We partner with Dentressangle and its portfolio companies to embed AI where it moves the needle. At group level, we run strategic workshops with partners to prioritise opportunities; at company level, we design and deliver tailored products that improve efficiency, quality and growth.
Dentressangle backs ambitious European businesses across multiple industries. To accelerate value creation, the group wanted a systematic way to:
identify high-ROI AI use cases per participation,
align executives on risks, governance and data prerequisites, and
turn strategy into working software, quickly and safely.
Our model blends board-level guidance with hands-on delivery:
Strategy & prioritisation with partners: executive briefings, opportunity mapping, and a scored pipeline of AI initiatives.
Delivery with participations: small, outcome-focused squads (Product, Data/ML, Engineering) shipping increments every few weeks.
Operating model & governance: data access, security, evaluation metrics, and change management embedded from day one.
We build bespoke AI products against each company’s systems and workflows, favouring explainable methods, measurable KPIs, and straightforward handover. Two examples below.
Who.
AplusA is a healthcare market-research institute that runs multilingual questionnaires in successive waves. Before analysis, teams prepare data through verbatim coding and quality controls, and a new DataHub is being introduced to centralise these operations. The objective is to automate the most manual steps and encourage teams to adopt the DataHub instead of relying on spreadsheets.
What we built.
We developed two modules that integrate directly with the DataHub:
First, an AI coding engine learns and maintains a consistent codeframe from raw answers, consolidates duplicate codes and classifies new responses, remaining robust across iterative waves of collection.
Second, a QC rule generator reads the existing Word “programming tables,” converts them into a structured JSON specification and automatically produces Python checks, which run with sandboxed tests and an auto-correction loop until all tests pass.
How.
We combine retrieval with reranking to provide the right semantic context, and we use LLM prompting to support code discovery and consolidation. An API layer exposes clear interfaces to the DataHub, and delivery proceeds in focused sprints with explicit acceptance tests.
Why it matters.
This approach delivers faster and more consistent coding, ensures quality checks are reproducible and auditable, and establishes a modern workflow that nudges teams to work inside the DataHub rather than Excel or Word.
Who
Coverguard Safety designs and sells PPE across Europe, and the leadership wanted a structured path from AI awareness to a set of prioritised pilots focused on the digital customer experience and commercial efficiency.
Engagement format
We ran a three-part session that opened with an executive-level primer on AI (history, capabilities and current state of the art), followed by a short presentation of Galadrim’s most relevant achievements, and concluded with a collaborative workshop to surface and qualify company-specific use cases.
Outcome
We produced a documented plan with near-term pilots and KPIs. The first stream proposes an intelligent product chatbot and recommender for the e-commerce site. The second targets automation of recurring customer requests pricing, stock and product specifications across channels. The third designs a tender assistant that maps client Excel lines to catalogue products and drafts proposal responses. Mid-term items include multilingual content workflows and a field-prescription assistant to support the sales team.
Next steps
We recommended confirming priorities, finalising scope and budget, and launching pilots on an iterative delivery cadence with clear success metrics and stakeholder checkpoints.
