AI and LLM Product Case Studies

Evidence checked:

This library lists 205 AI, machine learning and LLM case studies published by 21 Python engineering companies, for CTOs and engineering managers checking who has delivered comparable work. Of these, 93 name the client. All outcomes are firm-reported: the provider published them, and this site did not audit them.

Measure Value
Case studies 205
Companies 21
Industries represented 15
Cases that name the client 93 (45%)

Featured case studies

Project Company Task What was done Reported result What it proves
Claude in Production: Halving Merge Time at Embrace Accedia Embrace Accedia consultants worked inside Embrace’s Suite Core/Social team on a live .NET back-end and notification service, helped build an AI-driven SDLC, and adopted spec-driven, AI-assisted development on the existing codebase. Accedia reports that Embrace halved pull-request cycle time and that its consultants authored 86% of commits on the Social back-end and 94% on the notification service over eighteen months. AI-assisted development rolled out on a large legacy codebase with a partner embedded in the client’s own team and backlog.
Building a Multi-Agent AI Platform Apriorit Apriorit (internal project) Apriorit’s AI and ML team built an internal multi-agent platform that automates first-pass code review of Bitbucket pull requests against the company’s standards and security requirements. Apriorit reports reducing pull request processing time by 30% and speeding up code review by 60%. An internal tool built to the same process as client work, with two measured outcomes. It is the provider’s own project, not a client engagement.
The solution for email overflow that will save you 4 hours a week BEON.tech Superhuman BEON.tech placed one autonomous React engineer into Superhuman’s small AI team to work on the text-composition experience at the core of the product. No metric published. Team augmentation for an AI product: one engineer placed to contribute without a long ramp-up. The case does not publish a measurable outcome.
Porsche E-Mobility Smart Assistant DataArt Porsche AG DataArt built an assistant that monitors battery charge, predicts when and where a driver will need to charge, and recommends a nearby charging station, using anonymized charging data to build a driver-specific ML model. No metric published. ML product work at the intersection of automotive and energy, built for a competition run by the client. No outcome is reported.
Girls Who Code + Claude Code: An AI-Native Development Practice DataArt Girls Who Code DataArt ran a three-month program to build an AI-native development practice around Claude Code, with a defined baseline and measurement plan, covering unit tests, PR review and QA test cases. DataArt reports cutting average development lead time from roughly seven days to three and a half days within three months of adoption. AI adoption run as a phased program with a baseline, not an ad hoc tool rollout. The result is a before-and-after lead-time figure.
AI contract transformation platform for banks | DORA/NIS2 Innowise Leganta Innowise built the core parsing module that turns raw contract PDFs into structured, searchable objects, with AI for field classification and semantic content analysis, and took part in architecture decisions from day one. No metric published. LLM and semantic parsing inside a regulated document workflow. The case describes scope and ownership but reports no measured result.
AI-powered compliance ecosystem case study Innowise Vidda Solutions Innowise delivered a multi-module platform with an automated compliance assessment engine, a marketplace of compliance courses and RAG recommendations, applying responsible-AI practices for GDPR and DORA. No metric published. End-to-end new product build with AI at the core and regulatory constraints. No adoption or revenue figure is reported.
AI adoption in AgTech: Koppert reaches 94% shared-workflow adoption N-iX Koppert N-iX moved Koppert onto a Claude Team subscription and ran adoption as a structured program, starting with three tasks: unit tests with Claude Code, test-case authoring with Cowork, and business analysis. N-iX reports 94% shared-workflow adoption in nine weeks. AI adoption measured by how many people use the workflow, across engineering, QA and business analysis, not by code output alone.
AI adoption in field service SaaS: WorkWave reduces workflow time by 80% N-iX WorkWave N-iX ran a three-month GenAI acceleration program with four WorkWave teams (30 engineers), using a bottom-up champion model to shortlist 14 workflows and pilot five on real production tasks. N-iX reports reducing PR cycle time by 42% in three months and building 14 reusable AI workflows; the case title cites an 80% workflow-time reduction. AI enablement across several teams with before-and-after metrics. The case publishes its baseline, which lets a reader judge the change.
Lending AI Agent Development for Atlas Credit ScienceSoft Atlas Credit ScienceSoft ran workshops to scope agentic AI use cases for lending, then delivered solution concepts, requirements specifications and agent performance and acceptance metrics ahead of deployment. No metric published. Discovery and governance work for agentic AI in a regulated lender, by a provider that had already worked with the client for over five years. No deployment outcome is reported.

Case studies: AI, machine learning and LLM products

Case Company Client Industry Type of work Stack mentioned What was built
AI Agent Platform for Finance Ops Case Study Uvik Software VellumPay Operations Fintech & financial services AI / ML / LLM Python, FastAPI, AWS, Snowflake, PostgreSQL, LLM / GenAI Multi-agent AI platform for payment exceptions with audit logs and approval gates
Dedicated Python & AI Team Case Study Uvik Software HelioForge AI SaaS & technology Team augmentation Python, FastAPI, React, AWS, Kubernetes, LLM / GenAI Dedicated Python and AI squad shipping backend APIs and LLM workflows
Enterprise RAG Retrieval and Reranking Case Study Uvik Software deepset SaaS & technology AI / ML / LLM Python, FastAPI, Azure, Kubernetes, PostgreSQL, LLM / GenAI Rebuilt hybrid retrieval, reranking and grounding for enterprise RAG platform
Python SaaS MVP Rebuild Case Study Uvik Software LaunchGrid SaaS SaaS & technology Modernization or migration Python, Django, FastAPI, AWS, PostgreSQL, LLM / GenAI Rebuilt monolithic Django MVP into modular service stack with CI/CD, RBAC, and audit logging.
Financial Crime Screening Platform Case Study Uvik Software ComplyAdvantage Fintech & financial services AI / ML / LLM Python, FastAPI, AWS, Kubernetes, PostgreSQL Rebuilt entity matching, sanctions ingestion, and alert scoring for financial crime screening platform.
LangGraph Claims Automation Case Study Uvik Software ClaimArc Insurance Systems Fintech & financial services AI / ML / LLM Python, FastAPI, AWS, PostgreSQL, LLM / GenAI Built multi-agent claims automation with document intake, policy reasoning, and audit trails.
Legal AI Contract Review Case Study Uvik Software Robin AI Fintech & financial services AI / ML / LLM Python, FastAPI, AWS, PostgreSQL, LLM / GenAI Rebuilt contract review retrieval with clause segmentation, playbook matching, and evaluation.
MCP Enterprise Knowledge Assistant Case Study Uvik Software Meridian Knowledge Systems SaaS & technology AI / ML / LLM Python, FastAPI, Snowflake, Kubernetes, LLM / GenAI Built secure MCP enterprise assistant with permissioned tools, access control, and audit logging.
Cybersecurity AI Model Pipeline Case Study Uvik Software Darktrace SaaS & technology AI / ML / LLM Python, AWS, Kubernetes, Airflow Rebuilt model retraining and evaluation pipeline for threat detection platform
Generative AI Prototype to Production Case Study Uvik Software NovaBrief Labs SaaS & technology AI / ML / LLM Python, FastAPI, AWS, LLM / GenAI Turned generative AI prototype into validated production workflow with evaluation and guardrails.
Voice AI Real-Time Pipeline Case Study Uvik Software Dialpad Telecom AI / ML / LLM Python, FastAPI, GCP, Kubernetes Rebuilt real-time transcription and conversation intelligence pipeline for voice platform.
Building a Multi-Agent AI Platform Apriorit Apriorit SaaS & technology AI / ML / LLM Python, FastAPI, PostgreSQL, LLM / GenAI Built internal multi-agent AI platform automating Bitbucket pull request code reviews.
Case Study: Speeding Up DevOps & SRE Incidents Resolution with AWS Bedrock DataArt Fiberplane SaaS & technology AI / ML / LLM AWS, PostgreSQL, LLM / GenAI Built a GenAI assistant generating Prometheus queries for DevOps and SRE teams
A Mediphany Case Study Scopic Mediphany Healthcare & life sciences AI / ML / LLM Python, AWS, LLM / GenAI AI speech-to-text automating radiology report creation
YOUCAT's App Success: Boosting User Return to 60% Boldare YOUCAT Foundation Education Modernization or migration React, Node.js, LLM / GenAI Optimized app tech stack and ran user studies to improve retention
Kyriba: A Python code interpreter integrated with OpenAI GPT STX Next Kyriba Fintech & financial services AI / ML / LLM Python, LLM / GenAI Python code interpreter integrated with OpenAI GPT for payment formats
Linde Case Study: LLM Knowledge Retrieval STX Next Linde Manufacturing & industrial AI / ML / LLM Azure, LLM / GenAI Secure RAG knowledge retrieval tool for internal documents
ANWB Leverages MLOps To Achieve Personalization At Scale Xebia ANWB Travel & hospitality AI / ML / LLM AWS, LLM / GenAI Restructured and automated ML recommendations pipeline for personalized member offers
AXA Boosts Growth & Productivity With MLOps Xebia AXA Direct Assurance Fintech & financial services AI / ML / LLM Python, Databricks MLOps best practices and automated model retraining pipeline on Databricks
Automating Invoice Processing with AWS Bedrock for Equipment Rental Company​ DataArt Renta Group Manufacturing & industrial AI / ML / LLM AWS, LLM / GenAI GenAI invoice data extraction and maintenance chatbot for equipment rental company
DataArt client: Fiberplane DataArt Fiberplane SaaS & technology AI / ML / LLM AWS, LLM / GenAI GenAI-powered assistant for DevOps and SRE incident resolution using AWS Bedrock
GenAI and Voice Recognition for Pathology Reporting | AI & Azure Case Study DataArt Voicebrook Healthcare & life sciences AI / ML / LLM Azure, LLM / GenAI Explored GenAI and voice recognition to improve pathology reporting accuracy and workflow efficiency.
Optimizing Customer Support with Real-Time Call Sentiment Analysis Using AWS Bedrock​ DataArt Glia Fintech & financial services AI / ML / LLM AWS, PostgreSQL Deployed real-time call sentiment analysis using AWS Bedrock and Whisper transcription.
Speeding Up Maintenance Process with RAG Chatbot for Equipment Rental Company​ DataArt Renta Group Manufacturing & industrial AI / ML / LLM AWS, LLM / GenAI Built RAG chatbot for technical manual search in equipment maintenance
Expand Market Research with omnichannel – N N-iX Cleverbridge E-commerce & retail Modernization or migration AWS, LLM / GenAI Migrated desktop solution to web with new UX and AI-driven personalized email campaigns
Agentic RAG System Built with Claude on AWS Vention Vention Marketing & advertising AI / ML / LLM AWS, LLM / GenAI Built internal agentic RAG system on Claude and AWS Bedrock for marketing data retrieval.
AI Product Development for Comet Vention Comet SaaS & technology Support and extension of a live product Python, LLM / GenAI Enhanced ML platform with experiment management features, library integrations, and UX upgrades.
Web development services for EliseAI Vention EliseAI Real estate AI / ML / LLM React, AWS Built AI-powered web solutions and automated leasing replies for property management
Web development services for EliseAI Vention EliseAI Real estate AI / ML / LLM React, AWS Built AI-powered web solutions and automated leasing replies for property management
Setting a British telecom up for AI adoption Neoteric 4com Telecom AI / ML / LLM Python, GCP Validated AI hypothesis for lead management with predictive models in AI Sprint.
Podimo Case Study: ML-Powered Search STX Next Podimo Media & entertainment AI / ML / LLM LLM / GenAI ML-powered semantic and conversational search for podcasts and audiobooks
Tour Partner Group Case Study: AI Hotel Bookings STX Next Tour Partner Group Travel & hospitality AI / ML / LLM LLM / GenAI Implemented AI and custom LLM to classify emails, extract data, and generate hotel booking responses.
Wunderman Thompson Case Study: AI Brand QA STX Next Wunderman Thompson Marketing & advertising AI / ML / LLM Kubernetes Built ML-powered microservices for AI brand assurance platform, optimizing data collection and preprocessing.
Accelerating Customer Support Ticket Resolution with a RAG DataArt Studytube Education AI / ML / LLM LLM / GenAI Built GenAI RAG solution to automate customer support ticket replies
Case study: eCommerce Website Migration for Intersport DataArt Intersport E-commerce & retail Modernization or migration LLM / GenAI Migrated legacy e-commerce website to modern stack with CMS and personalization
Case Study: Enhancement and Automation of Security Document Search with RAG DataArt Security Compass SaaS & technology AI / ML / LLM LLM / GenAI Built RAG solution recommending security threats and countermeasures
DataArt Case Study: Cell Solution Implementation & Automation DataArt bit.bio Healthcare & life sciences AI / ML / LLM GCP Automated cell instance segmentation using deep learning on Google Cloud
DataArt Case Study: OpenStack Cluster Monitoring System Based on Machine Learning Algorith DataArt Canonical SaaS & technology AI / ML / LLM Spark Proof of concept OpenStack cluster monitoring system using machine learning algorithms
AI adoption in AgTech: case study – N N-iX Koppert Other / not stated AI / ML / LLM Azure Built shared AI delivery practice across engineering, QA, and business analysis
AI adoption in field service SaaS: WorkWave case study – N N-iX WorkWave SaaS & technology AI / ML / LLM LLM / GenAI Built reusable AI workflows across engineering teams to speed pull-request flow
Claude Code adoption in GovTech: Neogov case study – N N-iX Neogov Government & public sector AI / ML / LLM LLM / GenAI Secure Claude Code workflows implemented across engineering teams to accelerate debugging
Computer Vision Solution for Redflex – N N-iX Redflex Logistics & transport AI / ML / LLM Python Computer vision traffic management system detecting seat belt compliance
Effective traffic management powered by computer vision – N N-iX Redflex Government & public sector AI / ML / LLM AWS Developed computer vision models for traffic management and driver detection
StarBreach: Graphical Assets for Xenko Demo Game – N N-iX Silicon Studio Media & entertainment New product build LLM / GenAI Created graphical assets and art for a VR demo game
AI contract transformation platform for banks | DORA/NIS2 Innowise Leganta Fintech & financial services AI / ML / LLM Kubernetes Built AI contract parsing module converting legal PDFs into structured searchable data
AI-powered compliance ecosystem case study Innowise Vidda Solutions SaaS & technology New product build LLM / GenAI Built AI compliance ecosystem with assessment engine, course marketplace, and RAG recommendations
Claude in Production: Halving Merge Time at Embrace Accedia Embrace Real estate AI / ML / LLM Azure Scaled AI-assisted development and modernized intranet back-end while merging two products.
Aerial Image Analysis Automation for Vexcel Imaging Vention Vexcel Imaging Other / not stated AI / ML / LLM AWS Automated aerial image processing pipelines using computer vision to detect imperfections.
AngelQ: Engineering an AI-Powered Browser for Kids Vention AngelQ Education Support and extension of a live product LLM / GenAI Provided engineering support for AI kids browser, automating CI/CD and improving streaming and user management.
RAG Platform Development for Legal Department Vention Vention Other / not stated AI / ML / LLM LLM / GenAI Built internal RAG platform for legal document search and analysis
UrgentIQ: Scaling an AI-native urgent care platform Vention UrgentIQ Healthcare & life sciences AI / ML / LLM LLM / GenAI Scaled AI-native urgent care platform with AI clinical features and Claude Code workflows
VeraGen: Claude-First Insurance SaaS on Amazon Bedrock Vention VeraGen SaaS & technology AI / ML / LLM LLM / GenAI Built Claude-first generative AI architecture with permission-aware retrieval on Amazon Bedrock
Lending AI Agent Development for Atlas Credit ScienceSoft Atlas Credit Fintech & financial services AI / ML / LLM LLM / GenAI AI agents automating lending workflows with governance and compliance
MLOps for Medical Imaging Scopic MedCAD Healthcare & life sciences AI / ML / LLM AWS Implemented MLOps infrastructure for medical imaging ML models with audit trails
Transforming Dealership Marketing with RylieAI Scopic RylieAI Automotive AI / ML / LLM LLM / GenAI Built agentic AI platform unifying dealership data and automating lead engagement
Boosting Visibility of National Art Pass Torchbox Art Fund Media & entertainment Data platform or data engineering LLM / GenAI Rebuilt analytics infrastructure and implemented SEO strategies to boost National Art Pass visibility.
The solution for email overflow that will save you 4 hours a week BEON.tech Superhuman SaaS & technology AI / ML / LLM React Enhanced AI text-composition experience in email management app
Appoint.ly – 11 weeks to build an app prototype Neoteric Appoint.ly SaaS & technology New product build LLM / GenAI Designed and built SaaS meeting scheduling app prototype after market research.
Berry Recruitment: ML Document Classification STX Next Berry Recruitment Other / not stated AI / ML / LLM — Machine learning proof of concept classifying job seeker documents
Cambium: Agentic Engineering with Claude Code STX Next Cambium E-commerce & retail AI / ML / LLM — Claude Code agentic engineering pipeline for migration tickets with human review
Claude Code at Adweek: AI-Native SDLC Case Study STX Next Adweek Media & entertainment AI / ML / LLM — Applied Claude Code across engineering and QA jobs on publishing platform.
EssenceMediacom Case Study: AI Media Management STX Next EssenceMediacom Marketing & advertising AI / ML / LLM — Built OCTRA AI-driven campaign management solution improving media workflow efficiency and compliance.
https://www.stxnext.com/portfolio/how-cbmg-slashed-server-costs-by-50-and-taught-ai-to-del STX Next CBMG Healthcare & life sciences AI / ML / LLM — AI-powered personalized feedback system with reduced server costs
Permira Case Study: Agentic AI & n8n Training STX Next Permira Fintech & financial services AI / ML / LLM — Agentic AI and workflow automation training program and hackathon
AI Security Remediation With GitHub Advanced Security Xebia Global Financial Information Provider Fintech & financial services AI / ML / LLM — AI-assisted security remediation workflow using GitHub Advanced Security and Copilot Autofix
Global Bank Reduces Analysis Effort By 70% Using AI-Enabled Legacy Modernization Xebia Global Bank Fintech & financial services AI / ML / LLM — AI-enabled legacy modernization uncovering undocumented business logic with Xebia Ace
Xebia Helps A Leading Dutch Energy Consultancy Set Its AI Course Xebia ZICHT Consultants Energy & utilities AI / ML / LLM — Executive workshop defining shared AI ambition and strategic use cases
AI Telegram Bot for TBC Bank Marketing Campaign Innowise TBC Bank Uzbekistan Fintech & financial services AI / ML / LLM — Built bilingual AI Telegram bot generating selfies with football star for marketing campaign
How Vet-AI Built an AI-Powered Healthcare Solution with Accedia Accedia Vet-AI Healthcare & life sciences AI / ML / LLM — Built AI-powered pet healthcare solution with symptom checks, consultations and automated trial management.
UrgentIQ: Healthcare software delivery with Claude Code Vention UrgentIQ Healthcare & life sciences AI / ML / LLM — Integrated Claude Code and reusable AI Skills into healthcare platform engineering workflow
AI in Child-Safe Mobile Apps: Talking Anya Case Study Scopic Sprite Labs Media & entertainment AI / ML / LLM — AI integration for child-safe interactive children's app
Bringing Art to Life: Chroma Coloring Book’s AI-Driven Innovation Scopic Chroma Coloring Book Media & entertainment AI / ML / LLM — AI generative art feature creating personalized coloring pages from prompts
Revolutionizing Orthodontics: OrthoSelect's AI-Powered Treatment Planning Scopic OrthoSelect Healthcare & life sciences AI / ML / LLM — Built deep learning model for orthodontic treatment planning and 3D models
Transforming Talent Acquisition with Codeaid AI Interviewer Scopic Codeaid SaaS & technology AI / ML / LLM — Developed AI interviewer tool automating candidate assessment for recruitment
How the Migration Museum is Using AI | Case Study Torchbox Migration Museum Media & entertainment AI / ML / LLM — AI to connect dispersed museum collections and improve story discovery.
Blockchain R&D for V-Art Sigma Software Group V-Art Media & entertainment AI / ML / LLM — Designed blockchain protocol for NFT-based IP rights licensing and monetization
AI avatar interviewing platform for recruiters GlobalDev Globaldev Group SaaS & technology AI / ML / LLM — AI avatar platform automating first-round technical recruitment screenings
AI Health & Nutrition Assistant GlobalDev National Institute of Nutrition Healthcare & life sciences AI / ML / LLM — AI health and nutrition assistant analyzing meals and forecasting health metrics
Case study: Mobiry Integrio Systems Mobiry Marketing & advertising AI / ML / LLM — Built AI-powered loyalty platform with integrations and predictive consumer behavior engine.
Case study: Synnovis Integrio Systems Synnovis Healthcare & life sciences AI / ML / LLM — Implemented AI-driven pathology test optimization integrated into physician order entry systems.
Karus Case Study Oaks Lab Karus Fintech & financial services AI / ML / LLM — Built portal turning AI underwriting models into clear lending decisions with dashboards.
Karus Case Study Oaks Lab Karus Fintech & financial services AI / ML / LLM — Built brand, UI/UX, and real-time underwriting portal visualizing AI credit risk.
Chatbot for Data Analytics Development Innowise Client not named SaaS & technology AI / ML / LLM React, Azure, Spark, Databricks, LLM / GenAI Developed LLM-based chatbot for analyzing internal documents and answering questions.
Digital platform and AWS migration for tech leader – N N-iX Client not named Media & entertainment Modernization or migration React, Node.js, AWS, LLM / GenAI Rebuilt admin solution with modern stack and migrated to AWS
The use of industrial Machine Learning in driving logistics – N N-iX Client not named Logistics & transport AI / ML / LLM AWS, Azure, GCP, Kubernetes Refactored logistics platform using machine learning for warehouse inventory
AI Medical Advice App Development Innowise Client not named Healthcare & life sciences AI / ML / LLM Python, Django, AWS, LLM / GenAI Built cross-platform travel medicine app with AI medical advice and facility locations
How a leading UK Bank Uses AI Agents for Real-Time Fraud Detection? Accedia Client not named Fintech & financial services AI / ML / LLM Python, Azure, Spark, Databricks Added real-time AI fraud detection layer into automated credit decisioning workflow for UK bank.
Case Study: AI-Powered Radiology for Smarter Patient Care Capital Numbers Client not named Healthcare & life sciences AI / ML / LLM Python, React, Node.js, AWS Built AI-powered teleradiology solution analyzing medical images and generating accurate radiology reports.
Device Health Monitoring App Development Capital Numbers Client not named Energy & utilities New product build Python, React, PostgreSQL, LLM / GenAI Built real-time device health monitoring web app with dashboards, alerts, and GPT insights.
AI GlobalDev Client not named SaaS & technology AI / ML / LLM Python, Django, PostgreSQL, LLM / GenAI AI virtual assistant automating HR employee inquiries via NLP
Disrupting the bulk shipping industry with an app Boldare Client not named Logistics & transport New product build React, Node.js, PostgreSQL, LLM / GenAI Built digital marketplace platform for bulk shipping buyers and sellers
AI-powered platform modernization for energy company – N N-iX Client not named Energy & utilities Modernization or migration Python, Django, LLM / GenAI Modernized sales platform with AI-assisted delivery and expanded test coverage
Company lifts PR throughput 8x and cuts delivery cycle time 42% with structured AI adoptio N-iX Client not named SaaS & technology AI / ML / LLM React, Node.js, LLM / GenAI Structured GenAI adoption program accelerating engineering delivery across product lines
AI Text Processing Modules for a Content Management Platform Apriorit Client not named Media & entertainment AI / ML / LLM Python, Azure, LLM / GenAI Built AI text processing modules to translate, generate, and format content within a content management platform.
Claude Code in Production – Case Study Boldare Client not named Energy & utilities AI / ML / LLM React, AWS, PostgreSQL Embedded Claude Code into engineering team to raise test coverage and velocity
Governing Agentic AI At Enterprise Scale Xebia Client not named E-commerce & retail AI / ML / LLM Azure, LLM / GenAI Governed Super Agent orchestrating enterprise AI agents through secure interface
Developing an AI Language Tutor Apriorit Client not named Education AI / ML / LLM Python, LLM / GenAI Developed AI chatbot language tutor integrated into eLearning platform with speech and mistake detection
AI Dermatologist Skin Scanner App Innowise Client not named Healthcare & life sciences AI / ML / LLM AWS Developed AI skin scanner mobile apps and admin panel for preliminary dermatology diagnostics
Medical Research Software System Innowise Client not named Healthcare & life sciences AI / ML / LLM LLM / GenAI Upgraded medical research software with AI search, dashboards, and ontology integration
AI Boosts Customer Engagement by 30% for One of Europe's Top Banks Accedia Client not named Fintech & financial services AI / ML / LLM Python Integrated AI insights into debt collection, customer engagement and marketing for a European bank.
Building AIcademy: an AI-Powered Personalized Learning Platform Accedia Client not named Education AI / ML / LLM LLM / GenAI Built AIcademy personalized learning platform with LLM-assisted materials and teacher oversight.
Customer Story – Advanced Generative AI for E Accelerance Client not named E-commerce & retail AI / ML / LLM LLM / GenAI Generative AI for personalized product recommendations in e-commerce
Customer Story Accelerance Client not named Travel & hospitality AI / ML / LLM LLM / GenAI AI consulting to improve customer support for European travel company
Customer Story Accelerance Client not named E-commerce & retail AI / ML / LLM LLM / GenAI Generative AI and chatbots to streamline food brand shopping experience
Customer Story – AI & ML Boost Paper Manufacturing Efficiency Accelerance Client not named Manufacturing & industrial AI / ML / LLM AWS AI and ML predictive maintenance for paper manufacturing machines
Customer Story – AI Accelerance Client not named Education AI / ML / LLM LLM / GenAI Generative AI avatars on WhatsApp for Enem exam preparation
Customer Story – AI Transforms Customer Support Accelerance Client not named Travel & hospitality AI / ML / LLM LLM / GenAI GenAI knowledge assistant to enhance travel company customer support
Customer Story – AI Transforms Customer Support Accelerance Client not named Healthcare & life sciences AI / ML / LLM LLM / GenAI Generative AI to streamline healthcare insurer support ticket resolution
Customer Story – Data Retrieval for Drug Discovery Accelerance Client not named Healthcare & life sciences Data platform or data engineering LLM / GenAI Improved data retrieval from complex datasets to accelerate drug discovery research
Customer Story – Generative AI for Job Processing Accelerance Client not named Healthcare & life sciences AI / ML / LLM LLM / GenAI Used generative AI to streamline job description processing for recruiters
Customer Story – HealthTech Software Efficiency Accelerance Client not named Healthcare & life sciences AI / ML / LLM LLM / GenAI Developed generative AI solution automating unit test creation for developers
Customer Story – IT Service Management with GenAI Accelerance Client not named SaaS & technology AI / ML / LLM LLM / GenAI Generative AI automates IT service management ticket screening and issue resolution
Custom NLP in Medicine for Report Data Analysis Apriorit Client not named Healthcare & life sciences AI / ML / LLM LLM / GenAI Developed custom NLP model analyzing medical scientific reports and summarizing findings.
Xebia's Agentic Fraud Intelligence Helps US Life Insurer Release Genuine Claims The Same D Xebia Client not named Fintech & financial services AI / ML / LLM — Agentic fraud intelligence scoring claims to release genuine claims same day
AI Remote Patient Monitoring Software Innowise Client not named Healthcare & life sciences AI / ML / LLM — Developed AI remote patient monitoring app for accurate wound care assessment
Breast Cancer Detection with Federated Learning Innowise Client not named Healthcare & life sciences AI / ML / LLM — Implemented federated learning framework for collaborative mammography segmentation across clinics without sharing patient data.
Development of Automated Trading Software Powered with Data Science ScienceSoft Client not named Fintech & financial services AI / ML / LLM — Built data-science trading software with pattern recognition and automated trade execution
Development of Human-Like, Real-Time AI Scheduler ScienceSoft Client not named Healthcare & life sciences AI / ML / LLM — Built HIPAA-compliant AI voice agent for healthcare appointment scheduling
AI for Energy Storage Optimization Apriorit Client not named Energy & utilities AI / ML / LLM — Built AI smart grid system forecasting energy prices and managing battery charge and discharge.
AI-Powered LMS for Pharma Employees Apriorit Client not named Healthcare & life sciences AI / ML / LLM — Delivered AI-powered LMS generating personalized courses and quizzes, GDPR-compliant, integrated with existing LMS.
Intelligent chatbot development Sigma Software Group Client not named Telecom AI / ML / LLM — Built chatbot proof of concept automating first-line support using Microsoft services
Procurement planning software development Sigma Software Group Client not named E-commerce & retail AI / ML / LLM — ML procurement planning solution predicting sales of perishable product categories
Predictive maintenance software for wind turbines Boldare Client not named Energy & utilities AI / ML / LLM — Built ML web app forecasting wind turbine failures and planning maintenance
Custom AI development for construction safety Integrio Systems Client not named Manufacturing & industrial AI / ML / LLM — Built AI SaaS platform detecting construction hazards from photos and generating compliance reports.

Showing 124 of 205 case studies (balanced across companies; topic pages list more per topic). Industry, type and summary are this directory’s classification of each page; stack lists technologies the page names. Outcomes are firm-reported.

By industry

Industry Cases
Healthcare & life sciences 33
SaaS & technology 32
Fintech & financial services 30
E-commerce & retail 14
Manufacturing & industrial 14
Media & entertainment 13
Travel & hospitality 12
Education 9
Other / not stated 9
Energy & utilities 8
Marketing & advertising 7
Logistics & transport 7
Real estate 5
Automotive 5
Telecom 4
Government & public sector 3

By type of work

Type of work Cases
AI / ML / LLM 184
Modernization or migration 11
New product build 5
Data platform or data engineering 2
Support and extension of a live product 2
Team augmentation 1

What kinds of AI and LLM work appear most often in these cases?

The cases cluster around two kinds of work: putting LLM tools into an existing engineering organization, and building AI features into a product. The first group is visible in the Accedia case with Embrace, where the company reports that its consultants adopted a spec-driven, AI-assisted workflow on a live .NET codebase, and in the N-iX cases with Koppert and WorkWave, which describe structured adoption programs across engineering, QA and business analysis. The second group includes Innowise’s work for Leganta on contract parsing and for Vidda Solutions on a compliance platform, and ScienceSoft’s lending agents for Atlas Credit.

Several cases sit between the two. Apriorit built an internal multi-agent pull request review platform for its own engineering process, and DataArt worked with Girls Who Code to introduce Claude Code into a nonprofit’s development practice. Scopic’s Talking Anya case and DataArt’s Porsche E-Mobility Smart Assistant show AI built into a consumer or driver-facing product rather than into internal delivery.

Read the task column before the result column. A case about adopting AI tools inside a team answers a different question from a case about shipping an AI feature to end users, and the two are easy to confuse when a title mentions Claude or an LLM.

What does an AI adoption engagement actually involve?

The adoption cases describe programs with a baseline, a small set of target tasks and a measurement plan, not a tool licence handed to developers. N-iX reports that it ran a three-month program with four WorkWave teams representing 30 engineers, starting from a stated baseline of 2.8 pull requests per engineer per month, a 12.4-hour average PR cycle time and 0% AI-generated code in measured output. The same company reports that at Koppert it began with three high-value tasks rather than introducing AI across every activity at once.

DataArt reports that with Girls Who Code it treated the Claude Code rollout as a phased program with a defined baseline and measurement plan from day one, and that average development lead time fell from roughly seven days to three and a half days within three months of adoption. Accedia reports that its consultants contributed 279 commits to the workflow tooling in six months and then put their own codebase through it.

These cases share a pattern: name the workflows first, pilot a few, then measure. A buyer scoping similar work should expect the provider to ask for current throughput and cycle-time figures before proposing anything.

How do providers handle governance and compliance in AI products?

Regulated clients drive most of the governance detail in these cases. ScienceSoft reports that it delivered AI agents for Atlas Credit with a focus on governance, data security and compliance, and that the engagement began with workshops, scoped use cases, a requirements specification and a set of agent performance, quality and user acceptance metrics. Innowise reports that it built the Leganta parsing module to be DORA and NIS2 ready and that it applied responsible AI practices for GDPR and DORA compliance on the Vidda Solutions platform.

Child-safety requirements appear in Scopic’s Talking Anya case, where the company reports AI-driven safety filters as part of the work for Sprite Labs. The level of regulatory detail published varies widely between cases, so a buyer in a regulated sector should check whether the case names the specific framework or only refers to compliance in general terms.

Ask for the requirements document or the metric definitions behind a governance claim. A case that lists named frameworks and acceptance criteria gives a procurement team more to work with than one that states compliance as an outcome.

What team setup should a buyer expect for an AI project?

Team sizes in these cases range from one embedded engineer to a multi-person program team. BEON.tech reports that it placed one autonomous engineer with Superhuman, now part of Grammarly, to work on AI initiatives including the text-composition experience. Boldare describes a three-person team for the YOUCAT Daily app: a native Android developer, a native iOS developer and a Node.js back-end developer. N-iX describes a champion model at WorkWave, with one AI Champion from each of four teams working alongside its GenAI Adoption Leaders.

Accedia describes a different arrangement: senior consultants forming the back-end core of Embrace’s Suite Core/Social team, working on the client’s backlog and inside its delivery process rather than as a separate group. Innowise reports that its team took ownership of a significant portion of the Leganta product from the start and was involved in architectural decisions from day one.

The setup matters more than the headcount. A buyer should decide whether the work needs an embedded engineer inside an existing team, a standalone delivery team, or an adoption program that trains the client’s own staff, because the cases show all three.

How to read a case study

Signal Why it matters What to ask the provider
Client named 93 of the 205 cases in this library name the client; the rest do not. A named client lets you check the claim with a third party. Can you name the client or give a reference contact? If the client is under NDA, can you describe the sector, scale, and contract type?
Timeline stated A case without dates or duration cannot tell you whether the work was a short prototype or a long-running production system. When did the project start and end? How long did each phase take? Was the reported result measured at launch or later?
Metric with baseline A number without a starting point is hard to judge. “No metric published” appears in the table where the case gives no measurable result. What was the value before the project? How was it measured, over what period, and by whom?
Stack listed The stack shows whether the team used tools your engineers can maintain, and whether the work matches the skills you need to hire for. Which models, frameworks, and infrastructure did you use? Who owns the code and the model weights now?
Independent review exists Every outcome in this library is firm-reported: the provider published it and this site did not audit it. An outside review or client quote adds a second source. Is there a client testimonial, audit, or public reference we can contact? Can you share the original case URL?
Still in production A model or system that shipped but was later retired proves less about long-term delivery than one still running. Is the system still live? Who maintains it? What has changed since launch?

Frequently asked questions

What AI and LLM case studies are in this library?

The library holds 205 case studies from 21 companies, and 93 name the client. Of these, 184 are tagged AI, ML or LLM work. Examples include Accedia's Claude rollout at Embrace, ScienceSoft's lending AI agents for Atlas Credit, and Innowise's contract parsing platform for Leganta. Each row links to the provider's original case study. Outcomes are firm-reported and were not audited by this site.

Which companies have published AI adoption case studies with measurable results?

Several publish numbers. N-iX reports that WorkWave cut PR cycle time by 42% in three months, and that Koppert reached 94% shared-workflow adoption in nine weeks. DataArt reports Girls Who Code cut average development lead time from roughly seven days to three and a half days. Apriorit reports a 30% reduction in pull request processing time on its internal platform. All are company-published.

Do these AI case studies name the client?

Some do, many do not. Of the 205 cases, 93 name the client. Named examples include Embrace (Accedia), Porsche AG and Girls Who Code (DataArt), Atlas Credit (ScienceSoft), Leganta and Vidda Solutions (Innowise), and Koppert and WorkWave (N-iX). Where the client is not named, the table says "Client not named", which limits what the case can prove to a buyer.

How do I check whether an AI case study result is reliable?

Look for a named client, a stated timeline, and a metric with a baseline. N-iX's WorkWave case gives a baseline of 2.8 pull requests per engineer per month and 12.4 hours average PR cycle time before the program. A result without a baseline, such as a percentage improvement alone, is harder to verify. Ask the provider for the measurement method and whether the system is still in production.

What kind of AI work do these Python companies actually do?

The cases cover AI-assisted software delivery, agentic systems, and AI features inside existing products. Accedia and N-iX run Claude-based engineering programs; Apriorit built a multi-agent code review platform; Innowise built document parsing and a compliance ecosystem; Scopic added ChatGPT and speech synthesis to a children's app. Scope ranges from one embedded engineer to multi-team adoption programs.

Related pages

This site promotes Uvik Software's services, and Uvik Software is listed first in our rankings. The other providers are ranked by the evidence score described in our methodology. In this library Uvik Software’s cases appear on the same terms as every other company’s.

Get a shortlist

Python Team for Long-Term Product Development Bulletin