Generative AI Use Cases by Industry

Section 10.1.5 : Generative AI Use Cases by Industry

🎯 Learning Objective

Understand how each major industry leverages generative AI in early 2026. You will be able to identify concrete use cases, measure the economic impact of AI in your sector, and spot the most mature opportunities for your own activity.

The Economic Impact of Generative AI

Enterprise adoption of generative AI exploded between 2023 and 2026. According to the Global AI Survey 2025, 72% of large companies now use generative AI in at least one business function, up from 33% in late 2023. The question is no longer "should we adopt AI?" but "how to adopt it effectively and at what scale?"

Marketing and Communications

Generative AI has become the central tool for marketing teams. The impact is measured across 5 main areas:

Content creation at scale: a team of 3 can produce the content of 20. A content manager using GPT-5 + Midjourney v7 produces on average 8× more content than one without AI, with comparable editorial quality after iteration.

Hyper-targeted personalization: content adapted by segment, persona, funnel stage, and even by individual. Instead of a single follow-up email, AI generates 12 variants adapted to each segment's behavior.

Visual production: entire campaigns designed with DALL-E 3, Midjourney v7, or Flux. Advertising agencies report a 60-80% reduction in visual production costs for digital campaigns.

Massive A/B testing: dozens of title, hook, and visual variants generated and tested simultaneously. What took 2 weeks of brainstorming is done in 30 minutes.

Video marketing: creating ad spots with Sora 2 or Google Veo 2 for a few hundred euros vs tens of thousands in traditional production.

Case study — Coca-Cola: uses GPT-5 and DALL-E 3 to personalize advertising campaigns by market, reducing production time by 70%. In 2025, they generated their first video campaign entirely with Sora 2.

Case study — L'Oréal: developed a personalized AI beauty advisor that analyzes a photo of the customer's skin and generates product recommendations with a conversion rate 3× higher than generic recommendations.

Finance and Banking

The financial sector is one of the most advanced in generative AI adoption, driven by massive data volumes and regulatory requirements.

Regulatory document analysis: a compliance analyst spent 40h per week reading regulatory reports. Claude Opus 4.6 and GPT-5 summarize and analyze these documents in minutes, with automatic extraction of compliance points and alerts. Savings: 80% of time.

Investment reports: automatic generation of earnings call summaries, sector analyses, and recommendations. Bloomberg launched BloombergGPT, a finance-specialized LLM, to automate 30% of report writing.

Banking customer service: chatbots capable of answering complex questions about financial products. Klarna reports that its AI assistant handles the equivalent work of 700 customer service agents, with an equivalent satisfaction rate.

Enhanced fraud detection: generative models create synthetic fraud scenarios to train detection systems, improving detection rates by 25%.

Compliance scoring: automatic analysis of client contracts to detect risky clauses. A 200-page contract is analyzed in 3 minutes instead of 4 hours.

Case study — JPMorgan: developed LOXM, an AI system that executes trading orders 12% more efficiently than traditional methods. Their COiN (Contract Intelligence) program analyzes in seconds contracts that took 360,000 person-hours per year.

Healthcare and Pharma

The healthcare sector is undergoing a profound transformation, from drug discovery to patient monitoring.

Diagnostic aid: analysis of medical imaging (X-rays, MRI, retinal scans) with accuracy comparable to or exceeding specialists. Google DeepMind demonstrated a model capable of detecting over 50 eye diseases from OCT scans with 94% accuracy.

Drug discovery: reduction of R&D time from 10 years to 2-3 years for certain molecules. Generative AI identifies promising drug candidates by analyzing millions of molecular structures. AlphaFold (DeepMind) predicted the 3D structure of 200 million proteins — a feat that would have taken centuries in the lab.

Clinical documentation: automatic transcription and structuring of medical consultations. The doctor speaks naturally during the consultation, and AI automatically generates the structured report. Estimated savings: 2h/day for a general practitioner.

Clinical trials: identification of eligible patients, protocol optimization, side effect prediction through simulation.

Patient support: medical triage and information chatbots, available 24/7. Babylon Health demonstrated triage capability comparable to general practitioners for common conditions.

Case study — Insilico Medicine: used AI to identify a drug candidate for pulmonary fibrosis in only 18 months (vs the usual 4-5 years). The drug is currently in Phase II clinical trials.

Education and Training

Generative AI promises to transform mass education into personalized education — the dream of individual tutoring accessible to everyone.

Personalized tutoring: each student benefits from an assistant that adapts to their level, pace, and learning style. Khan Academy launched Khanmigo (based on GPT-4, then GPT-5), an AI tutor that guides students through Socratic questioning rather than direct answers.

Pedagogical content creation: generation of courses, exercises, quizzes, and training plans adapted to each level. A trainer using AI produces 5× more differentiated teaching material.

Accessibility: instant translation into dozens of languages, adaptation for disabilities (image descriptions for the visually impaired, text simplification), automatic subtitling of video courses.

Enriched assessment: detailed and immediate feedback on student work. Instead of "6/10," AI provides 500-word feedback explaining strengths, errors, and improvement paths. Duolingo uses GPT-5 to offer free conversations with personalized feedback.

Professional training: accelerated onboarding for new employees. A newcomer can query a "virtual expert" trained on all the company's internal documentation.

Law and Legal

The legal sector is being transformed by the ability of LLMs to understand and analyze dense text.

Legal research: analysis of millions of court decisions in seconds. Harvey AI (based on GPT-5) is used by 6 of the world's 10 largest law firms. A partner estimates that legal research has become 5× faster and 2× more exhaustive.

Contract drafting: automated first drafts of standard contracts (NDA, employment contracts, T&Cs). The lawyer focuses on customization and negotiation rather than basic drafting.

Due diligence: extraction of key information from thousands of documents for M&A. What took 3 weeks and 10 juniors now takes 3 days with AI + one supervising senior.

Case law summaries: synthesis of relevant decisions for a case, with identification of key arguments and jurisprudential trends.

Case study — Allen & Overy: launched Harvey AI in 2023, a specialized legal LLM. After a year of use, the firm reports a 30% reduction in time spent on research and initial drafting tasks.

Industry and Manufacturing

Generative AI is transforming design, production, and industrial maintenance.

Generative design: automatic exploration of thousands of design variants optimized for weight, strength, cost, or aerodynamics. Autodesk Fusion uses AI to propose designs that human engineers would never have imagined.

Predictive maintenance: anticipation of machine failures from sensor data (vibrations, temperature, acoustics). Siemens reports a 40% reduction in unplanned downtime thanks to predictive AI.

Quality control: automatic defect detection through AI vision on production lines. Defects invisible to the naked eye are detected with a 99.5% accuracy rate.

Supply chain optimization: demand forecasting and inventory management integrating external variables (weather, news, social trends).

Technical documentation: automatic generation of user manuals, technical data sheets, and maintenance procedures from CAD data.

Software Development

Coding is one of the domains most transformed by generative AI. The AI assistant has become the universal "pair programmer."

GitHub Copilot: now works with GPT-5, Claude Sonnet 4.6, and Gemini 3.1. Productivity increased by 55 to 75% according to studies (GitHub, 2025). 1.3 million paid developers in early 2026.

Code agents: Claude Code, Grok Code Fast 1, Amazon Q Developer write and deploy code almost autonomously. Claude Code can resolve entire GitHub issues: read the codebase, plan changes, write code, run tests, and submit a PR.

Automated testing: AI-generated unit and integration tests. Test coverage increased from 20% to 80% in a few hours.

Debug and refactoring: identification and correction of bugs, refactoring of complex codebases. A refactoring that would take an engineer 2 weeks is done in 2 days with a supervised AI agent.

Code review: pull request analysis and suggestions for security, performance, and readability improvements.

Impact on Employment: Transform, Not Replace

Studies converge: AI doesn't replace jobs, it transforms tasks. According to the World Economic Forum (2025):

• →92 million jobs were displaced or profoundly transformed between 2020 and 2025
• →But 170 million new roles are expected by 2030 (WEF Future of Jobs 2025 report)
• →The 5 most in-demand skills: critical thinking, creativity, AI tool proficiency, prompt engineering, adaptability
• →The "Prompt Engineer" role appeared in job postings starting in 2023, with salaries exceeding €100,000/year
• →"AI-augmented professionals" earn on average 25-40% more than their non-augmented peers in the same role (LinkedIn Workforce Report, 2025)

Practical Exercise: Map Your Sector

Duration: 15 minutes

1. →Identify 5 tasks you perform regularly in your work
2. →For each task, estimate the current time and the time with AI
3. →Calculate the total potential savings in hours/week
4. →Prioritize: which task to automate first for the best ROI?