stpierre.ai · Internal · One Source of Truth

The Operating Playbook

v1.1 — plain-English edition · July 27, 2026 · Built from a full audit of every page, plan, and dashboard the business has (11 AI agents read roughly 90 documents and graded each one)

§0The 7 Rules

These came straight from Akash on July 27. Every decision in the business gets tested against them.

The four numbers — the only weekly targets

1,500outreach touches per week
(300 a day × 5 days)
5sales calls booked per week
2new clients closed per week
5jobs delivered to each client per week

Targets get set every Monday, reviewed every Friday. Nothing else in the business gets a weekly target. If these four are on track, everything is on track.

§1How the Business Makes Money — the whole loop on one line

The business runs city by city, trade by trade. One "market" = one trade in one city (for example: concrete leveling in Dallas). Here's the loop:

Pick a marketone trade,
one city
Create demanda local web page +
ads aimed at homeowners
Real conversationschatbot qualifies homeowners;
Akash calls contractors
Sales callcontractor signs up
Client goes live15-step setup
checklist
Jobs deliveredleads flow straight into
the client's job board
Cash10% of the revenue
we generated, auto-billed

Why this compounds: every market we open starts collecting homeowner leads before a contractor has signed up there. That stockpile is what makes cold calls work — "I have three concrete-leveling leads in Dallas right now, do you want them?" converts far better than a bare cold call. Marketing opens markets. Sales sells the leads sitting in them. Product delivers them. The results fund and prove the next market.

Honest baseline as of July 27: the real sales pipeline is about 5 contacts and 1 booked client (Dan). Ads are mostly off. Every scoreboard shows the true small numbers — no inflating. We never claim 330 accounts (about 50 is real), or 10,000+ inspections, or 150+ markets, or $14M+. Those claims are retired.

§2The Offer — one sentence, said the same way everywhere needs Akash's sign-off

Recommended wording (approve it or edit it — then it goes on every page and every call):

"Nothing down — we front all the costs. Ten or more new jobs in 90 days or you don't pay. And you only ever pay 10% of the revenue we actually generate for you."

This merges the guarantee already on the website with the 10%-of-revenue billing model in the master plan. Internal price floors per closed job (not said publicly): concrete leveling $300 · crawl space $1,000 · foundation repair $1,500.

The problem the audit found: the business is currently making four different promises

§3Sales — turning conversations into clients Akash owns this

What goes in: people who responded to ads or applied on the website, plus the daily call list (sorted so cities where we already have leads waiting come first). What comes out: signed clients.

Lead arrives
Fast first touchwithin 5 minutes
when possible
The 300100 calls + 100 texts
+ 100 emails, 9:15–1:00
Real conversationtracked in the sales system,
every call gets an outcome logged
Call bookedreminder texts + short videos
in the 3 days before the call
Sales call4–7 PM window
Client signedsetup starts
within 48 hours
The 3 numbers Sales lives by
  • Input (fully in our control): 300 touches a day
  • Middle: 5 sales calls booked a week
  • Result: 2 new clients a week
Scoreboard to build

One page showing, every day: calls dialed → people who answered → real conversations → qualified → calls booked → showed up → closed. Daily bars against the 300 target, and the percentage lost at each step so we can see exactly where deals leak.

Tools we keep
  • The sales tracking system (sales.stpierre.ai) — the one screen for all deals
  • The 7:30 AM auto-generated call list
  • The call-practice trainer
  • The show-up reminder sequence for booked calls

Akash's non-negotiables in Sales: the 300 get done before any building work, every day · every new lead gets touched the same day it arrives · if 5 calls aren't booked by Wednesday, Thursday's 300 all go to warm follow-ups · the AI drafts messages, but Akash personally sends them (nothing goes out on its own).

§4Marketing — creating the conversations AI runs it · Akash approves

What goes in: ad budget plus the market schedule (which trade, which cities, in what order — currently: concrete leveling first). What comes out: qualified conversations for Sales, and stockpiled homeowner leads that make cold calls land.

Side one — homeowner ads (this is the product's fuel):

Market opens2 new city pages
built per day
Local page livecopied from the proven
Dallas template
Ads onthe proven winning ad,
city swapped in
Homeowner respondsform or message
Chatbot qualifiesasks the right questions,
24/7
Lead routedto a paying client — or held
as proof for sales calls

Side two — contractor ads (our own client acquisition): ads aimed at contractors in the cities we're calling → they land on stpierre.ai / start.stpierre.ai → they apply → Sales takes over. Those two pages are Akash's own builds — frozen, never rebuilt. And per rule 2: no new ad concepts. We clone what wins and swap the city.

The 3 numbers Marketing lives by
  • Input: markets opened (2 city pages a day)
  • Middle: cost per qualified conversation
  • Result: qualified conversations per day
Scoreboard to build

One page per market showing: money spent → people who saw the ad → clicked → became a lead → qualified → routed to a client. With a red line marking the most we're allowed to pay per lead in that trade (5% of the trade's average job size — e.g. a $3,000 concrete-leveling job means $150 max per lead; if we don't know a trade's job size we write "unknown," never a guess).

Tools we keep
  • The live-ads dashboard (what's running, what it costs)
  • The ad-copy standard — the writing rules only; the "make new ads daily" schedule is suspended
  • The master plan (the market-by-market strategy document)

Marketing's non-negotiables: two city pages a day, no misses · every ad passes the brand checklist before Akash ever sees it · no ad ever turns on without Akash seeing screenshots and approving first · every ad is labeled either "our own client-getting ads" or "homeowner ads run for clients" so the two never get mixed up in reporting.

§5Product — delivering results and keeping clients AI runs it · goal: first lead on day one

What goes in: a newly signed client. What comes out: cash collected and a client who stays because leads keep showing up.

Client signs
Setup15-step checklist +
a final quality check
Campaign readybuilt and paused —
Akash flips it on
Leads come inchatbot qualifies
each one
Deliveredlead drops straight into the
client's own job board
Jobs closeweekly proof report
sent to the client
Cash10% of won jobs,
billed automatically
The 3 numbers Product lives by
  • Input: leads delivered per client per week (5+)
  • Middle: how many booked homeowners actually show up
  • Result: clients kept (monthly revenue retained)
Scoreboard to build

One row per client: signed → setup step (x of 15) → live → first lead (goal: day one) → first booked job → first full week delivered → weeks retained. Any client getting fewer than 5 leads in a week turns red and becomes the day's top priority.

Tools we keep
  • The lead hand-off wiring (leads flow into the client's own software — proven with Dan's job board, no middleman)
  • The client-software setup guides
  • The simple billing tracker (deliberately basic — logs won jobs, drafts Monday invoices)

Product's non-negotiables: every new client gets their first lead within week one, no exceptions · Akash is never personally forwarding leads — the wiring does it · every delivered lead has a working phone number, verified · every client gets their weekly results report without asking.

§6The Daily & Weekly Schedule

Time (Mon–Fri)BlockThe rule
8:45–9:00Morning reportRead the overnight report, approve the queue. 15 minutes max — half the day's total "managing the machine" budget.
9:00–9:15Team huddleThe scoreboard is the agenda. Only red numbers get discussed.
9:15–1:00The 300Phone only. No Slack, no Claude, no building. Call order: people who missed calls → qualified people gone quiet → client emergencies → cold calls into cities where leads are waiting.
1:00–1:15Log the countFour numbers written down: dials, answers, real conversations, qualified.
2:00–4:00FlexOnly for client-delivery emergencies. If nothing's on fire, this is free time — not building time.
4:00–7:00MeetingsSales calls and client calls. The booking link only offers this window.
7:00–8:30Builder blockONE project, chosen the night before. If you're deciding what to build at 7 PM, last night's shutdown failed.
8:30–8:45ShutdownOne voice note closes the day. The other half of the managing budget.

Weekly: Monday — set this week's targets for the four numbers, plus the cash check (how many weeks of money are in the bank, said out loud). Wednesday — pace check: are 5 calls booked? If not, Thursday's 300 go entirely to warm follow-ups. Friday afternoon — review the week: predicted vs. actual on the four numbers, kill one thing, set next week's targets. Friday or Sunday — the two-hour planning slot, which is the only door through which new ideas enter the business.

§7What We're Stopping Akash: one yes/no pass on this list

The audit found these actively costing attention or actively lying to prospects. Each line names the rule that kills it.

#Stop / fixWhy
1Publish the corrected offer page. The live team-site offer page still sells the dead $9,000 package. The rewrite exists — publish it, minus its leftover deposit step.It poisons every sales call a prospect researches us before. (Rules 2, 6)
2Shut off the "$1 to start" checkout. Retired July 15, still live, still able to take money.A dead offer sitting on a live payment path. (Rule 2)
3Delete the stale "text-messaging approval pending" warnings. The carrier approval came through July 27, but 8 documents still list it as a blocker.A false blocker sitting on the most important delivery path. (Standing order)
4Suspend the "new ads every day" machine. Keep the winning ads running; stop manufacturing new ones.The ads are the ads. (Rules 2, 3, 4)
5Retire the old "four departments + 27 creatives a week" org chart still on the team site.We run three departments now, and that page steers the builder toward the least important work. (Rule 4)
6Park the personal-brand content kit (the Cancun launch plan, flashcard trainers, scripting masterclass, shoot scripts).It trains for a funnel we retired. Parked, not deleted. (Rules 2–4)
7Park the 33-lesson builder course and vision video script. The task board already carries the same information in checklist form.Polish for a team of one. (Rule 3)
8Merge the three ad dashboards into one. Two of the three are showing stale or broken data anyway.Three screens for one question. (Rule 3 — final retirement call is Akash's)
9Abandon the two half-finished funnels and the quiz waiting on domain hookups. Never connect them.New funnels are banned. (Rule 2)
10Clean up the leftovers: the old big-budget contractor ad plan (the market map survives, the ads don't), the oversized client-software rebuild proposal, the dashboard that mixes in another business's 40 clients, and 32 stale open decision records.All superseded. (Rules 3, 7)

§8House Rules — always true, fixed on sight

§9Who Does What

WhoOwnsNever does
AkashThe 300 daily touches · sales calls · closing · two 15-minute approval passes a day · the evening builder block · recruiting settersBuilding before the 300 are done · maintaining infrastructure · personally forwarding leads to clients
The AI (Victor + the agent fleet)Building city pages · running the chatbots · routing leads · writing reports · keeping dashboards current · drafting every message · keeping this playbook trueSending anything to a real person without approval · turning ads on · flipping any live switch
Collins (on hold)Builder projects from the agreed six-system list, re-ranked so closing-related systems come first — pulled from the task board one at a timeReceiving anything before scope, pay, and go-ahead are settled · contacting clients · being measured on creative volume

v1.1 · plain-English edition · replaces: the four-department org chart, the $9,000 package, the $1-start funnel, the guru-style pricing plan, the daily-creative schedule, and the "media company" identity. When documents disagree: this page first, the market-by-market master plan second, nothing else.