Developer Portal › Agent Recipes

Agent Recipes

Copy-paste prompt templates for common pricing intelligence workflows. Designed for Claude, GPT-4, and any MCP-enabled AI assistant.

Requires the Nesika MCP server

These recipes use Nesika MCP tools. Connect first in the MCP setup guide.

Alert When Undercut

monitoringbeginner

Automatically detect when a competitor drops below your price on any tracked product and summarise the gap.

MCP Tools:my_productsmy_price_positionmy_insights_summary

Prompt Template

You are a pricing intelligence assistant for my retail business.

Check all my tracked products using the my_products tool, then for each product use my_price_position to check my competitive position.

Report any products where:
1. A competitor is cheaper than me (I am being undercut)
2. The price gap is more than 3%

For each undercut product, tell me:
- Product name
- My current price
- The cheapest competitor and their price
- The gap as a percentage
- A recommendation: should I match, beat by $1, or hold?

Prioritise by the size of the gap (largest gap first).

Weekly Price Audit

reportingbeginner

Run a full competitive pricing audit across your portfolio and produce an executive summary with action items.

MCP Tools:my_productsmy_price_positionmy_insights_summarymy_price_changes

Prompt Template

You are my pricing analyst. Run a complete weekly pricing audit.

1. Use my_products to get my full product catalogue
2. Use my_insights_summary for the high-level overview
3. Use my_price_changes to see what moved this week
4. For my top 10 highest-revenue products, use my_price_position to check competitive rank

Produce a structured report with:

## Executive Summary
- How many products are competitively priced vs undercut vs at opportunity
- Overall portfolio health score (1 to 10)

## Priority Actions (top 3)
- Products requiring immediate repricing with specific recommendations

## Price Movements This Week
- What changed and why it matters

## Opportunities
- Products where I could raise price without losing competitive position

Export Scan Report to CSV

reportingbeginner

Turn your latest scan into a ready-to-share CSV (or JSON) report with the same columns as the scan-completed email.

MCP Tools:export_scan_reportmy_scan_results

Prompt Template

Export my latest scan report so I can share it with my team.

1. Use my_scan_results to confirm my most recent scan completed successfully
2. Use export_scan_report (scanId: "latest", format: "csv") to generate the report

The CSV uses the same columns as the scan-completed email report: SKU, Product Name,
Department, Category, Brand, Image URL, Owner, Customer Price, Product URL, Pricing
Position, then per-competitor Regular / Promo / Delta% / URL.

Then:
- Tell me how many products and competitors are included
- Call out any rows where a competitor is below my price (negative delta)

If I ask for a single category, re-run export_scan_report with the category filter
(for example "Give me a CSV of my Electronics scan results"). If I prefer machine-readable
output, re-run with format: "json".

Competitor Activity Overview

analysisintermediate

Map who you compete with most, how much your ranges overlap, and where competitors are undercutting you.

MCP Tools:my_competitor_overviewmy_productsmy_price_position

Prompt Template

Give me an overview of my competitors and where I stand against them.

1. Use my_competitor_overview to get the big picture: tracked product count, how many of my
   products are being undercut, and my top competitors
2. Use my_products to understand which categories those overlaps fall into
3. For the products where I am being undercut, use my_price_position to quantify the gap

Produce:

## Competitor Landscape
- Who are my top competitors and how many products do we overlap on?
- How many of my products are being undercut right now?

## Where It Hurts Most
- The categories and products where competitors undercut me most aggressively

## Recommended Response
- 3 specific actions I should take based on this analysis

Price Position Dashboard

monitoringintermediate

Generate a portfolio-wide competitive position snapshot with rankings and percentile gaps.

MCP Tools:my_productsmy_price_positionmy_competitor_overview

Prompt Template

Build me a competitive position dashboard for my entire product portfolio.

1. Use my_products to list all my active tracked products
2. For each product, use my_price_position to get my percentile and gap to the cheapest competitor
3. Use my_competitor_overview to see overall how many of my products are being undercut

Output a markdown table:

| Product | My Price | Percentile | Gap to Cheapest | Status |
|---------|----------|------------|-----------------|--------|

Then summarise:
- % of portfolio where I am the price leader
- % where I am competitive (within 5%)
- % where I am uncompetitive (more than 5% above the cheapest)
- Top 5 repricing opportunities with estimated impact

Daily Pricing Briefing

reportingbeginner

Start your day with an AI-curated briefing of overnight scan results, price movements, and competitor activity.

MCP Tools:my_scan_resultsmy_price_changesmy_competitor_overviewmy_insights_summary

Prompt Template

Generate my daily pricing intelligence briefing for today.

1. Use my_scan_results to check the latest scan completed overnight
2. Use my_price_changes to see what competitor prices changed in the last 24 hours
3. Use my_competitor_overview for a snapshot of key competitor activity
4. Use my_insights_summary to surface the highest-severity insight cards

Format as a morning briefing:

# Pricing Intelligence Briefing

## Overnight Scan
[Did my latest scan complete, and how many products / competitors it covered]

## My Portfolio in the Last 24h
[What changed in competitor prices and why it matters]

## Competitor Activity
[Notable competitor moves that could affect me]

## Action Required Today
[Numbered list of the 1 to 3 highest priority actions]

## Nothing to Worry About
[Brief note on what is stable]

Repricing Decision Engine

automationadvanced

Feed current competitive data to an agent that recommends precise repricing actions with margin guardrails.

MCP Tools:my_productsmy_price_positionmy_price_changesmy_insights_summary

Prompt Template

Act as my automated repricing engine. Apply these rules:

**My Constraints:**
- Minimum margin: 15% (cost is embedded in product metadata)
- Never price below [FLOOR PRICE] on any product
- Maximum price reduction in one move: 8%
- Target position: within the top 3 cheapest, or price leader if margin allows

**Process:**
1. Use my_products to get the full catalogue with current prices
2. For each product, use my_price_position to get my percentile and gap to the cheapest competitor
3. Use my_price_changes to factor in competitor moves from the last 7 days
4. Use my_insights_summary to respect any open recommendations
5. Calculate the optimal price adjustment for each product

**Output a repricing schedule:**
| SKU | Product | Current Price | Recommended Price | Change | Position Impact | Rationale |
|-----|---------|---------------|-------------------|--------|-----------------|-----------|

Then flag:
- Any product where hitting the top 3 would require going below the floor price (DO NOT REPRICE, flag for manual review)
- Any product where raising the price is safe (no competitive risk)

Total revenue impact estimate if all recommendations are applied.

Scan Health & Insights Triage

analysisadvanced

Verify your latest scan succeeded, triage the insight cards it produced, and export the evidence as CSV.

MCP Tools:my_scan_resultsmy_insights_summarymy_price_changesexport_scan_report

Prompt Template

Run a health check on my latest scan and triage what it found.

**Step 1 - Scan Health**
Use my_scan_results to confirm my most recent scan completed successfully. Report the scan id,
status, type, job counts, and product count. If any scan failed, list the failed scans.

**Step 2 - Insights Triage**
Use my_insights_summary to pull the insight cards. Group them by severity (high first) and, for
each high-severity card, summarise the product, what changed, why it matters, and the recommendation.

**Step 3 - Movement Check**
Use my_price_changes to cross-check which competitor price moves drove the new insights.

**Step 4 - Evidence**
Use export_scan_report (scanId: "latest", format: "csv") to attach the full report.

**Produce:**

## Scan Health
- Scan id, status, and coverage
- Any failed scans that need a re-run

## Priority Insights
- High-severity cards with recommended actions

## Supporting Evidence
- Confirmation the CSV export is ready and what it contains

Ready to run these recipes?

Connect the Nesika MCP server to your AI assistant in under 2 minutes. Get your API key from Settings and follow the setup guide.

Logos provided by Logo.dev