Selling Slow Movers with Gen-AI Powered decisions

Amorim Cork Solutions (ACS) operates across 8 industrial units and 14 business areas, delivering products engineered built for requirement, with each solution tailored to customer-specific technical characteristics that must be met (e.g. tensile strength, thickness, granularity).

Over time, the company accumulated a significant portfolio of slow-moving stock (more than a year-old product in stock). Matching existing stock to potentially new commercial opportunities was labor-intensive and technically challenging, as each product has its own technical properties, each use case and client has its own requisites and know-how is noteasily accessible. The key strategic question: Can we intelligently identify existing slow movers that meet the customer needs and the specific technical requirements?

Together, ACS and LTPlabs built SLATE, a GenAI platform, connected to purpose-built AI algorithms and a web application that matches client needs with slow movers. SLATE is both a new process and a new platform, combining traditional machine learning, similarity modeling, and generative AI into an intuitive web experience.

What SLATE does?

• Finds technical affinities between slow-moving stock and new orders by analyzing product specifications, BOMs, application requirements, and historical interactions.
• Ranks and recommends stock alternatives based on technical similarity, business rules, and operational feasibility.
• Provides a conversational interface that allows users to query stock naturally and obtain precise, explainable recommendations.
• Ensures consistent, reliable answers by orchestrating different specialized GenAI tools, avoiding hallucinations and ensuring predictable system behavior.

Behind the scenes, SLATE blends structured data, domain knowledge, prior user feedback, and a suite of AI models to generate relevant, actionable results.

Users interact with SLATE through a simple chat-based interface:

1. A prompt is interpreted (e.g., “Find similar impaired stock to product A for Home & Office”).
2. SLATE selects the right AI tool (e.g. A tool to find property similarity, to search within the product catalog, an algorithm to rank Slow mover affinity, or to search companies database for details).
3. Technical and commercial matches are computed.
4. The system returns a ranked list of viable slow-mover alternatives, enriched with details and ready for decision-making.

What once required technical know-how and manual cross-checking across SAP, Excel files, and technical sources is now achieved instantly. Then, the decision maker evaluates the suggestions and proposes to the client, if suitable.

SLATE was deployed through a collaborative pilot model that already delivered measurable strategic gains:

• Impaired and slow-moving items are now actively considered as first candidates for new orders, if they fit the technical specifications.
• Every order received within the customer service department is automatically analyzed by the AI algorithm.
• What used to require manual search and cross-checking is now executed automatically, with consistent logic.
• SLATE augments the decision-maker: it recommends; the human validates and decides.
• The architecture, data pipeline, and adoption model pave the way for broader AI applications at ACS.

‍

SLATE demonstrates how artificial intelligence can unlock hidden value, streamline complexity, and elevate decision quality across the organization. By transforming slow-movers into actionable opportunities, the company now operates with greater agility, better information, and a scalable platform for continued innovation.