Client Winning

Version 1.1.0 | $999 one-time file product

Client Winning turns supplied business evidence into an evidence-bound prospect review, competitor-ad teardown and spec-work pitch draft, with an optional evidence-to-creative production handoff. Start with the worked example, then replace its invented facts with your own. The product gives you an inspectable analysis and a defined human handoff.

In the box

• A complete operator procedure in SKILL.md, usable in an AI assistant or manually.
• A Python report engine that checks inputs and calculates the defined metrics offline.
• Input contract, copyable input, two worked cases (complete and missing evidence) and expected reports.
• Optional permitted-source ledger, original creative-brief alternatives, three channel handoff checklists and an experiment log.
• Tests for the core calculations and failure states, an explicit dependency contract and source/provenance manifest.

First run

Requires Python 3.10 or newer. No packages, keys or network connection are required. Unzip in a folder you control. From this product folder:

python3 tools/analyze.py --input examples/input.json --output my-first-report
python3 -m unittest discover -s tests -v

Open my-first-report/report.md. Structured results and the exact input SHA-256 are in report.json. Existing output directories are refused so a new run cannot silently replace a saved baseline. An invalid input returns exit code 2 and creates no report. A valid but incomplete input creates a clearly marked report; the business result is not silently promoted to ready.

What you provide

Business identities, at least two sourced facts per prospect, relationship/suppression state, operator fit ratings, ad/page observations, seller proof and a specific deliverable scope.

What to expect

The example creates one fictional pitch draft and holds one opted-out and one replied prospect. Their identical fit scores never override suppression or human control. Read START-HERE.md to replace the example safely. You are buying a reusable file product and offline analysis tool, not a hosted agent, a guaranteed business result or ongoing metered service. No recurring file-access charge applies. The included code runs independently of an AI subscription; interpreting evidence with an AI assistant uses your own account.

Scope

No contact discovery, enrichment, email sending, CRM writes, scraping or automated follow-up. Public ad presence does not prove conversion. Pitches are human-review drafts and require actual sending terms and applicable compliance before use. Inputs can be manually prepared from exports. The product does not include a direct API adapter. If you later add one, its permissions, vendor charges and credentials are yours; nothing here starts it or spends money.

Install and removal

No system installation is needed. Run from this directory. If you use an AI assistant, attach SKILL.md, INPUT-CONTRACT.md and your sanitized evidence deliberately. Do not auto-import unknown files as instructions. To uninstall, remove this product directory only after saving your own inputs/reports. There are no background jobs, runtime settings, home-directory edits or hidden files to remove.

See SUPPORT.md, DEPENDENCIES.md, EVIDENCE-AND-SAFETY.md and provenance.json. This package has offline synthetic acceptance tests. A passed fixture is not customer outcome proof.

Research-to-creative extension

Read CREATIVE-WORKFLOW.md, then run python3 tools/analyze.py --input examples/creative-input.json --output my-creative-report. The human or AI assistant authors the original alternatives from permitted observations; the offline engine verifies references, applies evidence/rights holds and produces channel-specific handoffs. No assets or performance results are generated. Existing v1 inputs remain supported.