The 9Q Grid
← Case Library
← Back to the Case Library
Unilever
Q6Consumer Goods · Shape forming
Researched as of April 2026
Shape Forming — Spiral proposed, not yet confirmed
Downgrading from the on-record Spiral: breadth is 5 cells with a dominant share of only 0.321 and structuralSuggestion itself reads Shield, which matches the narrative's own description of a broad multi-cell portfolio spanning Q3 manufacturing/HR automation, Q6 supply-chain optimisation (Project Sky, Aera, smart freezers), Q4 marketing/R&D acceleration, and emerging Q5 R&D co-discovery. There is real evidence of expansion over time (Project Sky 'launched 2022, expanding globally', DataLab's 20-fold speed increase, 23,000 employees AI-trained by end-2024), but this reads as broad institutional rollout rather than a demonstrated compounding loop, so it is diffuse Spiral-adjacent evidence rather than a clean trend. This is a genuine toss-up between Shield and Spiral.
AdaptAugmentAssist
Q3Automator
Q6Optimiser
Q9Conductor
Q2Tactician
Q5Partner
Q8Explorer
Q1Operator
Q4Weaver
Q7Navigator
StructuredDynamicComplex
5/9 cells occupied
Shape Read
Unilever operates as a genuinely multi-cell AI portfolio split between a back-stage optimisation spine and a front-stage creative-acceleration layer. Its strongest and most economically consequential deployments sit in Q6 Optimiser — Project Sky's Walmart Mexico forecasting engine, the Aera-powered autonomous supply chain, the 100,000-unit smart freezer cold chain, and satellite-based deforestation monitoring — systems that continuously adapt over large, live data flows with minimal per-decision human review. A parallel Q3 Automator layer handles structured, delegated execution in manufacturing and recruitment, while the company's most publicly visible AI investment, in Q4 Weaver, accelerates marketing content and R&D briefing without displacing human authorship. The most analytically interesting position is an emerging Q5 Partner loop in R&D, where scientists and Microsoft's Azure Quantum Elements co-discover new formulations through iterative simulation. Unilever's declared five-year Google Cloud partnership signals Q7 ambition for agentic commerce, but as of the assessment date this remains intent rather than deployed evidence, and the company's mature Responsible AI governance infrastructure functions as a ceiling-manager rather than a marker of higher autonomy.
Trajectory: Unilever's most consequential near-term move is operationalising Google Cloud agentic commerce workflows in 2026-2027, the first test of whether its Q7 ambitions convert into deployed, evidenced systems.
Inside the company · not a cell
Nothing quoted yet. A rate or a saving will sit here when it is on the record. It does not light the grid.
Initiative Map
| Cell | Initiative | Confidence | Evidence |
|---|---|---|---|
| Q3Automator | Safety Suite — AI workplace safety system (Aguaí, Brazil)Manufacturing / workplace safety | ✅ | Safety Suite integrates AI and image recognition at the Aguaí Personal Care factory to detect, store, and alert teams about potentially unsafe behaviours, delivering a 12% improvement in identifying potential on-site risks since introduction.As of April 2026 |
| Q6Optimiser | Project Sky — Walmart Mexico CPFR neural networkDemand forecasting / replenishment | ✅ | Neural network ingests Walmart Mexico data by SKU, store, and day, accumulating five years of historical data and accepting 3–4 months of promotional plans from Unilever. Generates 3.1 million forecast combinations daily and makes 12.5 billion computations per day, triggering replenishment of 20 million cases of Unilever products across Mexico. Results: 98% fill rates, 98% on-shelf availability, 12% sales growth in under one year, covering 30% of Unilever Mexico's business (Knorr, Hellmann's). Launched 2022, expanding globally.As of April 2026 |
| Q6Optimiser | Aera Decision Intelligence — autonomous supply chain decisionsSupply chain planning / logistics | ✅ | Decision Intelligence platform partner for autonomous supply chain decisions. Explicit ambition stated by Wendy Herrick (Head of Customer Operations, North America) to build an end-to-end, self-driving supply chain, with human-on-the-loop governance rather than per-decision review; success is measured as decision velocity. Project Sky is one programme within the Aera partnership, and the Digital Materials Planner is a separate AI agent delivering improvements in planning precision.As of April 2026 |
| Q6Optimiser | Weather-AI ice cream demand forecasting (Sweden pilot)Demand forecasting / ice cream cold chain | ✅ | AI analyses weather data alongside historical sales to forecast ice cream demand volumes, delivering a 10% improvement in forecast accuracy in the Sweden pilot. The system continuously incorporates real-time weather signals and historical data to optimise production schedules and distribution planning, and is operational across markets with ongoing refinement.As of April 2026 |
| Q6Optimiser | Smart freezer cold-chain AI — inventory and replenishmentIce cream cold chain / last-mile | ✅ | AI-enabled image capture deployed in 100,000 of 3 million global freezer cabinets provides real-time inventory data by SKU and automatically triggers replenishment orders — 75,000 daily orders processed. Sales impact: 8% Turkey, 12% US, 30% Denmark. Two operating layers: dynamic inventory sensing adapting to real-time stock data (Q6) and structured replenishment trigger once a threshold is crossed (Q3).As of April 2026 |
| Q3Automator | eB2B distributor automated resupply systemEmerging market distributive trade | ✅ | When an order is received on the eB2B platform, the distribution management system checks current stock levels against quantity ordered and generates automated resupply if any shortfall exists; it also manages settlements, credits, and returns, with a built-in route optimisation tool for last-mile delivery. The platform is live in Indonesia, Vietnam, Pakistan, Philippines, and Thailand (with Bangladesh in roll-out), processing 75,000 orders/day and supporting €2.5 billion in annualised sales.As of April 2026 |
| Q6Optimiser | Manufacturing digital twin — Hefei, China factory (Dove)Manufacturing optimisation | ✅ | AI and digital twins run process optimisation, predictive maintenance, and 3D printing for parts at the Hefei factory (Beauty & Wellbeing, Personal Care, Home Care). Outcomes: 8% increase in overall equipment effectiveness, 15% reduction in batch cycle time, up to 20% reduction in wastage. The digital twin continuously simulates and optimises factory operations in real time.As of April 2026 |
| Q6Optimiser | Manufacturing digital twin — Brazil facilityManufacturing optimisation | 🔶 | A Brazilian manufacturing facility's digital twin generated $2.8 million in cost savings while improving productivity by 3%, using real-time simulation of factory operations.As of April 2026 |
| Q3Automator | AI process control, predictive maintenance, and yield optimisation (general manufacturing fleet)Manufacturing / factory operations | ✅ | AI systems power autonomous process control, predictive maintenance to reduce downtime, and intelligent process control to increase yield and reduce waste across Unilever's manufacturing network, delivering up to 10% savings on high-value raw materials such as cocoa and vanilla through AI-driven process monitoring. The Unilever Manufacturing System provides a common OPEX framework.As of April 2026 |
| Q5Partner | DataLab + Microsoft Azure Quantum Elements (molecular simulation)R&D / ingredient discovery | ✅ | Unilever's DataLab (virtual digital laboratory) harnesses Azure Quantum Elements, Azure HPC, and Copilot in Azure for natural language querying of scientific information, performing thousands of computational simulations (density functional theory) in the time previously requiring tens of lab experiments — a 20-fold speed increase over the PySCF standard. It enables search across tens of thousands of ingredient candidates or molecular combinations, with confirmed applications in hair fibre bond restoration (Dove, TRESemmé), skin microbiome research, and carbon footprint reduction, over 2.5 years of partnership with Microsoft.As of April 2026 |
| Q5Partner | AI in-silico product formulation (Knorr Zero Salt Cube, Hellmann's Vegan Mayo)R&D / food product development | ✅ | AI analyses millions of flavour combinations to create the Knorr Zero Salt Cube (zero sodium, maintained taste) and predicts taste, texture, and stability for Hellmann's Vegan Mayonnaise without physical trials, establishing operational in-silico testing. Human scientists direct formulation goals and validate outputs, in a process described by Dr Samantha Samaras (Global VP Science & Technology BPC R&D) as enabling connections the human brain would never have been able to achieve.As of April 2026 |
| Q4Weaver | Odour Adapt technology (whole-body deodorant)R&D / Personal Care product innovation | 🔶 | Complex algorithms analyse sweat and bacteria to develop customised deodorant formulations for Unilever's whole-body deodorant range.As of April 2026 |
| Q4Weaver | DelphiAI — product innovation recommendation engine (PA Consulting / Azure)R&D / product innovation | 🔶 | Azure-hosted system draws on consumer behaviour data, market research, and ingredient/formulation expertise to recommend optimal ingredient combinations for product briefs. Launched in laundry (India) and beauty (US) in 2023, expanded to Health & Wellness (US) later in 2023, with nine business teams on-boarded by 2024. Future versions will predict commercial performance of launches.As of April 2026 |
| Q4Weaver | Digital product twins — NVIDIA Omniverse / Collective WorldMarketing content creation | ✅ | Pixel-perfect 3D digital twins of physical products are created using the NVIDIA Omniverse platform in Open USD format, with each twin containing all variants, labels, packaging, and language formats in a single file, integrated into content workflows from TV to digital commerce and removing production duplication from five steps to one. TRESemmé Thailand achieved 87% cost reduction, 2x faster content, and a 5% purchase-intent lift; Beauty & Wellbeing overall achieved up to 55% savings, 65% faster turnaround, doubled click-through rate, and three times longer attention hold.As of April 2026 |
| Q4Weaver | Smart Briefing pilot (Closeup marketing teams)Marketing productivity | 🔶 | AI-generated content enriches high-level marketing objectives submitted by the team, delivering a 14% improvement in brief quality, a 26% satisfaction boost, and up to 58% time savings, at pilot stage.As of April 2026 |
| Q4Weaver | Dove 'Change the Compliment' AI campaign analysisMarketing / consumer insights | ✅ | AI rapidly analyses consumer responses to creative messaging and identifies the highest-resonance content within the Dove campaign, which was activated across 25 markets with 100+ content pieces in six months from insight to execution, generating 700 million impressions in under 30 days with 94% positive sentiment.As of April 2026 |
| Q4Weaver | Accenture GenWizard platform — GenAI productivity (multi-year programme)Enterprise productivity / business processes | ✅ | A multi-year expanded strategic partnership (September 2024) deploys Accenture's GenWizard platform (350+ patents) to scale gen-AI powered productivity across Unilever's business operations globally, building on Horizon3 Labs concept-testing work. CEO Hein Schumacher named 500 AI applications in production and expressed intent to go deeper.As of April 2026 |
| Q5Partner | Beauty & Wellbeing AI consumer diagnostics (skin/hair tools)Consumer personalisation | 🔶 | AI-powered interactive digital diagnostic tools are deployed across Beauty & Wellbeing brands for skin analysis, hair diagnostics, and personalised product recommendations, with IBM Watson used for nutrition personalisation based on individual health data and dietary restrictions. Consumers provide iterative data input and AI generates evolving personalised recommendations, constituting a nascent feedback loop that meets early Q5 criteria, though scale and depth of iteration are not publicly verified.As of April 2026 |
| Q4Weaver | eB2B AI assortment recommendations (Asia emerging markets)Retail / trade execution | ✅ | Advanced AI modelling generates personalised assortment recommendations for sales reps during 15–20 minute retailer visits, modelling what a given retailer is likely to buy based on historical data and store profile. Sales reps use the recommendations to prioritise product focus, loyalty programmes, and promotion plans, while AI and image processing analyse in-store photos for stock-gap identification, across 500,000 retailers and 6,000+ sales reps in 5–6 markets.As of April 2026 |
| Q4Weaver | Latin America AI merchandising prioritisation (3,600 merchandisers)Retail / trade execution | ✅ | AI tools provide 3,600 merchandisers across 6,000 stores in Brazil with personalised prioritised action lists, replacing undifferentiated opportunity lists, using image recognition and sales data to proactively identify empty shelves and visibility gaps.As of April 2026 |
| Q6Optimiser | AI customer operations — Amazon, Walmart, Tesco data partnershipsCustomer operations / supply chain collaboration | 🔶 | The Customer Operations team uses AI integration and data-sharing partnerships with Amazon, Walmart, and Tesco for seamless point-of-sale access across planning, logistics, and customer service through seven global hubs, delivering €1.7 billion of value over 2.5 years; embracing agentic AI was explicitly noted as a highlight.As of April 2026 |
| Q3Automator | AI recruitment screening (Pymetrics, HireVue, Amberjack)HR / talent acquisition | ✅ | A multi-stage AI recruitment pipeline processes approximately 1.8 million annual applications: Pymetrics uses neuroscience games to assess cognitive and emotional attributes and matches candidates against successful employee profiles via machine learning; HireVue analyses video interviews for speech, facial expressions, and nonverbal cues; Amberjack provides scalable video interview scheduling. Results include a 75–90% reduction in hiring cycle time, over £1 million in annual savings, 50,000 hours recovered, a 16% improvement in workforce diversity, and an 82% offer acceptance rate.As of April 2026 |
| Q3Automator | Unabot — employee HR/IT chatbotHR / employee services | 🔶 | An AI chatbot serves as the front face for all employee questions — HR policy, IT systems, allowances, facilities, shuttle buses, and salary reviews — described in 2019 by Unilever HR leadership, though current operational status has not been reconfirmed in 2024–2025 sources.As of April 2026 |
| Q4Weaver | AI training programme — 23,000 employees trained in AI by end-2024HR / workforce development | ✅ | A systematic AI literacy programme spans the Unilever workforce; by end-2024, 23,000 employees had been trained in AI usage, and in Personal Care over 75% of office-based employees became regular users of personal AI productivity tools. The AI development programme includes function-tailored hands-on training, workshops, drop-in clinics, and AI champions, with the DigiOps programme specifically upskilling supply chain teams.As of April 2026 |
| Q6Optimiser | NDPE deforestation monitoring dashboard (satellite + AI)Sustainability / supply chain transparency | ✅ | The bespoke NDPE (No Deforestation, No Peat Conversion, No Exploitation) dashboard monitors agriculture, forests, and peatland across five forest-risk commodities (palm oil, paper, soy, tea, cocoa), continuously updated with fresh Google Earth satellite images. AI applied to satellite imaging detects tree-cover changes and generates deforestation alerts, which are channelled into the dashboard alongside supplier compliance information and supply chain links.As of April 2026 |
| Q7Navigator | Google Cloud 5-year partnership — agentic commerce and AI-first enterprise backboneAgentic commerce / brand discovery / enterprise data | ◻️ | A five-year strategic partnership announced February 2026 will see Unilever migrate key enterprise applications and data to Google Cloud, deploy Vertex AI and Gemini models for brand discovery, marketing measurement, and supply chain analytics, and develop agentic workflows for complex multi-step tasks across business functions to build an AI-first digital backbone targeted at conversational and agentic commerce. Tara Brady (Google Cloud EMEA President) described deploying models that reason, learn, and act; Willem Uijen (Chief Supply Chain and Operations Officer) said technology has moved to the core of value creation at Unilever. No specific production agentic system has been confirmed.As of April 2026 |
| Q5Partner | Quantum computing integration with Azure Quantum ElementsR&D / computational science | ◻️ | Unilever has begun experimenting with quantum computing through Microsoft Azure Quantum Elements. Alberto Prado (Global Head of R&D Digital & Partnerships) described it as a gamechanger in terms of speed and the amount of data it could allow the company to process; the initiative is not yet in production but active experimentation is confirmed.As of April 2026 |
| Q5Partner | DelphiAI commercial performance prediction (future capability)R&D / product innovation | ◻️ | Future versions of DelphiAI are announced to include prediction of the commercial performance of product launches, which would close the innovation loop from Q4 (recommendation) to Q5 (prediction-informed co-creation). This capability is declared by the PA Consulting case study but not yet deployed.As of April 2026 |
Strength
A deeply evidenced Q6 Optimiser spine in supply chain, cold-chain replenishment, and manufacturing that keeps a €60 billion global portfolio reliably on shelf.
Fragility
Its most publicised AI story is Q4 marketing acceleration, which is real but risks being read as a bigger operating-model transformation than the evidence supports; the eB2B and Aera systems also create supplier-retailer dependency concentration risk.
Failure mode
The Google Cloud agentic commerce partnership could remain declared intent rather than convert into production evidence, leaving Unilever's Q7 ambitions unrealised through 2027.
Governance ceiling
The Responsible AI framework and Holistic AI assurance process (150+ projects assured, EU AI Act preparation) function as mature ceiling-management infrastructure that keeps the company within its Q3–Q6 operating range.