Industry

Holiday Extras Deploys ChatGPT Dots as Autonomous Work Assistants

UK travel company Holiday Extras uses OpenAI's dots—persistent AI agents—to automate routine tasks and proactive reporting without explicit user requests.

Last verified:

Dots Operate Proactively Across Connected Enterprise Data

According to the OpenAI Blog, Holiday Extras—a UK travel technology platform specializing in airport logistics and trip ancillaries—has deployed ChatGPT Enterprise dots as persistent work assistants that operate without explicit task assignment. The distinguishing behavior: dots autonomously monitor connected data sources and surface findings before being asked. When Holiday Extras launched a new app feature, VP of AI Innovation Liam Thompson’s dot, which he names Beast, independently pulled performance metrics from Google BigQuery via ChatGPT’s plugin integration and delivered a report unprompted. This proactive capability—detecting events, gathering context, and surfacing summaries—marks a departure from query-response chatbot interaction.

Two-Hour Daily Voice Conversations Replace Email Triage

Liam Thompson reports spending at least two hours daily in voice conversations with Beast, using bidirectional audio as his primary interface for morning and evening briefings. The dot surfaces unanswered email replies, reconstructs past conversation context without manual scrolling, and maintains awareness of concurrent organizational activities. Thompson describes this as “orchestrator” behavior: Beast flags forgotten follow-ups, retrieves answers from prior discussions, and operates as an always-on oversight mechanism. Thompson granted Beast access to Slack, inbox, Google Drive, and his laptop—a permissions scope he said he would not have extended to competing assistants (OpenAI Blog references OpenClaw and Hermes as alternatives Thompson evaluated but rejected due to trust concerns about their enterprise environment handling).

Autonomous System Monitoring Surfaces Operational Issues

David Lee, Director of Innovation & Growth at Holiday Extras, names his dot Brian and uses it on his morning commute to preview upcoming meetings and request data-backed reports from company systems. According to the OpenAI Blog, Brian autonomously detected that deprecated API keys required retirement, then traced through Lee’s email history to surface affected stakeholders without manual reconstruction. This pattern—the dot identifying issues not yet surfaced to the human operator and pre-gathering relevant context—suggests dots function as continuous monitors rather than on-demand tools.

Why This Matters

The Holiday Extras deployment signals a shift in how enterprises operationalize ChatGPT Enterprise. Rather than a productivity layer for individual query handling, dots position AI agents as independent monitors of business systems, capable of identifying and contextualizing issues before human review. For organizations managing distributed systems (APIs, data pipelines, feature rollouts) or high-volume communication (email, Slack), this reduces triage overhead and latency for routine alerts. The two-hour daily voice engagement and cross-platform data access (Slack, email, Drive, BigQuery) indicate that trust in the AI platform’s enterprise security posture—not just capability—is a primary adoption driver. Teams evaluating ChatGPT Enterprise in 2026 should expect this autonomous, multimodal assistant behavior to become table-stakes for competitive positioning.

Frequently Asked Questions

What is a 'dot' in this context?

A dot is a persistent ChatGPT-powered AI assistant that operates within ChatGPT Enterprise. Unlike chat-based assistants, dots maintain access to connected data sources and can initiate work autonomously without explicit user prompts.

How does the proactive monitoring work?

According to Holiday Extras' deployment, dots monitor connected systems (email, Slack, BigQuery, Drive) and surface alerts or reports when they detect relevant changes—such as a feature launch or deprecated API keys—without being asked.

What trust or security model enables this access?

Holiday Extras VP Liam Thompson cited trust in OpenAI's enterprise environment as the reason he granted his dot access to sensitive systems like Slack, email, and Drive, in contrast to other assistants he evaluated but did not trust with the same level of access.

#enterprise-ai #agents #openai #automation #chatgpt-enterprise