Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    deal forecast and probability calibrator

    1

    Sales forecasts often rely on arbitrary stage probabilities that do not reflect reality.

    $2.99

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    deal forecast and probability calibrator

    deal forecast and probability calibrator

    Example session with this skill installed

    Calibrate our stage probabilities using the attached CSVs for North America. We use 20% for Discovery, 50% for Proposal, and 80% for Negotiation. Min sample 10. Reporting in USD.

    • Read your context and instructions
    • Compiled the deal forecast

    I have processed the historical deal data and open pipeline for the North America business unit. The analysis compares your assigned probabilities against observed outcomes from 450 closed records.

    Deal Forecast and Probability Calibration — Global Corp, 2024-05-20

    1. Scope, sources and assumptions

    • Sources: E1 (Stage Table), E2 (Closed Deals Q1-Q3), E3 (Open Pipeline).
    • As-of Date: 2024-05-20 | Currency: USD.
    • Access Scope: North America (granted by Sales Ops).
    • Calibration Basis: Amount | Min Sample: 10.

    2. Stage calibration

    StageAssignedClosedWonObserved rate95% intervalCalibratedBasisConsistency
    Discovery0.2000200300.1500[0.11, 0.19]0.1500amountOverstated
    Proposal0.5000150720.4800[0.42, 0.54]0.4800amountConsistent
    Negotiation0.8000100850.8500[0.79, 0.91]0.8500amountConsistent

    3. Back-test on closed deals

    StageExpected WinActual WinVariance
    Discovery$400,000$300,000-$100,000
    Totals$1,850,000$1,720,000-$130,000
    • Realised-to-Expected Ratio: 0.93
    • Assigned Brier Score: 0.184
    • Calibrated Brier Score: 0.142 (In-sample estimate)

    4. Open-pipeline forecast

    MetricValue
    Total Pipeline$5,000,000
    Stage-based Expected$2,450,000
    Calibrated Expected$2,210,000
    Difference-$240,000 (-9.8%)

    Analytical estimate — validation required

    5. Exclusions, consent and access

    SetReasonCountAmount
    EMEA RecordsOutside Scope45$900,000
    No ConsentPrivacy12$150,000

    6. Warnings and data gaps

    W04-LOW-SAMPLE: "Closing" stage had only 4 records; fallback to assigned 90% probability used.

    Information gap — organizational validation required.

    7. Interpretation, draft actions and items for human review

    • The Discovery stage is statistically overstated (computed result).
    • Analytical estimate: The current weighted forecast is likely aggressive by ~10% based on historical yield.
    • Draft Action: Update Discovery stage probability to 15% in CRM (requires_human_approval: true).
    • Management decision required: No CRM records were changed.

    This analysis supports, and does not replace, professional judgment and management's decision.

    Next steps

    1. Review the $240k forecast gap with the finance team.
    2. Verify if recent Discovery stage changes explain the lower observed win rate.
    3. Approve the draft probability adjustments for the next fiscal period.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Sales forecasts often rely on arbitrary stage probabilities that do not reflect reality. This skill compares your assigned CRM win rates against historical closed-deal outcomes to verify accuracy and provide a calibrated pipeline forecast.

    What it does

    • Probability calibration calculates actual win rates per stage with 95% confidence intervals.
    • Back-testing compares what assigned probabilities would have predicted against real-world revenue won.
    • Forecast adjustment generates an open-pipeline forecast using both assigned and observed historical probabilities.
    • Data auditing identifies stages with insufficient history for calibration and flags records missing consent or approved offers.
    • Statistical validation computes Brier scores to measure the predictive accuracy of your current forecasting model.

    How it works

    1. Material review identifies stage definitions, closed-deal history, and current open pipeline from your provided extracts.
    2. Request generation builds a request.json file mapping deal IDs, amounts, stages, and outcomes while enforcing business unit scopes.
    3. Deterministic execution runs a Python script to perform exact decimal arithmetic for win rates and expected revenue.
    4. Result synthesis produces a structured report comparing assigned vs. calibrated revenue across the pipeline.

    Frameworks & tools

    The skill uses a dedicated Python execution environment running a custom arithmetic script (scripts/run.py). It processes JSON-based deal exports and stage schemas.

    Why this beats prompting it yourself

    Standard LLM prompts often hallucinate win rates or struggle with precise decimal math over large datasets. This skill offloads all arithmetic to a deterministic script, ensuring 95% confidence intervals and Brier scores are mathematically sound rather than guessed.

    Use cases

    • Auditing CRM stage probabilities to see if a 50% "Proposal" stage actually closes at that frequency.
    • Generating a "calibrated" revenue forecast for an upcoming quarter based on the last 12 months of performance.
    • Identifying which parts of the sales funnel lack enough data for reliable statistical forecasting.

    Known limitations

    This skill does not score individual deal health, predict single deal outcomes, or plan sales quotas. It ignores records without explicit data use consent or approved offers.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Sales & Lead Gen.

    Does not predict single deal outcomes. Requires minimum 5 samples per stage for calibration. Excludes records without explicit consent or approved offers. No CRM write access.

    How to use

    1. 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. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    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 2 days ago

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Environment Variables
    Listed2 days ago

    What's inside

    Frequently Asked Questions

    This skill solves the problem of 'forecast fluff' by replacing arbitrary CRM stage percentages with mathematically verified win rates based on your actual historical deal outcomes.

    The skill requires a structured extract of your deal history and current pipeline in JSON format, including deal IDs, amounts, stages, and final outcomes (won/lost).

    No, this skill is designed for pipeline-level statistical analysis; it calculates the probability of success for groups of deals rather than predicting the outcome of a single specific opportunity.

    Unlike standard prompts that may struggle with large-scale decimal math, this skill uses a deterministic Python environment to perform precise arithmetic, ensuring your Brier scores and confidence intervals are mathematically exact.

    You will receive the core logic script, the stage schema for data mapping, and a comprehensive guide on how to format your CRM extracts for successful execution.

    For data privacy and accuracy, the skill automatically filters out any records that are missing explicit data use consent or lack officially approved offers.

    New in Sales & Lead Gen