Agency Sales System . Agency OS

What it does

Produces the machine your agency uses to win clients: a positioning statement and differentiation matrix, four outreach sequences (cold email, cold SMS, warm nurture, post-discovery ghost follow-up), a merge-field proposal template, anonymised case studies pulled from your completed client workspaces, eight objection scripts, an eight-stage GHL pipeline with six workflow automations and a Conversation AI qualifier, and two to four pricing tiers with revenue modelling. Everything is written GHL-ready so you can implement it the same day. This is an agency asset you build once in your own _default workspace, not a per-product build you repeat.

Say this to start

This skill has no button. You start it by saying what you want. Any of these will do it:

/sell-agency

> build my agency sales system

> client acquisition system

> cold outreach sequences

> agency proposal template

> objection handling scripts

When to reach for it

you deliver well but freeze on selling yourself
no repeatable way to book discovery calls
outreach is ad hoc and nothing follows up
every proposal is written from scratch
pricing conversations go badly or the retainer is too low
prospects go quiet after the discovery call

When NOT to use it

If you actually want	Use this instead
Closing the inbound leads your own funnel and ads already produced	`sales-conversion-engine`
Building a sales system for a client, inside a client workspace	`sales-conversion-engine`

Before you start

What you need	Why
Your `_default` workspace active, not a client workspace	this skill exists to get your agency more clients, not to help your clients get customers, and that distinction drives every decision it makes	Required
An offer document (`offer-output.md`)	the skill checks for it before Phase 1 and offers to run `/build-offer`, because outreach, proposals and objection scripts all anchor to the offer	Required
A GoHighLevel account for your agency, and admin access to it	Phase 5 builds the pipeline and six workflow automations, and they live in your main GHL agency account, never in a client sub-account	Required
Answers to the Phase 1 discovery questions	it asks what services you sell, who your ideal client is, what you charge, your current and target client count, your biggest acquisition challenge, whether you have case studies, and what is already set up in GHL; it will not re-ask for anything already in your workspace files	Required
A `verified-claims.md` registry with entries marked `consent: ok-as-stated`	results, numbers and testimonials in outreach and proposals only survive if they trace to it; anything else comes out as a `[PROOF GAP]` marker	Required
A name for this sales track	sales assets are multi-variant, so the skill asks for a name and files the deliverables under it	Required
GoHighLevel MCP connected	without it the pipeline, workflows and Conversation AI are still written out in full but as documentation you implement by hand	Optional
At least one completed client workspace with `ica-output.md` and `offer-output.md`	the case study generator cross-reads them; with none it skips case studies and offers to format results you supply verbally	Optional
A brand voice profile (`voice-profile.md`)	without it the cold outreach reads generic; the skill offers a quick voice build from a website URL or a two or three sentence description	Optional
Existing `positioning.md` and `audience.md`	used to pre-populate the differentiation matrix and ideal client profile; building them from scratch is normal on a first run	Optional
Web research tools available (Firecrawl, Perplexity, or WebSearch and WebFetch)	market research runs before Phase 1 and is blocking; it looks at competitor outreach, current response-rate benchmarks and which channels are being spam-filtered right now	Optional

How it runs

Phase 1: agency discoveryIt reads your workspace files first, then asks conversationally for what is missing: services, ideal client, current pricing, client count and target, biggest acquisition challenge, case studies, and GHL status. It plays the summary back and waits for you to confirm before building.
Market research and track namingA blocking research pass on competitor outreach templates and discovery scripts, proposal and pricing structures, current response-rate benchmarks, and channel state including what is getting flagged as spam. It also asks you to name this sales track.
Phase 2: positioning and differentiationIt builds a positioning statement to a fixed formula, then validates it: the audience must be specific, the differentiator must be hard for a competitor to copy, the outcome must be measurable. It also builds a differentiation matrix against DIY, other agencies and an in-house hire, plus your ideal client profile including disqualifiers.
Phase 3: outreach systemFour sequences, each with subject lines, body copy, CTAs, send timing and the GHL workflow configuration: cold email (5 to 7), cold SMS (3 to 5, each under 160 characters), warm nurture (5) for engaged leads who did not book, and post-discovery ghost follow-up (3).
Phase 4: sales toolsA merge-field proposal template covering executive summary through next steps, a case study generator that cross-reads your client workspaces read-only and anonymises by default, and eight objection scripts each with the psychology, a reframe, a verbatim response, a proof point and a persistent pushback line.
Phase 5: GHL pipeline and automationAn eight-stage pipeline from New Lead to Closed Won or Closed Lost, six workflow automations connecting the stages, and a Conversation AI agent that qualifies inbound leads through a five-step flow and routes them to the right stage.
Phase 6: pricing and packagingTwo to four tiers built around the $47 per month per-client GHL sub-account cost, with revenue modelling at 10, 20 and 50 clients, per-client economics down to net margin, and a packaging strategy that anchors high and makes the middle tier the obvious choice.
CloseoutProspect-facing assets are run past the hook, copy and claim graders before you see them. Then it scores the work, saves .docx files, updates the dashboard, asks whether you want presentation slides, writes the agency-sales-output schema, appends to assets.md, and asks how each deliverable landed.

What you get

An agency-sales-output.md schema file, with variant deliverables in their own subfolder
A positioning statement, a differentiation matrix against three alternatives, and your agency's ideal client profile including disqualifiers
Cold email sequence (5 to 7 emails with subject, body, CTA and timing)
Cold SMS sequence (3 to 5 messages, each under 160 characters)
Warm nurture sequence (5 emails) and post-discovery ghost follow-up (3 emails)
GHL workflow configuration for each sequence
A proposal template with merge fields and a registry of those fields, covering executive summary, problem analysis, solution, deliverables and timeline, investment, guarantee and next steps
Anonymised case studies in situation, challenge, solution, results and quote format, generated from your completed client workspaces
Eight objection scripts: too expensive, burned before, want in-house, guarantee demands, need to think, no budget, competitor comparison, scope reduction
An eight-stage GHL pipeline, six workflow automations and a Conversation AI qualification configuration
Two to four pricing tiers, a revenue model at 10, 20 and 50 clients, and per-client economics
A .docx of each deliverable, a dashboard section, and optional cumulative slides

Honest limits

Read this before you rely on it

The pipeline and automations need a GoHighLevel agency account. Without the GHL MCP connected everything is still written out, but as configuration documentation you implement by hand.
Case studies need at least one completed client workspace with ICA and offer data. With none, that part is skipped and you supply results verbally for manual formatting.
Nothing in the six phases sources or verifies a prospect list. The skill writes the sequences and the automations; the list, the sending domain and its reputation are yours.
Every number, result and testimonial in outreach and proposals must trace to verified-claims.md. Client data is anonymised by default and the skill asks permission before naming a client, but the consent conversation with that client is still yours to have.
The pricing tiers assume a $47 per month per-client sub-account cost and a $235 per month retainer floor. The skill does not know your actual labour cost or tool stack unless you tell it, so treat the revenue model as a frame, not a forecast.

Where people go wrong

The mistake	Do this instead
Running it inside a client workspace	Switch to `_default` first. This builds your agency's acquisition system, not a client's.
Building the pipeline in a client sub-account	It belongs in your main GHL agency account. Client sub-accounts are for client work.
Positioning that says 'small businesses' or 'we care about results'	The validation rules reject both. Name a specific audience, a differentiator a competitor cannot copy, and a measurable outcome with a timeframe.
Pricing a retainer without counting the sub-account cost	Retainers should be at least 5x the $47 per month sub-account cost, so $235 per month is the floor, and setup fees should cover the actual build cost.
Putting a client's name in outreach or a case study without asking	The skill anonymises by default. Get explicit permission before you name anyone in marketing material.
Expecting this to close the inbound leads your funnel produces	That is `/build-sales`. This one is for cold prospects who never raised a hand.

Worth knowing

Finish at least one client build before running this. The case study generator cross-reads completed client workspaces and writes anonymised case studies for you, so your proposals and outreach arrive with proof already in them instead of a [PROOF GAP] marker.