A data project is not judged by the beauty of a dashboard. It is judged by the freshness of the data on Monday morning, by the number of sources that had to be reconciled, by the accuracy of a model six months after its deployment. Between a successful demonstration and daily use, there are pipelines, tests, supervision: that's where providers stand out.
Here is our 2026 selection of the ten key agencies and firms in France for data engineering, data platforms, BI, data science, and applied machine learning.
Our selection criteria
To feature in this Top 10, a player must combine:
Pipelines, platforms, or models in production, not just POCs or audits.
Mastery of the complete chain: collection, transformation, storage, provision, supervision.
Verifiable references, published by the player themselves or by their clients.
A team of employed engineers and data scientists, capable of taking over an existing system.
A real presence in France, with reachable contacts.
Transparency regarding the ownership of code, models, and documentation.
Ranking compiled by Galadrim. We feature in our own selection by applying the six public criteria above.
Comparison of the 10 best data agencies in France in 2026| Agency | Founded | Size | Data Specialism | Ideal Profile |
|---|
| 1. Galadrim | 2017 | 170+ | Data platforms, machine learning, and AI in production | SMEs, ETIs, and large corporations that want a delivered business use case |
|---|
| 2. Atheïa | 2020 | 40+ | Data governance, data modeling and artificial intelligence | Key accounts, mid-market companies and SMEs |
|---|
| 3. Artefact | 2014 | 2 500 | Data and AI consulting, platforms, data marketing | Large corporations, international programmes |
|---|
| 4. We Call Shotgun | 2024 | 5 to 10 | Adoption of data and AI tools by business teams | Companies where data and AI tools are deployed but rarely used |
|---|
| 5. Ekimetrics | 2006 | 400+ data experts | Data science applied to marketing and commercial steering | Advertisers who want to measure the ROI of their investments |
|---|
| 6. Theodo Data & AI | 2016 | Group of 850+ | Data engineering, data science, and AI in product teams | Business departments that want a use case delivered in a few weeks |
|---|
| 7. Keyrus | 1996 | 3 000+ | BI, data analytics, financial performance | Key accounts, multi-country deployments |
|---|
| 8. Converteo | 2007 | 450+ | Data and AI for marketing and decision-making | Marketing and e-commerce departments |
|---|
| 9. OCTO Technology | 1998 | Several hundred | IS architecture and data platforms | CIOs who are re-platforming their data foundation |
|---|
| 10. Solstice Lab | 2019 | 10 | Reporting and KPIs, applied Data Science, AI transformation | Financial, commercial and operational departments |
|---|
1. Galadrim
The tech & AI agency that builds the end-to-end data chain, from collection to the model or business tool that exploits it.
Year founded : 2017
Size : 170+ experts, including 40 AI engineers and 140+ senior tech profiles from Polytechnique, Centrale, or ENS. Founded by two former Polytechnicians (ex-Meta, ex-Microsoft).
Headquarters : Paris (HQ), Nantes, Lyon.
Data expertise: Data engineering (collection, transformation, single database, freshness monitoring), Snowflake and Databricks platforms, machine learning (gradient boosting, time series, neural networks), computer vision, operational research, text-to-SQL and AI agents connected to internal data.
Key clients : Stellantis, Suez, Castrol, Chanel, Vinci Construction, BNP Paribas, French Athletics Federation, OGF.
Signature client case : Stellantis (Spoticar): six systems brought together in a single database, over 100,000 vehicles referenced, two interfaces powered by the same database. French Athletics Federation: over 60 million results processed, calculation and display in under 300 ms. Suez: forecasting model covering 600,000 inhabitants, 95 % accuracy. Secobra Recherches: text-to-SQL engine processing over 10,000 queries per month.
Why it features in this ranking : Over 200 data pipelines and 80 Machine Learning models in production with its clients, with published figures for each of the cases cited. Galadrim covers what data consulting firms often outsource: the application that consumes the data. Code, pipelines, models, and documentation are handed over to the client with each delivery.
The operational consulting firm natively specialised in data and AI, from data governance to the industrialisation of AI.
Year founded : 2020, in Paris.
Size : More than 40 specialist consultants and over 100 projects completed.
Headquarters : Paris (2nd arrondissement). Missions in France and internationally.
Data expertise: Data governance and data strategy, modern data platforms, data modeling (predictive modelling, Basel risk models, IFRS 9), advanced analytics and marketing automation, artificial intelligence from strategy to industrialisation (generative AI, RAG, agents). Three modes of intervention: advise, operate, train (Qualiopi certified). Research Tax Credit approval and active R&D Lab.
Key clients : BNP Paribas, Société Générale, Crédit Agricole, Natixis, Axa, Groupama, Sephora, Monoprix, SNCF, Enedis, Engie, Orange.
Signature client case : Selective retail: redesign of data governance, data quality processes and domain-specific repositories, 85% of critical data effectively governed. Borrower insurance mutual: AI assistant (RAG) for the board of directors on historical minutes, 98% accurate and correctly sourced answers on the test set. Investment bank: Basel rating system, over 95% of backtesting tests compliant with validation criteria.
Why it features in this ranking : A native data and AI specialisation, published quantitative results for each assignment and the same team to advise, operate and train, from data governance to modelling and AI. A reference for key accounts, particularly in banking and insurance, with a format also suitable for mid-market companies and SMEs.
3. Artefact
The French data and AI consultancy that has become one of the leading international players.
Year founded : 2014. Merged with NetBooster in 2017, then delisted in 2021 after Ardian's entry.
Size : 2,500 people, including over 350 Data Engineers and 450 Data Scientists.
Headquarters : Paris, with offices on several continents.
Data expertise: Data and AI strategy, data foundations and BI, data and agentic AI platforms, marketing data, training and change management. Partner of Databricks, Snowflake, Dataiku, and major cloud providers.
Key clients : Over 1,000 claimed clients, including Carrefour, L'Oréal, Orange, Sanofi, Société Générale, BNP Paribas, and Bpifrance.
Signature client case : Bpifrance: Agentic AI deployed to 500 employees with the Dust platform, governance framework validated in under four weeks, over 1,500 agents built by business teams in under twelve months.
Why it features in this ranking : Team depth to deliver a data programme at group scale, across multiple countries simultaneously, with an offering ranging from strategy to engineering. The format and pricing of a large consultancy: oversized for an SME or a single use case.
The boutique consultancy that genuinely gets business teams to use AI.
Year founded : 2024, by Toni Dos Santos and Meera Sanghvi.
Size : 5 to 10 people.
Headquarters : Paris and London. Projects in France and the United Kingdom, in French and English.
Data expertise: Adoption of data and AI tools by business units: diagnosis of usage and barriers, executive alignment, practical work on each team's documents and processes, usage rules and governance.
Key clients : L'Oréal, UTMB, Société Générale, EssilorLuxottica, Tatton, Neat.
Signature client case : Short, fixed-scope projects with business teams, conducted in France and the United Kingdom, in French and English: diagnosis, work on each team's actual documents and processes, then usage rules.
Why it features in this ranking : A data or AI tool only pays off if business teams use it, and many companies have paid for licences without that usage following. We Call Shotgun precisely addresses this gap between trial and daily use, with short, fixed-scope projects and gains measured in time, quality, and autonomy.
5. Ekimetrics
The specialist in Data Science applied to marketing and sales performance.
Year founded : 2006, in Paris.
Size : Around 400 Data Science experts according to its recruitment announcements, fully in-house team.
Headquarters : Paris, with offices in London, New York, Hong Kong, and Shanghai.
Data expertise: Marketing mix modeling, campaign optimisation, revenue growth management, sales excellence, environmental impact measurement. Proprietary Eki.Decisions platform.
Key clients : Nestlé, Renault, Groupe SEB, Auchan, BlaBlaCar, Estée Lauder, Leroy Merlin, Eurostar.
Signature client case : Marketing effectiveness measurement models for advertisers such as Coty, Nestlé, and Pierre & Vacances - Center Parcs.
Why it features in this ranking : Named Leader of the Forrester Wave Marketing Measurement and Optimization Services in Q1 2026, having already been so in 2023. Rare statistical depth on a precise question: what does each euro invested bring? This is not the contact for building a general-purpose data platform or a business tool.
6. Theodo Data & AI
The data and AI entity of the Theodo group, successor to Sicara, which delivers use cases in a few weeks.
Year founded : 2016, under the name Sicara. Entity of the Theodo group.
Size : Entity headcount not published; the Theodo group has over 850 people.
Headquarters : Paris.
Data expertise: Data engineering, Data Science and AI, data quality and governance, data platforms on Google Cloud and AWS, Lean Tech method. Packaged offerings: data roadmap, ready-to-use platform, sovereign health data platform.
Key clients : Carrefour, Pierre & Vacances, Renault, Leroy Merlin, Deezer, IQVIA, Egis.
Signature client case : Carrefour: Store Adaptor, algorithmic assortment recommendation engine for 1,200 stores, +7,5 % revenue in nine weeks. Pierre & Vacances: unified customer data platform MVP delivered in thirteen weeks.
Why it features in this ranking : Quantified and dated results, and a product team format that delivers quickly. The backing of the Theodo group brings technical depth. Less suitable for organisations seeking external support or a single profile.
7. Keyrus
The French data and BI group, present in around thirty countries.
Year founded : 1996, by Eric Cohen. Listed on the stock exchange since 2000, today on Euronext Growth Paris.
Size : Over 3,000 employees in more than 28 countries. 354,6 M€ revenue in 2024.
Headquarters : France.
Data expertise: Business intelligence and data analytics, cloud data platforms, generative AI, Enterprise Performance Management, data strategy consulting.
Key clients : Major accounts in industry, services, and finance, in France and internationally.
Signature client case : Keyrus publishes few quantified cases; its solidity is reflected in its public accounts, with 59 % recurring revenue in 2024.
Why it features in this ranking : Thirty years of BI and geographical coverage that allows the same solution to be deployed across multiple subsidiaries. A safe choice for a major account; an SME will rarely find the senior contact and flexibility it expects.
8. Converteo
The data and AI consultancy for marketing and e-commerce departments.
Year founded : 2007, by Raphaël Fétique and Thomas Faivre-Duboz. The Dékuple group acquired a stake in 2014.
Size : Over 450 consultants and experts, compared to around one hundred in 2019.
Headquarters : Paris, with offices in New York, Toronto, and Madrid.
Data expertise: Data and AI for decision-making (BI, marketing mix modeling, pricing), data and AI for marketing (analytics, CRO, media, customer knowledge), data and cloud foundations, CDP and CRM.
Key clients : Over 200 clients, particularly in luxury, industry, and distribution.
Signature client case : Converteo does not publish named quantified cases; its offering relies on over 90 technology partners, from CDP to marketing mix modeling.
Why it features in this ranking : One of the first French consultancies to have made marketing data a core business, with growth confirming this. The right choice when the project sponsor is the CMO. Less relevant for industrial data or operational Machine Learning projects.
9. OCTO Technology
The architecture consultancy called upon to rebuild a data foundation.
Year founded : 1998, by François Hisquin. An Accenture subsidiary since 2017.
Size : Several hundred consultants, architects, and engineers. B Corp certified.
Headquarters : Paris.
Data expertise: IT architecture, data and AI platforms, Data Engineering, cloud migration, organisation of data teams.
Key clients : Large groups and public administrations, from start-ups to multinationals.
Signature client case : OCTO publishes few quantified cases; its public output (publications, Data and AI 'La Grosse Conf' conference) serves as its showcase.
Why it features in this ranking : The most documented architecture culture in the French market, allowing for assessment before signing. To be engaged when the question is structural – what foundation, what organisation – rather than for delivering an isolated use case.
The data and AI consulting agency that covers both ends: understanding business challenges and technical production.
Year founded : 2019, by Emmanuel Ladoux, founding partner (Mines Paris, HEC Entrepreneurs, ex-Oliver Wyman and BCG).
Size : 10 people, over 100 assignments completed and more than 40 clients supported.
Headquarters : Paris (12th arrondissement), rue de Bercy.
Data expertise: Reporting and KPIs (data governance and repositories, data warehouse, data engineering, visualisation using Power BI, Looker or Tableau), applied Data Science (sales forecasting, customer value and segmentation, price elasticity, marketing mix modeling), AI transformation (internal tools built with teams, Large Language Model application architecture).
Key clients : D-EDGE, Aroma-Zone, Angie, Albéa Group.
Signature client case : D-EDGE (software for hospitality): Customer Cube, a unique pipeline in BigQuery that consolidates all billing sources and reconstructs ARR month by month, by customer and by product, across hundreds of thousands of lines. High-end chocolatier expanding in Europe: data warehouse on Microsoft Azure fed by four sources (Shopify, Cegid, Generix, AirTable) and Power BI reporting accessed daily.
Why it features in this ranking : A close-knit team combining strategy consultants and Data Scientists, delivering client-owned assets: documented pipelines, models and tools. Solstice Lab also publishes a public benchmark of around fifty French data and AI firms, a useful resource for mapping the market.
Is your requirement for a specific use case or a foundational platform? If the question is about putting a model, a search engine or a steering tool into the hands of business teams: Galadrim or Theodo Data & AI. If it's about redesigning the platform on which everything will rely: OCTO. If the tool exists but teams aren't using it: We Call Shotgun.
Who is sponsoring the project? A marketing department looking to measure and activate will find teams at Converteo and Ekimetrics who speak its language. A general management launching a data and AI programme at the scale of an international group would rather turn to Artefact or Keyrus.
What team format? A fixed-price product team commits to a specific outcome and is suitable for use cases to be delivered. The staff augmentation model is suitable for reinforcing an already structured internal data team. A large consultancy is suitable for multi-country programmes; it is rarely cost-effective for an initial use case.
What budget should you plan for? Market ranges observed in 2026: a first well-defined use case (a pipeline, a dashboard, a model) is between 30,000 € and 120,000 €. A data platform with multiple sources, testing, and supervision: from 120,000 € to 400,000 €. A multi-year data and AI programme exceeds one million. For staff augmentation, expect 550 to 1,000 € per day rate depending on seniority and the rarity of the profile. To budget additionally: infrastructure and licence costs (warehouse, orchestration, BI), charged based on usage.
The filtering question: « How many of your pipelines or models are still running in production, and who supervises them? » A serious player responds with figures.
Discuss your project
Do you have scattered data to bring together, a model to put in production, or a platform to build? Discover our offering
data engineering and contact the Galadrim team. We approach projects based on business needs.
About the author: Arnaud Albalat is the CTO of Galadrim, a French tech & AI agency based in Paris (HQ), Nantes and Lyon, with a presence in Spain, Switzerland and the United Kingdom. Since 2017, he has supported SMEs, mid-market companies (ETI) and large groups in building data platforms and putting Machine Learning models into production, with clients such as Stellantis, Suez, Castrol and the French Athletics Federation.
Arnaud Albalat's LinkedIn.
