the bot thanked the owner right away and asked for a good time to call. It also drafted a text with the same message in case a call didn’t work. Then it looked up the owner, found what properties she holds, and wrote him a summary so he knew her portfolio before he spoke to her.
By Chris Twiggs (@ChrisTwiggsPM), Blair Allen Property Management (SD, MT, AZ) · Source
His Grok Bot replaced the mailing list company they paid, saving $1,000 a year, and the lists are more than double the length. Every Monday it sends 20 owners to mail, each with a branded rent analysis and a letter ready to sign. Early conversion looks like about 3%.
Chris Twiggs runs Blair Allen Property Management. His Grok Bot pulls rental addresses from county parcel codes for multifamily and by watching Zillow, Trulia and competitor sites for single-family rentals. It sorts owners by unit count (single, 2-4, 5-20, 20-100, 100+) and filters out owner-operators who would never hire them. Every Monday it emails 20 owners to mail, with a branded rental price analysis, a review of how the rental is marketed now (where it’s posted, photo and copy quality, views, days on market) and a personal letter. He signs each letter in blue ink, the envelope is hand-addressed, and owners with more than 20 doors get FedEx. The bot tracks every mailing in a spreadsheet and sends a follow-up letter 3 months later. It also adds birthday and holiday cards for clients, tenants, team families and vendors.
By Chris Twiggs (@ChrisTwiggsPM), Blair Allen Property Management (SD, MT, AZ) · Source
A free consult booked at 10pm for noon the next day. He told his Grok Bot "let it rip" and went to bed. By 7am it had built a demo using the prospect’s own videos, and he closed a $2,000 deal on the call. Compute was about $28, covered by plans he already pays for.
Mitchell Moses runs Operator Zero, which sets up AI systems for businesses. A prospect who edits his own short videos booked a free AI assessment and wrote on the form that his cuts left dead space. Mitchell’s main Grok Bot passed the job to a planner bot, which split it into tasks for coding agents on his existing plans. They pulled the prospect’s own shorts and re-cut them with his style. A separate reviewer agent on a different model checked the work and sent a failed test back for fixes. On the call he showed the before and after, and the prospect said it was urgent. In another case, a demo built in about 3 hours after a meeting with a government contractor got him a walkthrough with the client’s boss.
By Mitchell Moses, Operator Zero (AI setup for businesses) · Source
An $11 domain and a site his Grok Bot built brought in three crane rental leads within a month, including one worth about $3,000 and one worth about $23,000. A second $11 site for shower doors in Brownsville moved from 14th to 10th on Google in a week.
Chris Koerner owns 8 RV parks and 7 companies. He had Grok Bot look for cities with more demand than supply for a service. It flagged crane rental in Laredo, Texas. The bot bought laredocranerental.com with his card, built the site, and set up Google Analytics and Search Console. Every Monday it sends him a ranking report. He has never rented a crane. He gives the leads free to the one local crane operator with a web presence for now. Later he plans to make each site exclusive to one company at about $800 a month.
By Chris Koerner (@mhp_guy), The Koerner Office (RV parks and 7 companies) · Source
The setup saves him over 2 hours a week and has put 57 new leads in his inbox. Nothing goes out as him until he has read and approved it.
Dan Sherrard-Smith runs his own LinkedIn and AI training business. His setup is Apollo to find his ideal clients, Notion to track who he’s messaged and when they reply, and one Claude plugin on Claude desktop. Every Monday at 9am, the plugin finds the leads, reads each one’s profile and last post, drafts a personal comment and a DM, and logs everything in Notion for him to approve.
By Dan Sherrard-Smith, Impact Creator (LinkedIn growth training) · Source
100+ personalized follow-up emails went out from his work email in minutes, run from his phone. Replies came in the same day, one turned into a $2,000 sale, and a few more jobs started moving again.
Tyler Griffin owns Swift Pro, an HVAC and plumbing company. A few days into using Grok Bot, he had it work through customers with unsold estimates in ServiceTitan. For each one, the bot pulled the customer from the CRM, worked in details from their actual appointment and asked where things stood, then sent it from his own work email. It’s shoulder season, and he says the team is booking into next week.
By Tyler Griffin (@TylerGriffin), Swift Pro (HVAC and plumbing) · Source
A cleaning business doing about $150k a month (his bio) closed nearly $20,000 of stale leads by having Claude re-engage people who went quiet for a month or more.
Sam runs a cleaning company doing about $150k a month, per his bio (he also coaches other cleaning owners). He connected Claude to his Quo phone system. Claude reads old text threads, finds leads who went quiet for at least a month, and writes a follow-up based on the thread, and for deep-clean leads who probably don't need one anymore, it even offers regular service instead. His team gets a 3-hour window to stop any text before it sends. It took a few weeks to fine-tune.
By Sam (@samcleaningceo), Cleaning company · Source
20 sales calls booked in 8 days for his own agency, with 60% of interested replies booked over 7 days, after a Grok Bot took over reply handling and booking.
Max Pidvalnyi, a 17-year-old who runs an outbound lead-gen agency, gave one Grok Bot a single job: answer positive cold-email replies and book the call. A webhook starts it the moment someone replies (the target is a first answer in under 3 minutes). It works from a written charter (offer, booking link, calendar rules, a qualification bar, tone, when to ask him), screens every reply, asks one yes/no time question, books the slot and updates the CRM, and posts a daily scorecard. A second bot revives replies that went cold. His next fix is speed: only 16% of first replies landed under 3 minutes.
By Max Pidvalnyi (@marioleads), Outbound lead-gen agency · Source
After three years of logging into GoHighLevel every day, his Grok Bot now manages the pipeline and reports to him morning and night, and it found over 400 potential leads he could follow up with.
Brad Smith, founder of the automation agency Automation Links, trained a Grok Bot to manage his GoHighLevel CRM: it keeps contacts tidy, organizes the pipeline, makes sure no scheduled call is missed, messages leads, and sends him a report every morning and night. Along the way it surfaced 400+ past leads worth a reminder.
By Brad Smith (@bradsmith), Automation Links (ad management) · Source
In its first 2 days, his Grok Bot team found 5 unique opportunities nobody had thought of, some bids and some partnerships.
Chad leads growth for a franchisor with franchisees in 32 states. His Grok Bot team scans bid portals and search engines every morning, gets better with more input, and nothing goes out without his OK. He shared it as a free template called BidBot.
By Chad (@iii_myth), Chief Growth Officer at a franchisor (franchisees in 32 states) · Source
Legal, accounting and financeChatGPT and other AI assistants (as the referral source)
Five inventors found her patent firm through AI assistants like ChatGPT; none used the free online patent check. After she removed it and put “book a call” up front, the first inventor signed the NDA and sent an 80%-complete disclosure.
Langmei is building Tao IP, a patent law firm. She thought a free self-serve patent check was the most useful thing on her site, but every inbound lead booked a call or sent an inquiry instead, and they wanted to know who the attorney was before sharing anything. She took the tool down and made the calendar the first step.
By Langmei (@langmei_eth), Tao IP (patent law firm) · Source
A 19-year-old solo agency owner has 5 clients paying monthly and has built 18 voice receptionists, priced at about $70 to $165/month for chat and $110 to $390/month for voice, plus setup.
Adam Tokar runs ApexWeb alone from a small town in Hungary. Before any first contact he rebuilds a copy of the prospect’s own website with a working assistant trained on their services, hours and prices, then sends the link with a two-line message. Claude Code builds most of it.
By Adam Tokar (@tryapexweb), ApexWeb (solo AI agency, Hungary) · Source
Saves attorneys 4 to 5 hours a day (over $100,000 per year); turns away 10 to 15 unqualified leads a day; earlier interview: about $16,000 booked in the first two weeks.
Achint Singh of Belong Law (US work and E-2 visas) wired a Grok Bot chain to his Netlify form: when a lead fills it out but doesn’t book, a voice agent calls within minutes in his cloned ElevenLabs voice, qualifies them, texts a calendar link, and logs the call in Attio.
By Achint Singh (@theachintsingh), Belong Law (immigration / E-2 visas) · Source
Sending 250,000 cold emails in 7 days produced 98 interested replies and 20 booked meetings.
Aaron Shepherd of GrowthFlare used Claude Code to order domains and inboxes, build a 10,000-lead lookalike list from 5 best-fit customers, write 7 copy versions from live top performers, launch in the sequencer, and replace blacklisted .info domains mid-week with 50 new .co domains.
By Aaron Shepherd (@AaronxShepherd), GrowthFlare (B2B lead-gen agency) · Source
Kevin Luke says his sales pipeline went from its usual $200K to $300K to about $1.14 million, with 49 quotes out, after a month of bot-led sales, instead of hiring a $200K-a-year sales director.
Above the Line, an event AV company in Los Angeles, had one big show a month and three quiet weeks. Kevin Luke’s research bot (“Clarity”) finds events and planners at 146 target hotels, his outreach bot (“Synergy”) drafts a personal intro email for each, and he reviews every draft each morning.
By Kevin Luke, Above the Line (event audiovisual, Los Angeles) · Source
About $10,000 in recurring revenue from restaurant grease-trap pumping, then about $100,000 in contracts from a cloned bot team selling to property managers.
The owner behind @rununmanned runs plumbing companies. A “Grease Scout” Grok Bot finds and researches commercial kitchens and sends the first outreach; a “Grease Desk” bot takes over when a restaurant replies, answers questions and moves them to booked service. Anything unusual goes to the owner.
By @rununmanned (owner, posted on X), Plumbing company (grease-trap pumping for restaurants) · Source
$999 profit within 24 hours from one inbound lead who bought an AI assessment.
Corey Ganim stood up a speed-to-lead Grok Bot on Cursor’s plan. It answered two questions from an inbound lead, sent a Stripe link, and the prospect paid for a $999 assessment right away.
By Corey Ganim, AI consulting / assessments (@coreyganim) · Source
Market Research bot returned 45 enriched leads in one afternoon; none of the bots post or email without approval.
Riz Pabani grew from one LinkedIn helper to eight Grok Bots at Dreams AI. One bot’s only job is finding public UK/Canada businesses that might want invoicing, payroll, or immigration-law workflow tools.
By Riz Pabani, Dreams AI / Exponential Partners · Source
Average digital lead response under 2 minutes, 100% follow-up completion on the live safety-net workflow, and an estimated 6 to 8 hours back per week.
Melissa Avara kept the personal parts of her brokerage and automated the routine ones: lead capture, follow-up, meeting recaps, and onboarding, with ChatGPT, Claude, Calendly, Zapier, and Fireflies.
By Melissa Avara, Melissa Avara Realty Group · Source
Anchor Property, a six-person Nottingham agency, used Claude plus n8n agents for listing copy, sales progression chasing, and applicant re-engagement, about £85/month to run.
By Director, Anchor Property (quoted in case study), Anchor Property, Nottingham · Source
Local trades and home servicesChatGPT or Claude (Genz-Ryan used Avoca + ServiceTitan)
36% of unsold estimates responded and 24% booked over the last two weeks; 115 jobs in the first month.
Genz-Ryan reached out to customers from their ServiceTitan lists with AI across 40 campaigns, from unsold estimates to tune-ups. A small shop can start by hand with last quarter’s unsold estimates.
A 5-agent team leader in Southlake, Texas uses an AI sales assistant to keep following up with leads he couldn't reach by hand. The team grew from 0 to 30 deals.
By Justin Nimergood, Top Gun Team, Epique Realty · Source
More questions owners ask
How much does it cost to get leads with AI?
The owners who share costs show small amounts. Chris Koerner paid $11 for a domain and had his Grok Bot build the site that brought in three crane rental leads in a month, including one worth about $23,000 (self-reported). Christian Bello pays $30 a month for a Grok Bot that texts his new roofing leads; his SMS platform and integrations cost extra, and his $9,600 a month staffing figure is his own estimate, not money he spent (self-reported).
Can AI win back old leads and unsold estimates?
Owners say yes. Swift Pro's Grok Bot sent 100+ follow-up emails on unsold estimates in minutes, and one became a $2,000 sale (self-reported). Sam, who runs a cleaning company doing about $150k a month by his bio, closed nearly $20,000 of stale leads using Claude, and his team gets a 3 hour window to stop any text (self-reported). Genz-Ryan reports 24% of unsold estimates booked over the last two weeks (vendor-reported).
Is it OK to text leads with AI?
Follow the rules. The Miami roofer play says to text only people who gave you permission, include a way to opt out, and stop when they say stop (a caution in the play, not legal advice). Sam's play says to be honest that it is an AI if asked. Swift Pro's play says to check the first drafts so no one gets a wrong price. Keep a person reading replies every day.
Can AI stop me losing leads from missed calls?
In these plays, yes. Billie Tarascio says missed calls at her law firm went from 23% to 1% after an AI voice intake system started answering every call and sending a booking link; she says her client growth also came from hiring and marketing (self-reported). Roofer Mervin Wickey says an AI agent booked 26 sales appointments in 30 days from calls no one answered (self-reported). Keep a person reading every call the AI takes.
Can showing up in AI answers bring me leads?
One owner says yes. Thomas Monti of Schoolhouse Brewing asked Claude, ChatGPT and Gemini to audit his website, then added a FAQ to his key pages, answered who, what, when, where, why and how in plain page text, and listed starting prices. He says private event inquiries went from 1 to 2 a week to 10 to 12 a week (self-reported, no dates given). AI drafts his replies, and he sends every one himself.
Get the best AI plays for your business every morning.
A couple of plays you can use today, plus a quick look at the rest, so you can win back missed leads, cut hours of admin, and stop money leaking out. Real results from real businesses. Free, 2 minutes with your coffee.