Command note: a pipeline full of open deals looks like momentum. Most of the time it is a graveyard. Nobody moves a card to Lost voluntarily. So stages fill with deals that stopped moving weeks ago, and the forecast quietly becomes fiction. The command I run: "List every open opportunity in this pipeline. Show days since last activity and current stage. Group them into moving, stalled over 14 days, and dead over 30." What comes back: • The stage deals actually die in, which is rarely the one you would guess • A short list worth a real follow-up today • A longer list that should be closed out or moved to nurture Do this with it: 1) Run it on every client pipeline once a week. Same day, same time. 2) Close out anything past 30 days. A clean pipeline beats a big one. 3) For the stalled bucket, ask what the stage was missing. Usually a task or an automation that never fired. 4) Fix the stage, not the deal. One fix clears the next fifty. Aging a pipeline is not a report you read. It is maintenance you perform. A pipeline you never prune stops being a forecast and starts being a story you tell yourself.

Command note: the account you break is always the one you were sure about. Every automation looks fine in the builder. The builder is not where it fails. It fails against real data: a contact with no last name, a tag that never got removed, a stage nobody uses anymore. So before anything runs on a list, run it on one. Make a single test contact in every sub-account you manage: • Your own email and phone, never the client's • Tagged so you can find it in two seconds • Left in the account permanently, not deleted after launch Then use it every time: 1) Publish the workflow and add the test contact first, nobody else. 2) Watch the full path end to end, not just the first action. 3) Check what actually landed: the SMS text, the merge fields, the timing. 4) Confirm it ends. The contact exits, or the goal fires. 5) Only then open it to the real list. This costs you four minutes. A bad send costs you the account. The same rule applies when AI builds it. AI can assemble a workflow faster than you can read it, which means the test contact matters more, not less. Speed of build raises the value of the check. One contact tells you the truth before four hundred tell your client.

Command note: the workflow you can't explain in one sentence is the one that will break at 2am. Most inherited accounts have three or four monster workflows doing everything at once. Booking, nurture, tagging, internal alerts, invoicing. When one part misfires, you can't tell which part, and you can't turn off the bad piece without killing the good ones. Give each workflow a single job: • One trigger, one outcome • A name that states the job ("Booked - confirm and remind") • An exit condition tied to that outcome, not to a side effect • Nothing in it that another workflow already does Split a monster today: 1) List every outcome the workflow produces. If there's more than one, that's your split line. 2) Pull each outcome into its own workflow with its own trigger and goal. 3) Chain them with a tag or a stage change, not with nested branches. 4) Turn the old one off. Don't delete it yet - watch it for a week. 5) Test each new one with a throwaway contact before it touches a real one. The payoff is diagnosis. When something goes wrong you get a name that tells you where, and you can pause one job instead of the whole account. A workflow that does five things is five bugs waiting for the same trigger.

Command note: the quietest leak in an inherited account is a form that collects leads and triggers nothing. Forms are easy to build and easy to forget. Someone spins one up for a promo, it stays live, and submissions land in the contact list with no tag, no notification, no follow-up. The command I run on every account I take over: "List every form and survey in this sub-account. For each one, show what workflow or automation it triggers, and flag any with no trigger, no notification, and no tag." What comes back: • Forms still live on pages nobody links to anymore • Submissions that create a contact and then stop • Two forms feeding the same workflow with different fields • Notifications routed to a staff member who left Do this today: 1) Run it on your newest inherited account first. Those leak the most. 2) For every orphan, decide: wire it or kill it. There is no third option. 3) For every form you keep, confirm one tag, one notification, one next step. 4) Re-run it monthly. Forms multiply quietly. Most agencies audit the ad spend and ignore the intake. The form is the front door of the account, and half the time nobody checks whether it opens onto anything. That lead did not go cold. It went nowhere.

Command note: we once tagged everything, and it made our segments worthless. One account had over 90 tags. Same idea spelled three ways. Tags for things the CRM already tracked. Nobody could build a clean list, so nobody trusted the lists. What went wrong: • We tagged activity (opened, clicked, visited) that GHL already records • Duplicates piled up: hot-lead, hot_lead, HotLead • Tags got added and never removed, so "current" meant nothing • Every new segment needed a guess about which tag was real The fix took an afternoon: 1) Export the tag list and count contacts per tag. Anything under 5 is noise. 2) Group the duplicates and pick one spelling. Write the naming rule down. 3) Delete every tag that describes activity the platform already logs. 4) Keep tags that answer one question: what does this person want, and where are they in the process. 5) Give every remaining tag a removal condition. If nothing removes it, it will lie to you eventually. We went from 90-plus tags to 14. Segments got obvious. The AI stopped guessing, because there was only one right answer per question. A tag is a promise about a contact. Keep fewer promises and keep them true.