AI-powered e-commerce analytics for every brand,
not just those who can afford a data analyst.
February 2026
Enterprise brands have data teams. Everyone else has spreadsheets and gut feel.
Shopify's built-in reports are disconnected. Critical insights require combining multiple reports manually. Hours of work for basic answers.
Hiring a data analyst: £35-50k/year. Analytics platforms with AI: $399+/month. Most brands under £500k revenue can't justify either.
AI analytics tools exist — but they're locked behind enterprise tiers. The businesses that would benefit most from automated insights can't access them.
A data analyst in every dashboard. Automated insights daily. Interactive AI when you need deeper answers. Starting from day one, on every tier.
Every morning, AI analyses each client's data overnight and surfaces what matters: anomalies, trends, opportunities. No tokens used — it's part of the platform. Every tier gets this.
Need a deeper answer? Ask Lucia. "Why did AOV drop last week?" "Which campaign is cannibalising organic?" Tokens scale by tier, with top-ups available anytime.
Interactive, real-time, and designed to answer questions at a glance.
Recharts-powered visualisations with hover tooltips, zoom, drill-down. Previous period shown as dotted overlays on area charts, ghost bars on bar charts.
Shopify data syncs through BullMQ job queues into Supabase. Live updates, not daily CSV imports. Multiple refreshes per day.
Works on desktop, tablet, and mobile. Tailwind + shadcn/ui component library — consistent, polished, fast.
Not just charts — structured reports that answer real business questions.
See the complete customer journey: first touch → intermediate steps → conversion. Which channels start journeys? Which close them? Where do people drop off?
Track customer groups over time. Which acquisition month produces the most loyal buyers? Where does retention drop off?
December cohort showing strongest retention — BFCM buyers coming back.
Predictive models with confidence bands. See where revenue is heading — not just where it's been.
Visit → Add to Cart → Checkout → Purchase. See exactly where customers drop off and by how much.
702 lines of interactive charting. Area charts with dotted previous-period overlays. Bar charts with ghost comparisons.
Three domain experts that analyse data and give structured, actionable output. Not a chatbot — a thinking engine.
Thinks like a fractional CFO
Thinks like a performance marketing director
Thinks like a causal analyst
Every response: a clear headline insight, supporting evidence from the data, business context, and a specific next action. No waffle. No ambiguity.
The brands that need AI most — the ones without data teams — should get it first. Not last.
| Starter | ✓ Daily insights + 50 prompts/mo |
| Pro | ✓ Daily insights + 200 prompts/mo |
| Enterprise | ✓ Daily insights + 1,000 prompts/mo |
| Any tier | + Buy more tokens anytime |
Daily Insights (Free)
AI runs overnight analysis for every client. Surfaces anomalies, opportunities, and trends. Delivered to their dashboard each morning. Zero cost to the user — it's the product.
Interactive Prompts (Tokens)
"Why did our AOV drop this week?" "Which campaign drives the highest LTV customers?" Deeper questions use tokens from their allowance.
The Economics
AI models get cheaper every quarter. Today: ~£7/month for daily insights across 10 clients. Tomorrow: half that. Margins improve automatically.
From Shopify webhook to dashboard widget in seconds. Not daily CSV imports — live data.
Order placed, product updated, customer created
Event hits our API → queued for processing
Data cleaned, normalised, enriched with UTM data
Stored with org_id, RLS-protected, indexed
Client sees new data on next page load
One database, unlimited brands. Every row tagged with organization_id. Row Level Security means Company A physically cannot query Company B's data.
Create org → OAuth connect Shopify → BullMQ pulls 24 months history → Lucia auto-configures → Client gets invite link.
Total time: minutes, not days.
10 companies, 30 users. What it costs to run.
~$13 per company
~$4.30 per user
Going from 10 to 50 companies adds minimal cost. Supabase stays at $25. Railway adds ~$15. AI scales linearly but gets cheaper every quarter.
50 companies ≈ $250/month
$5 per company
At $29/company (matching current Essential pricing):
10 companies: $290 revenue / $128 cost
50 companies: $1,450 revenue / $250 cost
100 companies: $2,900 revenue / $400 cost
Profitable from client #5.
AI API costs based on Claude Sonnet. Costs decrease as models improve — Sonnet 4.7 expected to deliver Opus-level quality at Sonnet pricing.
Built, tested, deployed. Working software — not wireframes or roadmap items.
The architecture is in place. These aren't rebuilds — they're wiring jobs.
The queue infrastructure (BullMQ + Redis) means adding a new data source is a pipeline task, not an architecture change. Each integration is a worker that feeds into the same schema.
Cron job → Lucia analyses each org's data → delivers morning summary. Architecture exists — needs scheduling layer.
Real-time order/product/customer events. Endpoint ready — needs Shopify app approval flow.
Lightweight JS snippet captures UTM on first visit → cookie → attaches to order. Designed, ready to build.
Track AI prompt usage per org. Enforce tier limits. Enable top-up purchases.
Campaign performance → attribution pipeline. Meta Marketing API, read access.
Search + Shopping campaign data → unified attribution.
Self-serve signup → plan selection → payment → auto-provisioned org. Schema already has Stripe fields.
Per-org branding: logo, colours, custom domain. Architecture supports it — needs UI.
AI-powered e-commerce analytics that works for a £29/month brand as well as a £399/month enterprise. Real-time data. Interactive visualisations. Three AI specialists from day one.
Everything in this presentation is deployed and running. Not planned. Not wireframed. Built.