Works with the AI Agents you already use
MICP Revenue Evidence & Decision Pack
Audit marketing revenue, reconcile CRM spend with finance, and generate evidence-based budget decision briefs.
$19.99
MICP Revenue Evidence & Decision Pack
Example session with this skill installed
Run the MICP Revenue Evidence & Decision Pack against its bundled synthetic sample using the documented three-stage Python workflow. Validate source records, separate annual-recurring and total-contract opportunity values, analyze funnel conversion and attribution evidence, and generate the executive decision brief and supporting CSV tables. Explicitly identify missing finance evidence. Do not invent missing data, claim causal ROI, or recommend a budget increase without sufficient evidence.
- Read your context and instructions
- Compiled the micp revenue evidence
Synthetic demonstration: Generated offline from the bundled sample dataset. This is not customer data or a live Agensi agent session.
MICP | Marketing Revenue Evidence and Decision Brief
Reporting scope: 2026-07-01 to 2026-09-30 | USD
Decision: RECONCILE BEFORE REALLOCATION
Executive conclusion
1 blocking source exclusion(s); finance booking export missing. Resolve those evidence gaps before budget changes.
The provided dataset records
10000.00 USD of marketing spend,
4 leads,
3 MQLs, and
2 qualified SQLs. Data quality review found
1 blocking exception(s).
Acquisition calculations
| Measure | Value | Definition |
|---|---|---|
| Lead to MQL | 75.0% | MQL events / eligible created-lead cohort |
| MQL to SQL | 66.7% | Valid SQL events / eligible MQL cohort |
| Marketing spend / lead | 2500.00 USD | Spend in scope / lead count |
| Marketing spend / SQL | 5000.00 USD | Spend in scope / valid SQL count |
These are descriptive cohort costs, not fully loaded CAC or incrementality.
Opportunity values (never combine ARR and TCV)
| Stage | Value basis | Amount (USD) | Supporting deal IDs |
|---|---|---|---|
| open | total_contract | 90000.00 | D2 |
| won | annual_recurring | 30000.00 | D1, D3 |
Supported marketing-source classifications
| Campaign | Stage | Value basis | Amount (USD) | Deal IDs |
|---|---|---|---|---|
| paid_social | won | annual_recurring | 20000.00 | D1 |
| partner | open | total_contract | 90000.00 | D2 |
A declared marketing-origin classification is not the same as a pre-deal touch model or causal revenue credit.
Finance booking crosscheck
Status: NOT_SUPPLIED. CRM amounts have not been checked against independent finance entries.
| Deal ID | CRM value | Finance value | Difference | Result | Finance record IDs |
|---|---|---|---|---|---|
| No eligible records |
Attribution evidence and ambiguity
Pre-deal touch lookback: 60 days. These policy-based source verdicts are not causal estimates.
| Deal | Join status | Evidence verdict | First eligible campaign | Last eligible campaign | Touch IDs |
|---|---|---|---|---|---|
| D1 | direct_deal | no_eligible_touch_found | unknown | unknown | |
| D2 | direct_deal | no_eligible_touch_found | unknown | unknown | |
| D3 | direct_deal | no_eligible_touch_found | unknown | unknown |
Fixed-maturity lead cohorts
Window: 30 days; right-censored: 0 leads.
| Campaign | Matured leads | MQL | SQL | SQL/lead | 95% Wilson interval | Sample caution |
|---|---|---|---|---|---|---|
| UNATTRIBUTED | 1 | 0 | 0 | 0.0% | 0.0% to 79.3% | small sample |
| paid_social | 2 | 2 | 1 | 50.0% | 9.4% to 90.5% | small sample |
| partner | 1 | 1 | 1 | 100.0% | 20.6% to 100.0% | small sample |
Campaign economics
| Campaign | Spend (USD) | Leads | MQL | SQL | Cost / SQL (USD) | Spend row IDs |
|---|---|---|---|---|---|---|
| UNATTRIBUTED | unknown | 1 | 0 | 0 | unknown | |
| paid_social | 6000.00 | 2 | 2 | 1 | 6000.00 | S1 |
| partner | 4000.00 | 1 | 1 | 1 | 4000.00 | S2 |
Reversible budget investigation
Scenario status: INSUFFICIENT_COMPARABLE_DATA.
No test budget is recommended: comparable mature cohorts or minimum SQL evidence are insufficient.
Compare controlled marginal response; do not forecast incremental SQL from average historical cost.
Exceptions and missing evidence
| Severity | Type | Dataset | Record ID | Explanation |
|---|---|---|---|---|
| review | SQL_WITHOUT_PRIOR_MQL | events | L4 | SQL excluded from cohort SQL metric |
| block | INVALID_DEAL | deals | D4 | unknown amount basis: no ARR/TCV conversion |
Suggested owner-led next steps
- Priority 1: Confirm excluded opportunity basis/currency and re-run. Owner: Finance + RevOps. Evidence: Deal IDs: D4. Stop/verification rule: Only verified deal values enter basis-specific totals.
- Priority 2: Supply a separate booking extract with stable deal IDs. Owner: Finance. Evidence: Only CRM-supplied values have been examined. Stop/verification rule: Stop financial claims until matching source values are supplied.
Priority 3: Verify event chronology and stage eligibility. Owner: Sales Ops. Evidence: Lead IDs: L4. Stop/verification rule: Do not backfill an MQL without timestamp evidence.
Validation and limitations
The numbers are derived from buyer-supplied records. They are neither authenticated against a live CRM nor evidence of cash collected, financial recognition or incremental marketing return.
- No live integrations or independent invoice/payment reconciliation.
- Open pipeline consists of open records created within the period, not historical as-of pipeline.
- Won ARR denotes supplied recurring deal value, not recognized or collected revenue.
- Marketing-influenced overlaps sourced; never add these buckets.
- No causal lift, ROI, fully loaded CAC or incremental ROAS established.
- Cohort stage conversions may be right-censored at period end.
- Missing channel spend is unknown, not zero.
- Inputs are buyer-provided and are not independently authenticated.
Input fingerprint SHA-256: 2530f9d2754206b64473af4dbdd0909868f96cdd7d96d0f221c5ab55dc324ec9 (proves repeatability, not source authenticity).
Connects securely to your tools. The creator never sees your data.
About this skill
Marketing teams often face conflicting CRM reports, unclear revenue attribution, incomplete funnel data, and unverified ROI claims. MICP Revenue Evidence & Decision Pack helps B2B marketers, RevOps teams, agencies, and business leaders understand what their marketing data actually supports before making investment decisions.
Combining three coordinated AI agent skills with deterministic Python calculations, MICP reconciles marketing and revenue records, diagnoses funnel performance, identifies data inconsistencies, and produces evidence-linked executive decision briefs. Rather than inventing revenue impact or attribution claims, it distinguishes verified calculations from assumptions and highlights what requires further validation.
What it does
- Reconcile Revenue Evidence: Compare campaign spend, CRM opportunities, and optional finance booking records. Identify discrepancies, distinguish marketing-sourced from influenced revenue, and separate recurring revenue values from total contract values.
- Diagnose Funnel Economics: Analyze lead-to-MQL-to-SQL conversion, campaign acquisition costs, matured cohorts, and incomplete conversion windows using 95% Wilson confidence intervals where applicable.
- Generate Executive Decision Briefs: Transform supported calculations into structured reports with evidence references, prioritized recommendations, validation requirements, and measurable stop conditions.
- Audit Data Quality: Detect duplicate records, missing identifiers, currency mismatches, invalid event sequences, attribution gaps, and unresolved financial differences.
- Evaluate Attribution Evidence: Examine first-touch and last-touch campaign relationships, compare available self-reported attribution, and identify where source classifications disagree.
- Propose Controlled Experiments: Identify potential marketing optimization tests with explicit assumptions, reversible actions, measurement requirements, and risk guardrails without claiming unproven revenue lift.
How it works
- Scope & Data Preparation: Provide authorized, preferably anonymized, marketing spend, lead-stage, CRM opportunity, attribution, and optional finance records in the supported structured format.
- Revenue Reconciliation: Execute the first Python script to validate source records, calculate eligible financial totals, and identify missing or conflicting information.
- Funnel & Attribution Diagnosis: Run the second script to analyze matured lead cohorts, campaign efficiency, financial comparisons, attribution lineage, and evidence limitations.
- Decision Brief Generation: Execute the reporting script to convert the results into readable executive reports and supporting evidence tables.
- Review & Next Steps: Examine the findings, resolve flagged exceptions, and determine which marketing decisions or controlled tests are justified by the available evidence.
Frameworks & tools
MICP combines three coordinated agent skills:
- Reconcile: Marketing revenue and financial evidence reconciliation.
- Diagnose: Funnel economics, cohort analysis, and attribution diagnostics.
- Decide: Evidence-based executive reporting and decision support.
The package uses independent Python scripts, structured JSON inputs, CSV-compatible data contracts, record-level references, and reproducible calculations.
Generated deliverables include:
- Executive decision brief in HTML and Markdown.
- Structured JSON evidence report.
- Campaign performance and acquisition-cost CSV.
- Finance reconciliation CSV.
- Attribution evidence CSV.
- Matured cohort analysis CSV.
- Data exceptions CSV.
Why this beats prompting it yourself
A conventional AI prompt may summarize marketing data without consistently validating record relationships, excluding invalid entries, or distinguishing correlation from causation.
MICP adds a repeatable computational framework that:
- Calculates supported metrics using executable logic rather than relying solely on language-model estimates.
- Preserves record-level evidence and identifies excluded or ambiguous data.
- Separates marketing-sourced, influenced, and closed-won values.
- Accounts for immature lead cohorts instead of treating incomplete conversion windows as final results.
- Compares supplied finance and CRM records without assuming that matching figures establish cash collection.
- Withholds unsupported budget recommendations when the available evidence is insufficient.
The result is a more structured and auditable foundation for marketing and revenue decisions.
Use cases
- Marketing Revenue Audits: Reconcile marketing-sourced pipeline, opportunity values, and supplied finance records.
- Campaign Performance Diagnostics: Identify funnel leakage, conversion inconsistencies, and acquisition-cost differences.
- Attribution Analysis: Compare campaign-source definitions and investigate missing or conflicting attribution evidence.
- Executive Budget Reviews: Prepare evidence-linked recommendations before reallocating marketing investments.
- Revenue Operations: Identify duplicate records, data-quality gaps, and reconciliation issues across marketing and CRM exports.
- Agency Reporting: Produce structured client-facing findings, exceptions, and supporting evidence from authorized datasets.
- Marketing Experiment Planning: Define controlled, measurable tests with clear assumptions and stop conditions.
Known limitations
- Requires Python 3.11+ and a compatible AI agent capable of executing scripts and reading/writing authorized files.
- Requires buyer-provided data in supported structured formats. Some data preparation or column mapping may be necessary.
- Does not connect directly to live CRM, advertising, analytics, or accounting platforms.
- Does not independently authenticate financial records, verify collected cash, or certify recognized revenue.
- Does not establish causal attribution, prove incremental marketing ROI, or guarantee revenue improvements.
- Does not autonomously modify campaign budgets or advertising-platform settings.
- Does not generate PDF files; executive reports are available in HTML and Markdown.
- Results depend on the accuracy, completeness, and consistency of the supplied records.
Requires Python environment for script execution. No live CRM/Ad platform API access. Does not generate PDF files. Requires pre-normalized JSON or specific CSV data structures.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 1 day ago
Requires