Ask AI

Diagnostic tool for 'why did revenue change?' questions. Bundles headline metric deltas, product mix shifts, and pre-generated ranked hypotheses into one call — so you don't have to stitch signals together yourself and risk asserting causation. Each hypothesis includes: what signal supports it, what else would be true if it's correct, and which tool to call next to confirm or rule it out. Use when

Anomaly detection across revenue, conversion rate, and refund rate for each of your stores (works for a single store too). Compares the most recent 7-day window against the prior 4 weeks and flags moves outside +/- 1.5 standard deviations or 25% — whichever is stricter. Returns ranked anomalies with store, metric, current vs baseline, severity, and a one-line explanation. High-severity anomalies c

Report the health of every connected data source for each of your stores (works for a single store too). For each provider/instance: last successful sync timestamp, record count, freshness flag (stale if > 36h), and key-field nullability rates (e.g. % of customers without firstOrderAt populated). Use this to diagnose 'why is metric X showing 0?' or to confirm data is current before reporting numbe

CALL THIS FIRST to see which data sources are connected and have data. Returns connection status and record counts for: Shopify (always connected), Triple Whale, Klaviyo, Gorgias, Recharge, Google Search Console, Google Analytics, Microsoft Clarity, and YouTube. Use this to understand what data is available before making other queries. If a source shows 'not_connected', those tools will return emp

Get a unified dashboard with key metrics from ALL connected data sources in one call. Returns a lean snapshot of revenue, traffic, marketing, email, organic search, UX, support, reviews, subscriptions, and loyalty — only for sources that are connected. Use this for 'Give me an overview' or 'How is the business doing?' then drill into specifics with individual tools. Note: in multi-store mode, the

BEST FOR COMPARISONS AND THE CANONICAL CONVERSION RATE: Get core e-commerce metrics (Revenue, Sessions, CVR, AOV, Revenue per Session) with period-over-period comparisons. CVR here uses Shopify TrafficStat sessions (Plus/Advanced) — or GA4 sessions as fallback — divided by REAL Shopify orders. This is the right source for 'what's my conversion rate?' / 'how's my CVR trending?' questions. Do NOT us

Get metrics with MULTIPLE comparisons at once. Perfect for questions like 'Show me revenue WoW and YoY' or 'Compare this week to last week and same week last year'. Returns { period, metrics, comparisons, charts, presentation }. `charts` is an array of grouped-bar specs (one per headline metric) with 3+ bars each (current + each comparison period). DEFAULT TO RENDERING THE RELEVANT CHART for multi

Month-by-month year-over-year comparison for a single metric. SINGLE-CALL ANSWER for 'how is revenue this year vs last year by month?' / 'show me 2026 vs 2025 monthly trends'. Returns { metric, currentYear, compareYear, rows, totals, charts, presentation, seeAlso }. `rows` is one entry per month with current/compare/change. `charts[0]` is a ready-to-render grouped-bar spec (seriesField='year') — d

Forecast a metric — answers 'what should I expect?' / 'are we on track this month?' / 'project next month'. Returns a per-store `stores` array (each in its OWN currency — never summed across currencies). Each store has: `currentMonth` (month-to-date actual + projected month-end with low/high band, built from the live daily run-rate), `horizon` (future full months with expected/low/high), `method`,

Store a merchant's GOAL for a metric over a calendar period ('we need £80k this month'). TRIGGER: whenever the user states a target/goal/budget for a metric. This saves the merchant's own target in our DB (NOT a write-back to any connected tool). Pacing (are we on track?) is computed later on read by get_targets — never stored. Money metrics (revenue, aov, ltv, …) MUST be scoped to one store via `

List the merchant's active targets with LIVE actual-vs-target pacing, recomputed from ground truth on every call (never stored). Each target returns a `pacing` block: status (ahead / on_track / at_risk / behind / too_early / not_paceable), actualToDate, projected month-end with its interval, progressPercent, gapToTarget, currentDailyPace vs requiredDailyRunRate, and days elapsed/remaining. Live mo

Get daily/period numbers from one of three domains — pass `domain` to pick which, `report` to pick the report within it. `report:"daily"` exists in ALL THREE domains and means a DIFFERENT payload in each — always pass both `domain` and `report` together, never assume "daily" behaves the same across domains. • domain:"funnel" — the daily series WITH sessions/CVR/funnel steps. Reports: daily (defau

Get customer behavior and retention insights. All reports are per-store in multi-store mode — use the store param to target a specific store. Reports: new_vs_returning (revenue/orders/AOV split by first-time vs repeat buyers), repeat_metrics (repeat purchase rate, avg days between purchases, BOTH avgLifetimeValue and medianLifetimeValue, lifetimeValueSkewRatio + distribution note — LEAD WITH MEDIA

Get refund/return analytics. Reports: summary (total refunds, refund rate, % of revenue — supports compare for WoW/MoM/YoY trend), top_products (most refunded products by amount or rate), by_reason (classifies refund notes into buckets: size_fit_too_small, size_fit_too_large, quality_defect, wrong_item, shipping_late_or_lost, color_style_mismatch, comfort, changed_mind, duplicate, out_of_stock — w

Get discount code analytics. Reports: summary (usage rate, AOV impact), top_codes (best performing codes by revenue/usage), code_details (specific code performance — requires 'code' param), by_product (which discount codes were applied to which products — optionally filter by 'code').

Get sales-velocity analytics from one of two angles — pass `angle` to pick which, `report` to pick the report within it. • angle:"inventory" — the INVENTORY-management angle: how fast stock is moving and what to reorder. Reports: summary (sell-through rate, turnover, health status), by_product (products sorted by sales velocity), restock (reorder quantity recommendations from velocity, lead time,

Get marketing performance dashboard — INCLUDING the canonical checkout funnel. ALSO USE THIS for 'checkout funnel' / 'where do customers drop off' / 'how is my funnel performing' questions; do NOT manually stitch a funnel from get_analytics + get_store_summary. Returns: `funnel` (sessions, addToCartRate, checkoutRate, purchaseRate, cvr — sessions sourced from Shopify TrafficStat when available, GA

Reconcile ad-spend numbers between Triple Whale's pixel attribution and the ad-platform APIs (Google Ads, Meta, TikTok). Same channel, different numbers — pixel typically captures 10-30% of true spend due to ad blockers, consent banners, and iOS ATT. Returns per-channel platform-API spend, TW pixel spend, ratio, severity (match / moderate / severe / critical), trueRoas (TW revenue / platform spend

Get products with marketing channel attribution showing which channels drive sales for each product. Attribution uses Triple Whale last-click data joined to Shopify order line items — requires Triple Whale to be connected. When Triple Whale is not connected, all orders appear under 'unattributed'. Use for questions like 'Which channels drive sales of product X?' or 'What products does Facebook sel

Co-purchase / market-basket analysis. Given an anchor product (productTitle), returns other products frequently bought with it, ranked by lift. Reports: basket (same-order co-purchase — best for bundles, 'frequently bought together' widgets, post-checkout upsells), lifetime (same-customer across all their orders — best for email flows and longer-horizon recommendations; also returns avgDaysToCoBuy

Get overall store metrics from Shopify order data ONLY: revenue, orders, average order value, items sold, discounts, unique customers — all for the specified period. Includes percentage changes vs prior period. Also returns all-time totals (orders, products, customers). Use this for high-level store performance questions. For an overview spanning ALL connected sources (traffic, email, support, rev

List individual orders matching filters: date range, financial status, customerId, min/max order total, order tags — sortable by date or total. Returns one row per order with name, status, fulfillment, orderTotal (formatted), items count, country, tags, and date. Use for 'show me recent orders', 'orders over $500', 'this customer's pending orders', 'orders tagged wholesale'. DISAMBIGUATION: for AG

Look up a specific order with full line item details (product, variant, SKU, quantity, price, vendor). Use orderName for user-facing IDs like '#1001', or orderId for Shopify numeric IDs. Also supports orderIds (array) to look up multiple orders at once — useful for cross-referencing with Triple Whale attribution data.

Get order analytics by dimension. Reports: by_status (order counts/revenue by financial status), by_country (geographic sales breakdown — supports compare for WoW/MoM/YoY trend, returns top rising/falling countries), by_product (BEST FOR: 'how many customers bought product X?', 'which products have the most unique buyers?', product-level customer counts. Returns uniqueCustomers, orderCount, unitsS

Get top selling products ranked by revenue or quantity. Set level to 'variant' to break down by variant (size, colour, etc.) — includes SKU, discount, and profit margin when cost data is available. Set level to 'category' for a revenue / COGS / gross-profit / margin-% rollup by product TYPE (e.g. 'Sewing Machines vs Fabric') — the structured answer to category-P&L / margin-by-category questions (m

Get product catalog with inventory levels, pricing, cost, margins, status, and SEO metadata. Returns: title, vendor, productType, price, cost, margin, marginPercent, totalInventory, seoTitle (Shopify global.title_tag), seoDescription (Shopify global.description_tag), productUrl (full storefront URL — use this directly instead of guessing). Use for questions about products, pricing, profitability,

Did THAT campaign actually work? Cross-source synthesis tool that measures the real impact of a specific Klaviyo campaign by comparing the post-send window against a DAY-OF-WEEK-ALIGNED baseline (same N days one week earlier). Returns: the campaign metadata, Klaviyo's own attribution claim, the actual Shopify orders/revenue/new-customers in the post-send window, the baseline counterfactual, the li

CROSS-SOURCE PRODUCT HEALTH SCAN — one call that returns per-product reviews + refunds + sales velocity + inventory + a composite 'needs attention' score (0-100). Use this for 'which products need fixing?' / 'what should I look at?' / 'are there any product issues?' / 'which products are performing badly?'. Saves the LLM from stitching get_top_products + get_refunds + get_reviews + get_inventory m

Get Shopify product sales analytics — units sold and revenue per product from ShopifyQL. Requires Shopify Plus or Advanced. NOTE: Views, add-to-cart, and purchase funnel counts are no longer available from ShopifyQL (Shopify removed these fields) — use get_marketing_performance for funnel analysis or get_analytics(ecommerce) for GA4-based product funnel data. Reports: summary (total units/revenue)

Get product collections and their contents, including SEO metadata. Reports: list (all collections with product counts plus seoTitle/seoDescription), products (products in a specific collection with pricing, inventory, SEO, and optional sales data). seoTitle/seoDescription correspond to Shopify's global.title_tag and global.description_tag metafields. Use for 'how many products in the sandals coll

Get customer analytics by dimension. Reports: segments (one-time/returning/VIP/at-risk), top (by spend or orders — includes customerId for each customer), by_country (geographic distribution — country is derived from order shipping address, not customer records; Shopify does not expose customer country without protected data access approval), customer_history (full order history with line items fo

Get inventory analytics. Reports: summary (activeProducts count, totalVariants, outOfStock variant count, lowStock variant count with >0 and ≤10 units, totalUnits, active locations count), low_stock (products needing reorder — threshold defaults to 10 units), out_of_stock (zero-inventory variants), by_location (per-warehouse breakdown with location name, total SKUs tracked, units on hand).

Get weekly revenue, orders, and items sold as a time series. Data is aggregated live from the Shopify Order table. Returns { rows, charts, presentation }. `rows` is the table — each row: weekOf, revenue (formatted), orders, itemsSold. `charts` is an array of three ready-to-render chart specs (revenue, orders, items — separate charts because units differ) in long/tidy format with metadata. DEFAULT

Get data from one connected third-party connector. Pass `connector` to pick which one and `report` to pick the report within it — valid `report` values differ per connector; this is the quick index, call get_help(<topic>) for the full picker + gotchas per connector. • connector:"triple-whale" — marketing attribution (channel mix, PPS, journey events). Reports: summary, attributions, by_channel, p

ESCAPE HATCH: ask Triple Whale's own AI assistant (Moby) a free-form question, answered by Triple Whale's servers against their full dataset. Use ONLY when no structured tool covers the question — prefer get_connector_data(connector:'triple-whale') / get_marketing_performance / get_spend_reconciliation first: they're faster (Moby takes 20-60 seconds), deterministic, and carry confidence/freshness

Get Shopify Markets data — geographic market definitions and per-market revenue performance. Reports: markets (list of configured markets with their assigned countries and currencies), performance (revenue, orders, AOV, and revenue share broken down by market for a date range). Use this to understand which geographic markets are configured and how revenue is distributed across them. Note: market d

Join Google Search Console queries with Shopify product sales — answers questions that pure SEO or pure sales tools can't. Reports: by_url (given a product URL, return its top driving organic queries plus the matching product's sales/orders/refunds and an implied click-to-purchase rate), unconverting_pages (URLs that received >= minClicks organic clicks but the matching product sold <= maxSales un

Get product review analytics (Yotpo, Reviews.io, or Judge.me). Reports: summary (average rating, distribution), trends (daily volume/score), top_rated (best reviewed products), lowest_rated (worst reviewed), recent (latest reviews), by_product (all reviews for a specific product — requires productId or productTitle), search (keyword search in reviews), timing (days from purchase to review), sales_

Get cart abandonment analytics. Reports: stats (trackedAbandonmentRecords, abandonedCheckouts, recoveredCheckouts, recoveryRate %, abandonedValue, lostRevenue, abandonedAgeDistribution — how long ago the still-unrecovered carts were abandoned: 0-24h, 24-48h, 2-7d, 7-30d, 30d+), top_products (most frequently abandoned products with count, quantity, and total value). IMPORTANT UNIVERSE: figures cove

Get Google Analytics (GA4) website traffic and e-commerce data. **DO NOT USE THIS FOR CHECKOUT FUNNEL OR CVR QUESTIONS** — call `get_marketing_performance` instead, which uses Shopify TrafficStat (the canonical session source) and Shopify orders. GA4 over-counts sessions (~2-3× higher than Shopify TrafficStat for many stores due to subdomain/bot inclusion) which produces a CVR roughly half the rea

Run a real GA4 funnel report — step-by-step conversion through a sequence of events, with optional breakdown and next-action analysis. This is a LIVE call to the GA4 Data API (not synced data), so it can answer 'where do users drop off between viewing a product and purchasing?' with GA4's own funnel engine. Each step is an event name (e.g. page_view, add_to_cart, begin_checkout, purchase) or a ful

Discovery tool: list the GA4 property's custom dimensions and custom metrics (with apiName, uiName, description), plus the standard dimension/metric API names available on the property. Call this to find valid field names before building a get_ga4_funnel report or interpreting custom events. LIVE GA4 Admin/Data API call.

Get the business improvement checklist (the FINDINGS to act on). If insights exist, returns the current checklist with status. If no insights exist (or refresh is requested), returns a data bundle with all key metrics from connected sources — use this data to generate actionable business recommendations (typically 3–8, as many as the data supports), then call save_insights to store them. You can a

TRIGGER: Whenever you surface a problem or opportunity from the data, OFFER to save it as an insight and confirm before writing — don't auto-save. Make the offer concrete and inline: 'Want me to save this as an insight to track?'. Findings vary in significance; the user decides what belongs on the checklist. Insights are OBSERVATIONS (findings to act on) — when a fix is actually SHIPPED, that's a

Update an insight (a FINDING) — set its status, or edit its content. Allowed status: open / addressed / superseded / dismissed. Note: you normally don't set 'addressed' by hand — that happens automatically when set_intervention links this finding to a shipped fix. An insight carries NO fix verdict; 'did it work' lives on the linked intervention (see get_interventions). Use this to dismiss/supersed

Record one or more structured metric snapshots for time-series analysis. Use this whenever you cite a metric in an insight, weekly digest, or benchmark — instead of (or in addition to) burying it in prose. Each snapshot is queryable later via query_metric_snapshots, e.g. 'show CVR for UK over the last 6 months'. Always pass canonical metric keys from the registry (call query_metric_snapshots with

Query recorded metric snapshots as a time series. Use to answer 'show me all CVR snapshots for UK over the last 6 months', 'how has ROAS evolved by month?', etc. Returns canonical snapshot data (matches the dashboard and monthly digest exactly) plus a `charts` array of ready-to-render line chart specs (one per metric; multi-series when the query spans multiple stores or segments). DEFAULT TO RENDE

TRIGGER: Call WITHOUT asking whenever the user applies or ships a fix. 'Fix' covers any shipped change — SEO/copy/ads/UX/ops, AND code/schema/connector changes. Distinct from save_insights (which captures observations) — this is for tracking actions: 'I changed X, hypothesised Y, will measure at dates [a,b,c]'. Auto-captures the most recent matching snapshots as baseline (for each {store, metric}

TRIGGER: Call WITHOUT asking once an applied intervention's check date has passed. Find candidates with get_interventions(dueForCheckIn: true). Attaches the latest RECORDED post-fix snapshot for each tracked metric (a snapshot whose periodStart is on/after the fix's appliedAt) as the 'post' side, computes deltas vs baseline, and flags whether each metric met its expected delta. It does NOT capture

TRIGGER: Call WITHOUT asking when an existing intervention's details need correcting — a check date that should move, a note to add, a hypothesis to sharpen, or a threshold watch to attach. Interventions could previously only be created and closed, so a wrong check date stayed wrong and surfaced as due on the wrong day; closing and re-creating is NOT a workaround, because it discards the captured

List interventions (the actions/fixes that were shipped) and report their measured outcomes. This is the SINGLE source of truth for 'did it work' — insights (findings) never carry a verdict. Use report: 'status' for a one-call picture of 'what fix-work is open' (due for check-in / in flight / recently closed + win rate — best for 'where did we get to?' / session start); report: 'list' (default) fo

Return all merchant-attached LLM context notes for a store (or all stores if omitted). Notes are authoritative caveats the merchant has added — e.g. 'one B2B customer skews retention', 'Q2 budget freeze, don't suggest more ad spend', 'pre-2026-03 data is partial'. The same notes are auto-injected into every store-scoped tool response as _storeNotes, so calling this directly is only needed when you

Merged, date-ordered timeline of everything that happened to the business: interventions (changes the merchant applied, and when their verdict landed) plus store-context notes (campaign launches, migrations, budget freezes, data quirks the merchant recorded). Use this to line dated events up against a metric movement — 'revenue dipped on the 12th, what changed around then?' — instead of calling ge

Save a new store-context note from something the USER told you that future sessions should know. Only call this when the user reveals business context the data alone wouldn't show (e.g. 'one B2B customer is a reseller', 'Q2 budget is fixed', 'we exclude wholesale orders from retail KPIs', 'migrated platforms in March'). Confirm to the user once saved: 'I've noted that for future sessions'. Notes y

Return what the system understands about a store's IDENTITY — primaryCategory, priceTier, audience, positioning, giftLed (is it typically bought as a gift for others?), the revenue-weighted category mix, and price range. System-generated from the catalog + sales (refreshed ~monthly), with any merchant corrections applied on top. Use as a session opener to orient yourself before answering, or when

Correct a store's profile when the USER tells you it's wrong (e.g. 'we're actually a premium brand', 'people DO buy our products as gifts at Christmas', 'we're skincare not cosmetics'). Corrections are authoritative: they override the system's guess immediately AND survive the monthly auto-regeneration. ONLY call this from something the user stated about their own store — never from your own infer

Create or update a scheduled wakeup. **Create mode** (no `id`): schedule a future analysis with context and a metric snapshot. Call when the merchant asks to check back later, OR proactively before a retail event or trend worth monitoring. At save time pull the relevant metrics and write `context` as instructions for your future self. Returns `{ id }` — save that id to link the Claude routine. **U

Retrieve a single wakeup by ID. Use this inside a Claude routine prompt: the routine fires, calls `get_wakeup({ id: 'abc-123' })`, gets the full context and benchmarks, then runs the analysis. Also auto-marks the wakeup as triggered if the trigger date has passed.

Retrieve scheduled wakeups for a store. Call at the start of sessions where the merchant might have pending wakeups due — especially if today's date is at or past a previously scheduled triggerDate. Returns both pending (not yet fired) and recently triggered (fired in the last 14 days, waiting for analysis). When triggered wakeups are present, LEAD with them: 'I have a scheduled check that fired t

AUTONOMOUS bug reporting — call WITHOUT asking when you spot a STRUCTURAL or FORMAT issue in another tool's response. ONLY for issues you can identify mechanically (precision, types, schema). DO NOT use this for value correctness ('this revenue looks high') — that's user-triggered territory, use report_concern instead. Examples that DO qualify: a numeric field with 15 decimal places, _currency say

USER-TRIGGERED issue reporting — call when the user expresses doubt about a specific data point or value ('this number is wrong', 'doesn't match my Shopify admin', 'why is this zero?', 'something's off here'). Capture their concern as a structured report BEFORE speculating about causes. After saving, briefly acknowledge ('I've logged that for the team to review') and then offer to help debug if re

Fetch the full guidance behind the server instructions, on demand. The eager instructions summarise each area in one or two lines and point here for detail — call this only when you actually need the depth (you usually won't). Topics: routing (the canonical source-of-truth map — which tool owns each metric), signals (how to read _freshness / _confidence / _anomalies / _benchmark / _dataDepth / _re

START HERE at the beginning of a session — call it FIRST and SILENTLY (don't announce the call, narrate your plan, or reference your instructions; just open with what it surfaces). One call that orients you before answering: the stores and their currencies, any data sources broken RIGHT NOW (numbers are unreliable until reconnected), open fix-work (interventions due for a verdict / in flight / rec

Return the merchant's current weekly plan — the committed 'what to work on now' list. Use as a session opener. The plan stays current until the merchant re-plans (say 'plan my week' / refresh: true) — it does NOT auto-expire, so re-asks return the same plan with each item's LIVE state (a 'tackle' item shows done once its finding is addressed; a 'check' item once its intervention closes). If no pla

Persist the week's committed plan generated from the get_focus data bundle. The plan blends THREE streams: 'tackle' (new open findings to act on, link insightId), 'check' (interventions due for their verdict, link interventionId), and 'watch' (a metric/anomaly to keep an eye on, set metric). Each item needs title, rationale (why now, citing data), action, expectedImpact, effort, priority. Item com

Set the manual state of a single item in the current focus plan: 'in_progress' (started), 'dropped' (decided not to do it this week), or 'planned' (clear the overlay). NOTE: you can NOT set an item to 'done' here — completion is DERIVED from the linked finding becoming addressed or the intervention closing. To complete a 'tackle' item, ship the fix and call set_intervention (which addresses the fi