Seasonal Intelligence Platform

Every season. Every retailer. Seventeen years.

Plan seasonal campaigns with 17 years of retail intelligence. Ask questions in plain English, compare retailers across events, and make decisions with data no competitor can replicate.

0
Seasonal images from a 1.2 million image archive, captured weekly since 2009
0
Events
0
Years
0
Retailers

Used by category teams at a FTSE 100 retailer

The Return

Do the maths.

Select a seasonal event. Adjust the improvement. See what better seasonal intelligence is worth to each retailer. Built on publicly available grocery market data.

value created
Improvement 1.0%
0.1% 5.0%

One event pays for the platform many times over. Now multiply it across 57. Request the full ROI model

Ask Delphi

Ask anything. Get answers grounded in evidence.

Every query draws on 348,000 seasonal images, 17 years of data, and promotional observations across 357 retailers. These are real answers from the platform.

Try asking

How do premium retailers differentiate Valentine’s displays from discounters?
What promotional mechanics dominate Easter across the multiples?
When do retailers typically launch Back to School ranging?
How has Christmas gifting space allocation evolved since 2019?
Which retailers invest earliest in Halloween decorations?

Built For

Two audiences. One platform.

CPG Brands

  • Plan seasonal campaigns using 17 years of competitor evidence
  • Benchmark your shelf presence against category leaders
  • Build launch calendars from observed retailer behaviour
  • Track price positioning across retailers and events

Retailers

  • Optimise seasonal category allocation with historical data
  • Time seasonal launches based on competitive intelligence
  • Benchmark your seasonal execution against the market
  • Understand promotional positioning relative to competitors

Inside the Platform

See what you get.

The Platform

Three pillars. One archive.

Everything in Delphi draws on the same dataset, the same models, and the same continuously growing archive. The detail emerges when you need it.

Intelligence

Ask questions in plain English, track trends, and generate briefings from 17 years of seasonal retail data.

  • Natural language queries with AI-generated answers
  • Weekly rollout tracking across every retailer
  • Longitudinal trend analysis over the full archive

Comparison

Compare any retailer against any other, across any event and any year. Browse the photographic evidence behind every data point.

  • Side-by-side retailer comparison for any event and year
  • Retailer-by-event intensity matrices
  • Filterable image gallery with channel breakdowns

Planning

Build launch calendars from 17 years of observed reality. Know when competitors move before they move.

  • Data-backed trade plans from historical patterns
  • Promotional mechanics: was/now, multi-buy, percentage off
  • Price positioning and depth analysis across retailers

The Intelligence Behind Ask Delphi

Trained by an expert. Not scraped from the internet.

Ask Delphi is powered by Cartwright, an AI trained on the expertise of Steve Dresser — twenty years of retail consultancy, thousands of store visits, and direct relationships with CEOs and executive teams at retailers and CPG brands across the world.

Cartwright draws on two decades of presentations, client briefings, market analysis, and in-store observations. When you ask Delphi a question, you are drawing on genuine retail expertise grounded in real-world evidence — not a generic language model guessing from public data. Client data is never used to train shared models. Your competitive intelligence remains yours.

Domain expertise, not general knowledge

Twenty years of trade presentations, commercial strategy, and direct retail relationships

Grounded in evidence

Every answer draws on 348,000 images and 17 years of observed retail behaviour

Continuously learning

New imagery, new observations, and new market context feed the models every week

How It Works

Every price. Every promotion. Captured at shelf level.

Six connected AI models detect products, read labels, and match observations to a database of 54,851 known products. A normalisation layer validates every extraction against 6,563 known brands, filters price outliers, and merges retailer sub-brands into canonical names.

87%
Extraction Yield
54,851
Products Matched
6,563
Verified Brands
6
Connected Models
Detect
Finds every shipper display and promotional unit in the image
Crop
Individual products and shelf-edge labels are isolated
Classify
Identifies the product category from 21 retail categories
OCR
Reads prices, promotions, and brand names from labels
Match
Validates against 54,851 known products and 6,563 brands
Intelligence
Trends, positioning, and promotional patterns emerge
20+
Years of in-store retail consultancy across UK grocery, general merchandise, and clothing
5
Masters programmes supported at King's College London with academic research partnerships
1,000+
Store visits per year, every year, since 2009. Purpose-captured imagery from real retail environments

Every partner interaction improves the models for all partners. The platform compounds.

The Moat

You cannot buy time.

Delphi is built on 1.2 million proprietary images captured weekly since 2009. Not scraped. Not licensed. Purpose-captured by a team that has been in UK retail stores every single week for seventeen years. The seasonal archive of 348,000 images is a curated subset — the full dataset covers weekly trading activity across every category. The 2009–2020 data no longer exists anywhere else. Every week that passes, the archive grows and the moat widens.

All imagery is GDPR-compliant with automated face detection and blurring. Human-reviewed quality assurance at every stage. Zero automated deletions.

Work in progress!
2009 2026

Get Started

See Delphi in action.

We are onboarding a small number of CPG brands and retailers as launch partners. Early partners receive priority access, dedicated onboarding, and direct influence over what ships next. Pricing scales with seasonal grocery revenue.

Or email me directly at steve@kanops.ai