A free Overyield tool. Everything you need to install a missed-lead recovery system for one local business, prove what it recovered, and get paid to run it.

You do not need code. You need a spreadsheet, a chat model (Claude, Gemini, or ChatGPT), and one business that already buys leads and wastes them. Work top to bottom. It takes a weekend the first time, and about an hour a week after that.

What you are building (the 60-second overview)

One simple system: pull a business’s old leads, missed calls, and dead web forms into a Sheet, score each one hot, warm, or cold with AI, draft the follow-up in the office’s voice, hand the front desk a daily call-back list, and track what it recovers. You are not selling a chatbot. You are selling recovered revenue, with the receipts.

Your whole operation lives in one Google Sheet, six tabs. Set these up empty now; the steps fill them:

  1. Raw Leads: their export, pasted in untouched.

  2. Scored Leads: the AI output: HOT / WARM / COLD, a reason, the next action.

  3. Message Queue: the drafted SMS and email per lead (status: draft, approved, sent).

  4. Call Log: HOT-lead calls and what happened.

  5. Booked: appointments that came back.

  6. Scorecard: the 30-day before/after that renews your retainer.

Step 1) The 5-minute lead audit (do this on the sales call)

Ask the owner four questions. Write down the answers. These become your pitch.

  1. “Where do new inquiries come from?” (calls, web forms, Google, Instagram, referrals)

  2. “When someone reaches out and you do not close them, what happens to their info?” (the answer is almost always “nothing”)

  3. “Roughly how many of those do you get a month, and what is a new customer worth to you?”

  4. “Who would run a daily follow-up list if I handed them one already written?” (usually the front desk)

Then do the math out loud, conservatively. Say they get 50 inquiries a month and close 20. That leaves 30 going cold. You do not need all 30. Recover even two a month at $1,500 each and that is $3,000 a month they are leaving on the table, every month. That is the number you sell back to them. Use their real close rate and customer value, and always round down. Believable sells better than impressive.

Step 2) The exact data to request

Ask for a CSV export from their CRM, booking tool, or inbox. You need, at minimum:

  • Name

  • Phone and/or email

  • Date of first contact

  • What they asked about (free text, notes, last message, messy is fine)

  • Status (booked / no-show / never replied / unknown)

If they have no CRM, ask for: missed-call logs from their phone system, web-form email notifications (a Gmail search export), and any “contact us” inbox. Tell them: “Just send me the mess. Cleaning it up is my job.” Paste it all into your Raw Leads tab.

Step 3) The lead-scoring prompt

Paste this into Claude, Gemini, or ChatGPT. Put the result in your Scored Leads tab.

You are a lead-recovery analyst for a [TYPE OF BUSINESS, e.g. chiropractic office]. Below is a CSV of old leads, missed calls, and web forms. For each lead, output a row with: name, contact, a score of HOT / WARM / COLD, a one-line reason, and the single best next action. Scoring rules: - HOT: asked about price, booking, availability, insurance, or described a specific problem/pain. Recent. - WARM: general question, downloaded something, replied once, or older but specific. - COLD: vague, spam, out of service area, wrong number, or already a customer. Sort HOT first. Be conservative: when unsure between WARM and COLD, choose WARM. Do not invent details that are not in the data. DATA: [paste the CSV]

You now have a prioritized money list. The HOT rows are your 90-second demo: this is what closes the next client.

Step 4) The follow-up message prompt

Once you have the scores, ask Claude, Gemini, or ChatGPT. Put the result in your Message Queue tab.

Write follow-up messages for these leads in the voice of [BUSINESS NAME], a [TYPE] in [CITY]. Tone: warm, human, low-pressure, never salesy. For each HOT and WARM lead, write one SMS (under 160 characters) and one short email. Reference what they originally asked about. End with one easy next step (a question or a booking link). Do not use the words “AI,” “automated,” or “follow-up sequence.” Sound like a real person at the front desk who remembered them. LEADS: [paste the HOT and WARM rows]

Sample output (chiropractor): “Hi Amanda, this is Dr. Patel’s office. You reached out a while back about neck pain after your accident. We had a couple of openings this week and wanted to see if you still want to get it looked at.”

Step 5) The front-desk call script

For HOT leads, a call beats a text. Hand the receptionist this:

“Hi [name], this is [name] from [business]. You’d reached out about [their issue] a little while back and I wanted to personally follow up. Are you still looking to get that taken care of? … Great, I have [day] at [time] or [day] at [time], which works better?”

If voicemail: “Hi [name], [name] from [business] following up on your question about [issue]. We’ve got a couple of openings this week, give us a call back at [number] and we’ll get you in.”

Step 6) The client offer (copy, paste, fill the brackets)

What I do: I take your old leads, missed calls, and web forms that never got followed up, and turn them into a daily list your front desk can actually work, with the messages already written. What you get: a prioritized call-back list every morning, ready-to-send texts and emails in your voice, and a 30-day report showing exactly what it recovered. Price: $3,500 to set it up, $1,500/month to run it. If we recover one [customer] a month, it has paid for itself. The deal: give me your last 90 days of leads. I’ll build the first list free so you can see it before you decide.

Step 7) The client-safe QC checklist (run before anything sends)

  • [ ] Every message reviewed by a human before sending. No exceptions on the first month.

  • [ ] No medical, legal, or financial claims in the copy.

  • [ ] Texts only to numbers that opted in or are existing/past customers. Include the business name. Honor STOP. (US: this is TCPA territory; when in doubt, email or call instead of cold-texting.)

  • [ ] Emails include the business’s real name and a way to unsubscribe. (US: CAN-SPAM.)

  • [ ] No invented details, no fake urgency, no “we tried to reach you” if you did not.

  • [ ] The owner has seen and approved the voice and the offer.

This checklist is the product. It is why they pay you and not a $20 chatbot.

Step 8) The 30-day before/after scorecard (this is your renewal)

Fill it on day 0 and day 30. It turns “trust me” into proof. But proof is only proof if the numbers are honest, so define every term before you start.

Attribution rules (use these exactly):

  • Follow-up sent: a message you actually sent, not one you drafted.

  • Conversation: the lead replied or answered a call. A voicemail does not count.

  • Appointment booked: a real slot on the calendar.

  • Showed: they actually attended. Booked is not money; showed is closer.

  • Revenue recovered: only a customer who booked, showed, and can be traced to a follow-up you sent. Tag every recovered customer in your Sheet so you can defend the number. When unsure, do not count it.

Track these, before → after:

  • Old leads sitting unworked: ____ → ____

  • Estimated value of that pile (conservative): $____ → ____

  • Follow-ups actually sent: 0 → ____

  • Conversations started: 0 → ____

  • Appointments booked: 0 → ____

  • Showed: 0 → ____

  • Revenue recovered (traceable only): $0 → ____

Screenshot the “before.” Email the “after.” That one honest image is your testimonial, your case study, and the reason your $1,500/month never gets questioned.

See it work: one chiropractor, start to finish

Here is the whole system on a made-up but realistic client, so you know exactly what you are delivering.

The client: a chiropractor with 312 old leads sitting in a booking tool nobody has touched in months. A new patient is worth about $1,500.

A slice of the Raw Leads export:

  • Amanda R. | 6 weeks ago | web form | “neck pain after a car accident” | never replied

  • Mike T. | 3 months ago | missed call | voicemail: “do you take my insurance?” | no follow-up

  • Priya N. | 8 days ago | web form | “lower back pain, can I come in this week?” | never replied

  • Janelle P. | 5 weeks ago | Instagram | “do you treat sciatica?” | replied once, went cold

  • Carlos M. | 4 months ago | web form | “prices for adjustments?” | never replied

  • Sarah K. | 7 months ago | referral | “scheduling for my son” | booked, then no-showed

  • Dan W. | 1 year ago | web form | “do you do massage?” | out of scope

  • (unknown) | 2 months ago | missed call | no voicemail | unknown

After the scoring prompt (Scored Leads tab):

  • Amanda R. → HOT: specific injury, recent, accident often means insurance and urgency → text + call today

  • Mike T. → HOT: insurance question is high intent → call today

  • Priya N. → HOT: recent, urgent, ready to book now → text + call now

  • Janelle P. → WARM: specific condition, engaged once → text

  • Carlos M. → WARM: price question but old → text

  • Sarah K. → WARM: was already booked, just dropped → call, easy rebook

  • Dan W. → COLD: service not offered → skip

  • (unknown) → COLD: no info to act on → skip

Across all 312 leads: about 28 HOT, about 90 WARM, the rest COLD. You work the 28 first.

A drafted message (Message Queue tab), Amanda, HOT:

SMS: “Hi Amanda, this is Dr. Patel’s office. You reached out a while back about the neck pain after your accident. We’ve had a couple of openings come up this week, want me to hold one for you?”

The front desk reviews it, approves it, and sends. Human in the loop, always.

The filled 30-day scorecard (conservative, traceable only):

  • Old leads sitting unworked: 312 → 0 (all triaged)

  • Estimated value of that pile (conservative): ~$15,000

  • Follow-ups actually sent: 0 → 138

  • Conversations started: 0 → 21

  • Appointments booked: 0 → 9

  • Showed: 0 → 6

  • Revenue recovered (traceable): $0 → $9,000 (6 patients who showed, at $1,500)

You charged $3,500 to set it up and $1,500/month to run it. Month one recovered $9,000 it can prove. That is the renewal conversation, and the testimonial that gets you client number two.

Numbers here are illustrative and deliberately conservative. Real results vary by niche, how old the list is, and how fast the front desk works it. Never promise a number. Show the method and let the scorecard speak.

Your 7-day install plan

You do not need a month. You need a week and a willing front desk.

  • Day 1: Run the 5-minute audit on the sales call. Get the verbal yes and the data.

  • Day 2: Pull the export. Build the six-tab Sheet. Paste Raw Leads.

  • Day 3: Run the scoring prompt. Fill Scored Leads. Send the owner the HOT count as your demo.

  • Day 4: Run the message prompt for the HOT leads. Owner reviews and approves the voice.

  • Day 5: Front desk sends the first batch and calls the hottest five.

  • Day 6: Log replies and calls. Book what books.

  • Day 7: Send the owner the week-one numbers and set the monthly rhythm (about an hour a week from here).

That is the install. Then you repeat it for the next business, and the one after that.

Want the done-for-you version?

The free kit is the runnable starter. The Pro version is the Missed-Lead Recovery Operator’s Edition, plus the full Vault beside it: the finished six-tab system, the tuned scoring and follow-up prompts, the client offer and share-of-savings contract templates, and the objection scripts for the sales call, each ready to run. Beyond this play, Pro is the whole library of complete plays like it, a new one banked every Monday, a members-only deeper play each week, the forward-only track-record scorecard, and the founding community.

Founding rate: $129 a year, locked forever. It closes when the first 25 members are in.

This kit is from Overyield: make money with AI, one play at a time. Educational, not legal or business advice. Follow the rules in your area for outreach (TCPA, CAN-SPAM), and keep a human in the loop. Numbers in examples are illustrative.

Keep reading