Holiday Extras successfully deployed ChatGPT Enterprise across their organization, demonstrating how enterprise-wide AI adoption can transform business operations and culture. The implementation led to significant measurable outcomes including 500+ hours saved weekly, $500k annual savings, and 95% weekly adoption rate. The company leveraged AI across multiple functions - from multilingual content creation and data analysis to engineering support and customer service - while improving their NPS from 60% to 70%. The case study provides valuable insights into successful enterprise AI deployment, showing how proper implementation can drive both efficiency gains and cultural transformation toward data-driven operations, while empowering employees across technical and non-technical roles.
Holiday Extras is a European leader in the travel extras market, providing services such as airport lounges, hotels, parking, and travel insurance to an international customer base. With over 40 years in the industry, the company has maintained a focus on innovation to stay competitive. This case study, published by OpenAI, documents their enterprise-wide deployment of ChatGPT Enterprise and subsequent development of customer-facing AI products using the OpenAI API.
The case study presents Holiday Extras as a success story for broad organizational AI adoption, though it should be noted that this is a promotional piece from OpenAI’s perspective. Nevertheless, it provides useful insights into how a mid-sized company approached the operational challenges of deploying LLMs across diverse business functions.
Holiday Extras faced several operational challenges that motivated their AI adoption:
Multilingual Content at Scale: As a company serving multiple European markets, they needed to produce marketing copy in numerous languages including German, Italian, and Polish. Their single marketing team was responsible for all markets, creating a bottleneck in content localization.
Data Fluency Gap: While the company had a data-driven culture with widespread dashboard usage, there was a significant gap between technical employees comfortable with SQL and less technical staff who struggled to perform independent data analysis. This limited the ability of some team members to contribute meaningfully to data-driven discussions.
Design Team Quantification: The UI/UX design function was looking to become more metrics-driven and rigorous, moving beyond purely qualitative assessments to demonstrate measurable impact within the organization.
Customer Support Scalability: Like many companies, they faced the challenge of scaling customer support efficiently while maintaining satisfaction levels.
Holiday Extras deployed ChatGPT Enterprise across their entire organization, making it available to hundreds of employees. This represents a broad horizontal deployment rather than a narrow vertical application, which presents its own operational considerations.
The internal deployment focused on providing employees with a powerful general-purpose AI assistant. Key aspects of the implementation include:
Translation and Localization Workflows: The marketing team integrated ChatGPT Enterprise into their content workflows for translating hundreds of strings into multiple languages. According to the Head of Growth (Europe), tasks that previously took weeks now take hours. The company verified translation quality using native speakers within the organization, which represents a sensible quality assurance approach for production content.
Data Analysis Democratization: Non-technical employees began uploading CSV files directly to ChatGPT for trend identification and analysis. This lowered the barrier to data analysis and allowed a broader group to participate in data-driven discussions. For employees already familiar with SQL, ChatGPT reportedly enabled them to write queries 80% faster.
Engineering Support: The development team found varied use cases depending on seniority. Junior engineers used ChatGPT as a thought partner for approach validation, while senior engineers leveraged it for communication tasks, particularly for high-stakes presentations. Code debugging times were reportedly reduced by 75%.
Custom GPT Development: The innovation team built a “UX Scoring GPT” that captures research about UI/UX principles from authoritative sources along with internal guidelines and recommendations. This custom GPT provides quantified scores for designs, identifies specific issues, and delivers actionable feedback. This represents an interesting approach to creating domain-specific tools within the ChatGPT Enterprise framework.
Beyond internal productivity tools, Holiday Extras has deployed AI in customer-facing contexts:
Syd AI: The company released a customer-facing bot called Syd AI that helps travelers understand their travel insurance policies. This represents a production deployment using OpenAI’s API, though specific technical details about the implementation architecture are not provided in the case study.
AI-Powered Customer Support: The company reports that 30% of customer service inquiries are now handled by an AI bot before escalation to human agents. This has reportedly reduced customer support costs while simultaneously improving NPS (Net Promoter Score) from 60% to 70%. The improvement in customer satisfaction alongside automation is notable, though the case study doesn’t detail how they achieved this balance or what safeguards are in place for the AI interactions.
Future Development: The company is developing what they describe as an “AI-powered, ultra-personalized Holiday Extras super app” using OpenAI’s API, aimed at providing unique trip recommendations for each customer.
The case study provides several metrics around their deployment:
These metrics should be viewed with appropriate caution as they come from a promotional case study, and the specific measurement methodologies are not disclosed.
The case study highlights an organic adoption approach. According to the Chief Growth Officer, early ChatGPT users became advocates within the organization: “Early ChatGPT users in the organization were so proud of the work they were doing, they couldn’t help but tell colleagues. Employee word-of-mouth on the quality of work became an instant driver of ChatGPT adoption.”
This grassroots approach to adoption appears to have been effective for their context, though it’s worth noting that this may not translate to all organizational cultures.
From an LLMOps perspective, this case study illustrates several patterns and considerations:
Horizontal vs. Vertical Deployment: Holiday Extras chose a broad horizontal deployment, giving the same tool to all employees rather than building specialized applications for specific workflows. This approach has tradeoffs—it’s simpler to deploy but may not optimize for specific use cases as effectively as purpose-built applications.
Quality Assurance: The mention of verifying translations with native speakers suggests some level of human-in-the-loop quality assurance for production content. This is a sensible approach for customer-facing materials where errors could impact brand perception.
Custom GPTs as Lightweight Applications: The UX Scoring GPT represents an interesting middle ground between a general-purpose assistant and a fully custom application. It allows domain expertise to be encoded without requiring custom development infrastructure.
Measurement and ROI Tracking: The company appears to have invested in measuring the impact of their AI deployment, tracking metrics like hours saved, adoption rates, and customer satisfaction. This suggests a mature approach to evaluating their AI investments.
Progression from Internal to External: The company’s journey from internal productivity tools to customer-facing applications (Syd AI, the planned super app) illustrates a common progression pattern where organizations build internal expertise before deploying AI to customers.
As a promotional case study from OpenAI, several important details are not disclosed:
The case study presents an optimistic view of the deployment, and real-world implementations typically involve more complexity and challenges than are represented here.
The Holiday Extras case study demonstrates an enterprise-wide LLM deployment spanning internal productivity tools, custom GPTs, and customer-facing AI applications. The reported results are impressive, though they should be viewed within the context of a vendor promotional piece. For organizations considering similar deployments, this case study provides useful reference points for adoption patterns, use case diversity, and measurement approaches, while the operational details remain largely undisclosed.
Snorkel developed a specialized benchmark dataset for evaluating AI agents in insurance underwriting, leveraging their expert network of Chartered Property and Casualty Underwriters (CPCUs). The benchmark simulates an AI copilot that assists junior underwriters by reasoning over proprietary knowledge, using multiple tools including databases and underwriting guidelines, and engaging in multi-turn conversations. The evaluation revealed significant performance variations across frontier models (single digits to ~80% accuracy), with notable error modes including tool use failures (36% of conversations) and hallucinations from pretrained domain knowledge, particularly from OpenAI models which hallucinated non-existent insurance products 15-45% of the time.
Stripe, processing approximately 1.3% of global GDP, has evolved from traditional ML-based fraud detection to deploying transformer-based foundation models for payments that process every transaction in under 100ms. The company built a domain-specific foundation model treating charges as tokens and behavior sequences as context windows, ingesting tens of billions of transactions to power fraud detection, improving card-testing detection from 59% to 97% accuracy for large merchants. Stripe also launched the Agentic Commerce Protocol (ACP) jointly with OpenAI to standardize how agents discover and purchase from merchant catalogs, complemented by internal AI adoption reaching 8,500 employees daily using LLM tools, with 65-70% of engineers using AI coding assistants and achieving significant productivity gains like reducing payment method integrations from 2 months to 2 weeks.
Agoda transformed from GenAI experiments to company-wide adoption through a strategic approach that began with a 2023 hackathon, grew into a grassroots culture of exploration, and was supported by robust infrastructure including a centralized GenAI proxy and internal chat platform. Starting with over 200 developers prototyping 40+ ideas, the initiative evolved into 200+ applications serving both internal productivity (73% employee adoption, 45% of tech support tickets automated) and customer-facing features, demonstrating how systematic enablement and community-driven innovation can scale GenAI across an entire organization.
