Case studies
Enterprise AI and technology leadership, with the evidence laid out.
From customer problem and technical validation to production system, measurable outcome and reusable capability. Skim the list, then open the ones that matter to you.
Every case in one row
- Jaden Data Bootstrapped to €1M+, team 0→10, two certifications €1M+ annual revenue, bootstrapped, no external capital MeasuredProduction Own company
- APG (€500B+ pension fund) Member-facing assistant, on-premises, ministry-recognised 4 languages at launch — Dutch, Polish, Turkish and English MeasuredProduction Client
- A German development bank and a humanitarian NGO it funds Seven document-heavy workflows, one spec, one evaluation template ~1 day saved per ministry monitoring report (out of ~90 per year) Customer estimateProof of concept Client
- Bosch (multinational automotive and industrial group) Hackathons on our platform, prompt engineering coached across Bosch teams 2 hackathons run for Bosch employees on our own platform MeasuredClient
- An enterprise software vendor SaaS-to-on-premises migration and AI skills across teams €190,000+ to be saved in SaaS costs alone by moving from SaaS to on-premises ProjectedIndependent
- A global automotive group Four tendering and contracting workflows in production 4 workflows taken from proof of concept to a production MVP MeasuredProduction Client
- anita.legal Security, audit readiness and AI-accelerated development Covered the engagement, from the long-term AWS savings alone QualitativeIndependent
- entAIngine Multi-tenant AI platform for 50+ organisations 50+ organisations running AI processes on one multi-tenant platform MeasuredProduction Own company
- RP Group (emergency lighting) Matching customer inquiries to catalogue products 60% less time to prepare an offer, as reported by the customer Customer estimateProduction Client
- Red Bull Startup screening agents for a venture team Thousands of startups analysed for the venture team QualitativeProduction Client
- Taufolio AI equity research from filings and calls ~3 min for a full report; about twenty seconds for a snapshot MeasuredProduction Own company
- idealo · ING (through SAS Institute) Six-country login rollout, real-time bank pipelines 1.4M accounts in three months, 700% over forecast, as key contributor MeasuredProduction Employment
- One of the world's largest pharmaceutical companies Employee chatbot and graph-based research platform Client
- Stellantis Foundation-model evaluation; global challenge won with Patronus AI Winner of the Foundation Model Evaluator challenge, entered with Patronus AI QualitativeProof of concept Client
- TÜV Regulation-change analysis across certifications Client
- Norges Bank, a German development bank and EU-funded consortia Cross-ledger bridge and sovereign data platforms 4 systems delivered: a bridge, two platform MVPs and an open-source marketplace MeasuredMVP Client
- IMF AI processes for writing documents Client
- Kenstone Image description and report extraction for valuers Client
Enterprise AI deployments
-
Pensions · 2023 – 2024
A pension-letter assistant: one-day prototype, on-premises production in twelve sprints
APG (€500B+ pension fund)
A customer-facing assistant that explains pension letters in plain language, defines difficult terms, lists the actions a member must take, and offers translation and audio for non-native speakers; prototyped in a day, then integrated and validated over twelve sprints into an on-premises production deployment that has since been replaced by another solution.
4 languages at launch — Dutch, Polish, Turkish and English MeasuredProduction Client
Read the case -
Development finance · 2025
A repeatable way to find, build and evaluate AI workflows with domain experts
A German development bank and a humanitarian NGO it funds
Seven proofs of concept across ministry reporting, partner-report checks, contract analysis, evaluation reports, technical knowledge retrieval and humanitarian self-service, each built as a multi-step process and validated by subject-matter experts on real documents.
~1 day saved per ministry monitoring report (out of ~90 per year) Customer estimateProof of concept Client
Read the case -
Automotive · 2024 – 2025
AI supplier to Bosch: hackathons, prompt engineering and safe AI-accelerated development
Bosch (multinational automotive and industrial group)
Engineering enablement for Bosch: two hackathons run on our own platform, my team and I teaching Bosch employees prompt engineering and AI-accelerated development with Cursor, Claude Code and Copilot, security standards and responsible-use guidelines for enterprise AI adoption, and MCP practices and custom skills set up inside their internal tooling.
2 hackathons run for Bosch employees on our own platform MeasuredClient
Read the case -
Enterprise software · July 2026 – present
From SaaS to on-premises, and AI skills built inside the teams
An enterprise software vendor
An ongoing independent engagement with an enterprise software vendor: a migration from SaaS to on-premises set to save the company more than €190,000 in SaaS costs alone, and AI skills built across the teams with best practices in their internal systems.
€190,000+ to be saved in SaaS costs alone by moving from SaaS to on-premises ProjectedIndependent
Read the case -
Automotive · December 2023 – present
Four procurement workflows, from scoped use case to production
A global automotive group
AI-assisted tendering and contracting for a global automotive group: drafting requests for proposals, evaluating proposals against predefined criteria and negotiating contracts through template exchanges, integrated with the procurement systems and used in production by the process owners.
4 workflows taken from proof of concept to a production MVP MeasuredProduction Client
Read the case -
Automotive · Not published
Foundation Model Evaluator: winning a global challenge on choosing between models
Stellantis
A global challenge sponsored by Stellantis asked for a way to assess and evaluate the performance of foundation models and improve how they are selected for practical AI use cases. Jaden Data entered with Patronus AI and won, with a proof of concept on retrieval-augmented generation evaluation using Ragas.
Winner of the Foundation Model Evaluator challenge, entered with Patronus AI QualitativeProof of concept Client
Read the case
Production AI and ML products
-
Enterprise AI platform · 2022 – 2024 as CTO; consulting partner since
entAIngine: one multi-tenant AI platform behind every client deployment
entAIngine
The platform every client deployment runs on: multi-tenant, event-driven on AWS, with a provider-independent model layer, role-based access control, a retrieval service over client documents, and its own tooling for prompt optimisation and pipeline evaluation.
50+ organisations running AI processes on one multi-tenant platform MeasuredProduction Own company
Read the case -
Manufacturing · 2025 – 2026
RP-Matcher, a machine-learning product matcher that takes the expert out of the critical path
RP Group (emergency lighting)
RP-Matcher, a supervised machine-learning system that matches incoming customer inquiries to the right products in the RP Group emergency-lighting catalogue, so offer preparation no longer waits for a domain expert to pick the parts by hand. It runs in production.
60% less time to prepare an offer, as reported by the customer Customer estimateProduction Client
Read the case -
Venture capital · 2023 – present
Screening thousands of startups for a venture team
Red Bull
Flowhive VC screens deal flow for Red Bull's venture team: pre-configured agents for company due diligence, competitive analysis and market mapping, with portfolio tracking around them. It has analysed thousands of startups and is still running.
Thousands of startups analysed for the venture team QualitativeProduction Client
Read the case -
Equity research · 2025 – present
Taufolio: equity research drafted from primary sources in minutes
Taufolio
My current independent build: an AI equity research platform that reads SEC filings and earnings-call transcripts, drafts a report with several models in parallel, reconciles them and anchors every claim to its source. A snapshot takes about twenty seconds, a full report about three minutes.
~3 min for a full report; about twenty seconds for a snapshot MeasuredProduction Own company
Read the case
High-scale and regulated systems
-
Legal technology · Not published
Security, audit readiness and AI inside the development process
anita.legal
An independent engagement with anita.legal, a German AI legal-research platform for law firms: security best practices and a comprehensive tech-stack review, audit, pentest, due-diligence and certification preparation with the CTO, an optimised AWS setup, a modernised architecture, and AI-accelerated development including a daemon that implements tickets and reviews pull requests.
Covered the engagement, from the long-term AWS savings alone QualitativeIndependent
Read the case -
E-commerce and banking · 2016 – 2018 (SAS Institute / ING) · 2020 – 2021 (idealo)
Login for six countries, and real-time decisioning inside a bank
idealo · ING (through SAS Institute)
Two employments at the scale end of the work: a price-comparison portal with up to 18 million monthly users, where I was a key contributor to the single sign-on internationalisation behind a six-country login rollout, and a banking programme processing thousands of customer events per second in real time.
1.4M accounts in three months, 700% over forecast, as key contributor MeasuredProduction Employment
Read the case -
Central banking, development finance and supply chain · 2021 – 2024
Regulated distributed systems: data sovereignty for institutions that cannot centralise
Norges Bank, a German development bank and EU-funded consortia
Four deliveries for institutions that need shared data without a shared owner: a cross-ledger bridge for Norges Bank, a development-finance platform on federated context brokers, an auditable carbon-capture supply chain, and a marketplace where models are trained on data nobody hands over.
4 systems delivered: a bridge, two platform MVPs and an open-source marketplace MeasuredMVP Client
Read the case
Every outcome carries its evidence type: measured, customer estimate, projected, qualitative validation or technical benchmark.
Which of these looks like your problem?
If one of them does, the fastest route is to send me the constraints you cannot move. If you are hiring instead, the leadership page has the same evidence arranged for that decision.
Remote-first, on-site when it matters; NDA on request