Article 03 · October 12, 2026
Why Explainable AI Can Still Produce Bad Decisions
Making an AI recommendation easier to understand is not the same as making a manager better able to judge it.
Research on AI and decision authority in hotels
How should hotels divide decision authority between managers and AI?
Hotel Decision Lab is a research-informed publication on how hotel managers and intelligent systems should share decision authority in operations. The current focus is AI-supported labor decisions, managerial judgment, calibrated reliance, and human oversight.
Explore the research agendaResearch focus
-
Comparative Capability
How AI and human alternatives perform for a given decision class and operating regime.
Explore the research -
Human Contribution
When manager knowledge adds information or catches errors, and when it adds noise or bias.
Explore the research -
Decision Rights
How authority, modification bounds, pause triggers, escalation, monitoring, feedback, and accountability should be assigned.
Explore the research
Featured article
Who Controls the Decision When Hotel AI Recommends It?
The accountability question starts before the manager clicks approve.
When a hotel uses algorithmic forecasts, scheduling recommendations, or work-allocation systems, the usual accountability question shows up late: did the manager follow the recommendation or override it? The better question starts earlier.
Read Article 04
Latest Articles
View all articles
Article 02.5 · September 30, 2026
When Hotel AI Works, the Next Problem May Be Orchestration
A capable AI model does not complete a hotel decision by itself: what has to happen between a recommendation and an operating outcome.
Article 02 · September 4, 2026
When Local Knowledge Beats the Forecast
When manager-supplied local information deserves decision weight, and when intervention adds noise.
Start here
A research gateway, an agenda, and the flagship argument.
AI in Hotels
What the term means, where AI enters hotel operating decisions, what current evidence supports, and what remains unknown.
Explore AI in hotelsResearch Agenda
The canonical question, evidence position, decision architecture, research boundaries, and five priority questions.
Read the research agendaManager Overrides
Why a human approval step should not be assumed to make an AI-supported hotel labor decision safer or more accurate.
Read the first flagship article
Current research question
How should decision authority be divided between hotel managers and intelligent systems?
The current application is AI-supported hotel labor decision-making. The recommendation instance remains the primary unit of analysis, but authority should be designed before a live recommendation appears.
Methodology and evidence boundaries
How Hotel Decision Lab handles evidence.
Hotel Decision Lab uses a structured integrative review to synthesize relevant hotel and adjacent-domain evidence, identify contradictions and gaps, and develop bounded propositions for future testing.
This work is not presented as a systematic review, peer-reviewed research, original empirical research, or a validated framework.
Read the methodology- Direct finding What a source reports. The statement stays close to the study’s population, setting, measures, comparison, and outcome.
- Transferred evidence Transferred evidence comes from adjacent settings such as forecasting, medicine, public-sector decisions, or general algorithm-use research. It can identify mechanisms and boundary conditions, but transfer to hotel operations is labeled and not presumed.
- Author interpretation How the source authors explain their finding. An interpretation is attributed to the source rather than restated as a demonstrated fact.
- Hotel Decision Lab inference A bounded synthesis across evidence and hotel context. The reasoning is made explicit, including where it transfers from another domain or depends on hotel operating logic.
- Proposition A claim designed for future testing. Propositions are not findings. Framework elements and decision modes remain proposed unless empirical work validates them.
- Open question What the evidence does not yet resolve. Unknowns remain visible instead of being filled by search demand, professional intuition, or rhetorical certainty.
Founder and author
Drew Potter
Drew Potter has a background in hotel operations and hospitality technology and is completing an MBA focused on business analytics and digital marketing.
Professional experience helps identify meaningful operating questions and interpret practical implications. It is not presented as empirical research evidence.
About Drew and Hotel Decision Lab