14. Marketing Technology, Analytics and Artificial Intelligence

Introduction and Context

Hospitality marketing has moved from periodic advertising and reservation reporting toward continuous, data-informed relationship management. A hotel may now receive a guest through an online travel agency, identify the booking in its central reservation system, record the stay in the property management system, communicate through a customer relationship platform, and evaluate campaign profitability through an analytics dashboard. These systems create substantial managerial potential, but only when their data, objectives and decisions are connected.

Marketing technology, commonly called martech, refers to the digital systems used to plan, execute, measure and improve marketing activity. In hospitality, martech includes the property management system (PMS), central reservation system (CRS), customer relationship management (CRM), customer data platform (CDP), marketing-automation tools, digital advertising platforms, revenue systems and business-intelligence applications. Artificial intelligence (AI) extends these capabilities by identifying patterns, generating predictions, automating communication and supporting managerial decisions.

The purpose of technology is not simply to collect more data. Its strategic value lies in converting data into insight, insight into action and action into measurable guest and business outcomes. A hotel that sends more emails but cannot determine whether they produce profitable direct bookings has increased activity without necessarily improving performance. Conversely, a well-designed data architecture can help a hotel recognize guest intent, personalize offers, forecast demand, allocate sales effort and protect customer trust.

This chapter examines the integration of hospitality data systems, the use of dashboards and analytical models, experimentation and attribution, predictive analytics, sales productivity, and responsible AI. It treats technology as a managerial capability rather than as a substitute for judgement or service quality.

Learning Objectives

After studying this chapter, the learner should be able to:

1. Explain the roles and relationships of PMS, CRS, CRM, CDP and marketing-automation systems.

2. Design a basic hospitality marketing-data flow and identify common data-quality problems.

3. Interpret funnels, dashboards, attribution measures and campaign performance indicators.

4. Apply experimentation and predictive analytics to marketing, revenue and sales decisions.

5. Evaluate the benefits and limitations of AI-enabled personalization and forecasting.

6. Recommend governance, privacy, security and ethical safeguards for hospitality data.

7. Assess a hotel’s technology strategy in relation to guest experience, profitability and organizational capability.

From Operational Records to Marketing Intelligence

A hotel’s data environment begins with transactions and interactions. The PMS normally contains operational information such as reservations, room types, rates, stay dates, check-in and check-out details, charges, cancellations and room status. The CRS manages availability, inventory and reservations across distribution channels, including the hotel’s direct booking engine and, in many organizations, connections to global distribution systems and online travel agencies.

The CRM is designed to support relationship management. It may store contact details, communication history, preferences, complaints, loyalty status, sales activities and campaign responses. A CDP has a related but distinct purpose: it attempts to create a unified, persistent customer profile by ingesting identifiable data from multiple sources. It is especially useful when the organization must connect anonymous digital behaviour, known reservations, email engagement, loyalty records and post-stay interactions. The distinction is not absolute because vendors use these labels differently; managers should evaluate actual functions rather than product names.

Marketing automation applies predefined rules or event-based triggers to communication and workflow. Examples include a pre-arrival email, an abandoned-booking reminder, a birthday offer, a post-stay review request or a lead-routing notification to a sales manager. Automation can improve consistency and speed, but poorly designed automation can produce irrelevant messages, duplicate contact or embarrassing errors, such as sending a promotional message to a guest who has just complained.

A simplified information flow is:

Guest interaction → data capture → identity resolution → segmentation → decision or prediction → communication or sales action → response measurement → learning.

Identity resolution is the process of deciding whether records from different systems refer to the same person or organization. It may use an email address, loyalty identifier, booking reference, telephone number or probabilistic matching. Errors at this stage are consequential. Two people may be incorrectly merged, or one guest may be represented by several incomplete profiles. Data integration therefore requires common field definitions, validation rules, synchronization schedules and clearly assigned ownership.

A useful distinction is between first-party, second-party and third-party data. First-party data is collected directly through the hotel’s own interactions, such as bookings, website behaviour, surveys and loyalty activity. Second-party data is another organization’s first-party data shared through a partnership. Third-party data is acquired or inferred from external providers. Hospitality organizations increasingly emphasize consented first-party data because it is more relevant to the relationship and less dependent on external tracking practices.

Architecture, Integration and Data Governance

Technology integration may occur through application programming interfaces (APIs), scheduled file transfers, middleware, webhooks or direct database connections. APIs allow systems to exchange information in defined formats. For example, a booking engine may transmit a new reservation to the CRS, which updates the PMS and triggers a confirmation journey in the marketing platform. Middleware can translate fields and coordinate workflows when systems use incompatible structures.

Integration should be evaluated across four dimensions. First is completeness: are all important interactions represented? Second is timeliness: does information arrive quickly enough for the intended decision? Third is accuracy: are fields correct and consistently formatted? Fourth is accessibility: can authorized staff interpret and use the information without excessive technical dependence?

Data governance provides the organizational framework for these questions. It assigns responsibility for data ownership, quality, access, retention and use. A governance program commonly includes a data dictionary, naming standards, permission levels, audit trails, retention schedules and procedures for correcting records. For example, “guest value” should not be calculated differently by marketing, finance and revenue management without an agreed definition.

Data quality is often described through dimensions such as accuracy, completeness, consistency, validity, uniqueness and timeliness. A campaign based on incomplete country codes may misdirect language-specific communication. A revenue dashboard that combines gross room revenue with net commissionable revenue may produce apparently precise but misleading results. Analytical maturity therefore depends as much on operational discipline as on sophisticated software.

The Analytical Journey: Metrics, Funnels and Dashboards

A metric is a measured quantity; a key performance indicator (KPI) is a metric linked to an important objective. A hotel should begin with the decision it wishes to improve rather than with the data available. If the objective is profitable direct acquisition, relevant measures may include direct booking conversion, cost per acquired booking, net revenue after discounts and payment costs, repeat booking rate and contribution margin. Website visits alone are not sufficient evidence of success.

A marketing funnel represents movement from awareness to consideration, booking and retention. A hospitality funnel may contain:

• reach and qualified traffic;
• website or booking-engine engagement;
• search of dates, rooms or packages;
• booking initiation;
• completed reservation;
• arrival, consumption and ancillary purchase; and
• repeat stay, advocacy or referral.

Funnel analysis identifies where potential demand is lost. A high rate of room searches but low booking completion may indicate price shock, poor mobile design, unavailable inventory, payment friction or a technical defect. A low post-stay repeat rate may indicate weak relationship value rather than weak acquisition.

Dashboards should be designed for decisions and audiences. An executive dashboard may show occupancy, average daily rate, revenue per available room, direct-channel share, customer acquisition cost and forecast variance. A campaign manager may require delivery, click-through, conversion, revenue, unsubscribe and incremental-lift measures. A sales leader may need pipeline value, stage movement, response time, win rate, account production and forecast accuracy.

A dashboard becomes harmful when it creates false precision or excessive attention to vanity metrics. Good dashboards show definitions, time periods, comparison bases, filters, data freshness and exceptions. They distinguish leading indicators, such as qualified inquiries and booking-engine searches, from lagging indicators, such as realized revenue and repeat stays. Visualization should support interpretation rather than decorate data.

Attribution and the Problem of Causality

Attribution assigns credit for an outcome, usually a booking or revenue, to marketing touchpoints. A last-click model gives all credit to the final interaction before booking. It is simple but may undervalue brand exposure, organic search, email and earlier research. A first-click model has the opposite limitation. Linear attribution distributes credit equally, while time-decay models assign greater weight to recent interactions. Position-based models emphasize selected first and last contacts.

These models are useful descriptive devices, but none automatically proves that a channel caused the booking. A guest may click a retargeting advertisement after already deciding to book. The advertisement receives credit even though its incremental effect may be small. Incrementality asks a stronger question: what additional bookings occurred because of the intervention compared with what would have happened without it?

Incrementality can be estimated through randomized holdout groups, geo-experiments, matched markets or carefully designed quasi-experimental methods. The practical challenge is that hospitality demand is affected by seasonality, events, price, availability, competitor actions and distribution restrictions. Managers should therefore avoid reallocating major budgets solely on the basis of platform-reported attribution.

Experimentation as a Management Discipline

Experimentation converts assumptions into testable propositions. A hotel might test whether a transparent package description improves direct booking completion, whether a loyalty-member rate increases net contribution, or whether a shorter inquiry form improves qualified lead volume.

A sound experiment identifies a hypothesis, the treatment, the control or comparison condition, the target population, the primary outcome, the test duration and possible risks. The primary metric should be selected before reviewing results. For booking pages, conversion rate may be relevant, but cancellation rate, net revenue and guest satisfaction should also be monitored. A treatment that increases bookings while attracting highly cancellable or unprofitable demand may not be beneficial.

Randomized A/B testing is generally stronger than comparing two non-equivalent periods. Tests should be large and long enough to reduce random fluctuation, while avoiding unnecessary exposure to a harmful treatment. Statistical significance does not guarantee managerial importance; a very small improvement may be statistically detectable but economically irrelevant. Conversely, an apparently promising result from a small sample may not be reliable. Practical significance, confidence intervals, cost and operational feasibility should inform the decision.

Predictive Analytics in Hospitality

Descriptive analytics explains what happened, diagnostic analytics investigates why it happened, predictive analytics estimates what may happen, and prescriptive analytics recommends an action. These categories often overlap, but the distinction helps managers select appropriate tools.

Demand forecasting estimates future room nights, occupancy, rate or segment production. Inputs may include historical bookings, pickup pace, cancellations, length of stay, lead time, events, holidays, weather indicators, search behaviour and competitor prices. Forecasts support pricing, staffing, inventory allocation and campaign timing. Forecast error should be measured consistently, and managers should examine bias: a model that systematically under-forecasts high-demand periods may lead to premature discounting.

Propensity models estimate the likelihood of an action, such as booking within a specified period, responding to an offer, cancelling, upgrading or purchasing an ancillary service. Churn models estimate the likelihood that a valuable customer will become inactive. Lead-scoring models rank corporate or group prospects according to fit, intent and expected value. These models are decision aids, not facts about individuals.

Customer lifetime value (CLV) estimates the expected economic contribution of a customer over a relationship period. A simplified model may be represented as:

CLV = expected contribution per stay × expected number of future stays − relationship and acquisition costs.

A more advanced calculation discounts future contribution to present value and incorporates probability of retention, ancillary revenue, service cost and channel commission. CLV should not be used to deny service or create unfair treatment. Its legitimate purpose is to guide relationship investment, offer design and portfolio decisions.

Model performance should be evaluated on data not used for training. Managers may examine accuracy, calibration, precision, recall and error by customer or market segment. A model that performs well on average but poorly for smaller geographic or demographic groups requires investigation. Data drift is also important: changes in distribution, consumer behaviour, privacy settings or market conditions can reduce model reliability.

Artificial Intelligence and Marketing Automation

AI in hospitality includes machine-learning models, natural-language processing, generative systems, recommendation engines, computer vision and conversational interfaces. Applications include demand forecasting, dynamic audience selection, automated response drafting, chatbot assistance, review analysis, upselling recommendations, sales-call summarization and anomaly detection.

Personalization may operate at several levels. Basic personalization inserts a guest’s name or destination. Rule-based personalization selects messages according to known attributes, such as a family package for a family segment. Predictive personalization uses estimated preferences or propensities. Real-time personalization adapts content to current context, such as device, search dates, availability and browsing behaviour. Greater sophistication increases both potential relevance and the need for governance.

Generative AI can help produce campaign variants, summarize account information, translate drafts and assist employees in retrieving knowledge. It may also generate inaccurate statements, disclose confidential information, reproduce biased language or create content that does not reflect the hotel’s brand promise. Human review remains essential for public claims, contractual communication, sensitive service recovery and high-value sales proposals.

AI should augment hospitality expertise rather than eliminate accountability. A revenue manager must understand why a forecast changed; a sales manager must be able to challenge a lead score; and a guest-service leader must provide a route to a human employee. Automation should be proportionate to risk. A low-risk subject-line test can be highly automated, whereas an automated decision affecting complaint resolution, pricing fairness or access to a loyalty benefit requires stronger oversight.

Sales Productivity and the Revenue-Marketing Connection

Marketing and sales systems become more valuable when they share definitions and workflows. A qualified group lead should have agreed criteria, such as event dates, room-night potential, budget fit, decision authority and purchase timing. Lead routing can then send the opportunity to the appropriate salesperson, trigger a response-time standard and record the outcome in the CRM.

Sales productivity is not equivalent to the number of calls or emails. More meaningful measures include qualified pipeline created, response time, opportunity-to-proposal rate, proposal-to-contract conversion, sales-cycle duration, win rate, account profitability and forecast accuracy. Automation can reduce administrative work by recording activities, generating reminders, preparing account briefs and identifying stalled opportunities.

The revenue-marketing connection is particularly important in hospitality because demand, price and availability interact. A campaign promising a discounted package may be attractive from a communications perspective but damaging if it displaces higher-value demand during a compression period. Marketing, sales, distribution and revenue teams should therefore coordinate calendars, inventory rules, rate fences and campaign evaluation.

Responsible Use: Privacy, Security and Trust

Hospitality data often includes personally identifiable information, payment-related information, travel patterns, preferences and sometimes sensitive special requests. Organizations must comply with applicable privacy and data-protection laws, contractual obligations and payment-security requirements. Exact legal duties vary by jurisdiction, but sound practice includes lawful collection, clear purpose, data minimization, appropriate consent where required, access and correction mechanisms, secure processing and controlled retention.

Consent should be meaningful rather than hidden in confusing language. A guest who provides an email to receive a booking confirmation has not necessarily agreed to every promotional use of that email. Preference centers can allow guests to choose communication types and frequency. Suppression lists must work across connected systems so that an opt-out is not defeated by an outdated database.

Security controls include role-based access, strong authentication, encryption, vendor due diligence, network protection, logging, staff training, incident response and tested backups. Human error remains a major risk. Employees should not paste guest records, confidential contracts or payment data into unapproved generative-AI tools.

Responsible AI requires transparency, fairness, accountability, privacy and human oversight. Managers should document the purpose of a model, training data, performance measures, limitations, decision rights and escalation procedures. Bias can enter through historical purchasing patterns, unequal data coverage, proxy variables or business rules. A model that predicts high value from past spending may reproduce unequal access to recognition or offers. Ethical design asks not only whether a model is accurate, but also whether its use is justified and consistent with the organization’s values.

Illustrative Case: The Meridian City Hotel

Background

The Meridian City Hotel, a fictional 280-room business hotel, receives reservations through its website, an online travel agency, corporate contracts and a call centre. It operates a PMS, CRS, CRM and email platform, but these systems were implemented at different times. Marketing reports campaign revenue using platform-attributed bookings, while finance reports net revenue after commission and discounts.

Situation and Problem

Direct website bookings have increased, yet contribution from the direct channel has not improved proportionately. The hotel also sends duplicate pre-arrival messages, and sales managers complain that many CRM leads lack dates, budget information or a clear next action. Senior management proposes purchasing an AI personalization product immediately.

Analysis

An initial audit finds three customer records for some guests because email addresses were entered differently. The PMS records country inconsistently, and the marketing platform does not receive cancellation updates promptly. Attribution credits many bookings to retargeting advertisements, although repeat guests often visit the website directly after receiving an email. The hotel has no agreed definition of campaign revenue and no holdout design.

The analysis concludes that the principal constraint is not the absence of AI but weak data foundations and unclear measurement. A predictive system trained on inconsistent and incomplete data could personalize confidently but incorrectly.

Approach and Solution

The hotel first creates a shared data dictionary for booking value, net revenue, direct booking, repeat guest and qualified lead. It establishes a master guest identifier, standardizes country and consent fields, and synchronizes cancellation and stay-status events. A suppression rule prevents promotional messages during unresolved complaints.

Marketing then redesigns the funnel dashboard to show traffic, booking-engine searches, initiated bookings, completed bookings, cancellations, net contribution and repeat behaviour. A randomized holdout group is used to estimate the incremental effect of a retargeting campaign. The sales team introduces lead criteria and a response-time workflow in the CRM.

Only after these changes does the hotel pilot a recommendation model for a limited segment of consenting repeat guests. The pilot recommends room upgrades and late check-out when inventory and policy permit. Human review is required for unusual recommendations, and the hotel monitors conversion, incremental revenue, cancellation, complaints and opt-outs.

Outcome

The hotel discovers that retargeting receives substantial attributed credit but produces modest incremental bookings. Email suppression reduces duplicate communication, while standardized records improve campaign reporting. Sales managers spend less time searching for information and more time developing qualified accounts. The pilot produces useful ancillary recommendations, but the hotel decides not to automate all pricing or service decisions because the evidence and governance are not yet sufficient.

Learning Points

The case demonstrates that technology investment should follow a decision problem and a data audit. Integration, definitions and experimentation create the conditions for useful AI. It also shows why gross bookings, attributed bookings and profitable incremental contribution must not be treated as interchangeable measures.

Challenges, Limitations and Common Mistakes

Several recurring mistakes undermine hospitality analytics. Organizations may purchase disconnected tools because each department selects its own platform. They may collect excessive data without a clear use, confuse correlation with causation, optimize click-through rather than profit, or assume that a dashboard is accurate because it is visually sophisticated. Another problem is “model worship”: employees accept a score or forecast without examining inputs, uncertainty and changed market conditions.

Operational adoption is equally important. If front-office or sales employees do not enter reliable information, integration will not repair the underlying gap. Staff may also resist systems that increase surveillance or administrative burden. Training should explain how data improves work and guest service, while performance expectations should focus on meaningful outcomes rather than mechanical activity counts.

Technology cannot resolve poor positioning, unattractive products, inadequate service or unrealistic commercial objectives. Nor can personalization compensate for an unavailable room, an inflexible cancellation policy or a broken booking journey. Managers should retain a systems perspective: customer experience, commercial policy, technology and human capability interact.

Current Developments and Best-Practice Direction

Contemporary hospitality organizations are moving toward composable technology architectures, event-based data exchange, privacy-enhancing measurement, unified commerce and embedded AI assistants. Real-time decisioning is becoming more feasible, although speed is not always desirable. First-party data strategies are gaining importance as tracking environments change. Digital identity, consent management and clean-room approaches may support collaboration while limiting unnecessary disclosure.

Strong practice can be summarized as follows:

1. Begin with a commercial or guest-experience decision, not a software purchase.

2. Establish common definitions, identifiers, consent rules and ownership before advanced modeling.

3. Measure net and incremental value, not only platform-attributed activity.

4. Combine quantitative evidence with employee and guest insight.

5. Test changes with clear hypotheses and monitor unintended effects.

6. Use AI proportionately, document limitations and retain human accountability.

7. Secure data throughout its life cycle and provide meaningful guest choice.

8. Review models and dashboards regularly for drift, bias and operational usefulness.

Chapter Summary

Hospitality marketing technology connects operational, commercial and relationship data so that organizations can recognize demand, communicate relevantly and evaluate business outcomes. The PMS and CRS provide reservation and inventory foundations; CRM and CDP systems support identity, relationship and profile management; marketing automation activates event-based communication and workflows. Their value depends on integration, data quality, governance and shared definitions.

Funnels and dashboards help managers locate losses and distinguish leading from lagging indicators. Attribution describes touchpoints but does not by itself establish causality; experimentation and incrementality provide stronger evidence. Predictive analytics supports forecasting, propensity assessment, lifetime-value estimation and sales prioritization, but models require validation and continuing review. AI can improve personalization, content, service support and productivity, yet it introduces risks involving privacy, security, bias, opacity and over-automation. Responsible practice combines human judgement, lawful data use, transparent governance and measurement of net, incremental and guest-centred value.

Key Terms

Application Programming Interface (API): A defined mechanism through which software systems exchange data or functions.

Attribution: The assignment of credit for a conversion or outcome to marketing touchpoints.

Artificial Intelligence (AI): Computer-based methods that perform tasks involving learning, prediction, language, perception or decision support.

Customer Data Platform (CDP): A system designed to unify identifiable customer data from multiple sources into persistent profiles.

Customer Lifetime Value (CLV): The expected economic contribution of a customer over the relationship period, after relevant costs.

Customer Relationship Management (CRM): A system and managerial approach for recording and developing customer and account relationships.

Data Governance: Policies, roles and controls governing data quality, access, use, retention and accountability.

Experimentation: Structured testing of alternative interventions to estimate their effects.

First-Party Data: Data collected directly by an organization through its own customer interactions.

Incrementality: The additional outcome caused by a marketing intervention compared with a suitable counterfactual.

Marketing Automation: Technology that executes communication or workflow based on rules, schedules or events.

Model Drift: Decline in model performance caused by changes in data, behaviour or operating conditions.

Personalization: Adaptation of content, offers or experiences to a customer’s attributes, preferences or context.

Predictive Analytics: Use of historical and current data to estimate future outcomes or behaviours.

Property Management System (PMS): The principal hotel operating system for reservations, stays, room status and related transactions.

Qualified Lead: A sales opportunity meeting defined criteria for fit, potential and purchase intent.

Revenue Attribution: Allocation of reported revenue to channels or touchpoints under a selected attribution rule.

Central Reservation System (CRS): A system managing reservations, rates and availability across relevant distribution channels.

Identity Resolution: The process of determining which records belong to the same customer or organization.

Review Questions

Short Answer Questions

1. Distinguish the primary functions of a PMS, CRS, CRM and CDP.

2. Why is identity resolution important in hospitality marketing data?

3. Explain the difference between a metric and a KPI.

4. What is incrementality, and why is it stronger than last-click attribution?

5. State two principles of responsible AI use in hospitality.

Long Answer Questions

1. Explain how data should flow from a guest interaction to a measured marketing decision in an integrated hotel technology environment.

2. Discuss the design of a useful hospitality dashboard for senior management and identify the dangers of vanity metrics.

3. Examine the role of predictive analytics in demand forecasting, customer-value estimation and sales prioritization.

4. Analyse the benefits and risks of marketing automation and AI-enabled personalization.

5. Develop a data-governance framework for a hotel that uses PMS, CRS, CRM, CDP and external marketing platforms.

Analytical / Application Questions

1. A campaign has a high click-through rate but produces few profitable bookings after cancellations and commissions. What additional measures and tests should the hotel use before expanding the campaign?

2. A hotel’s AI model recommends premium offers mainly to guests with historically high spending. Identify possible ethical and analytical problems and propose safeguards.

3. A sales team receives many leads but has a low conversion rate and poor forecast accuracy. Design a CRM workflow and dashboard that could improve qualification, response and accountability.

22. Multiple Choice Questions with Answers

Question 1

Which system is primarily responsible for managing hotel reservations, rates and availability across distribution channels?

A. Customer relationship management system

B. Central reservation system

C. Customer data platform

D. Marketing-automation platform

Correct Answer: B. Central reservation system

Question 2

A hotel wants to determine whether retargeting caused additional bookings rather than merely receiving credit for bookings likely to occur anyway. Which method is most appropriate?

A. Last-click attribution

B. Equal credit attribution

C. Randomized holdout testing

D. Counting advertisement impressions

Correct Answer: C. Randomized holdout testing

Question 3

Which situation most clearly indicates an identity-resolution problem?

A. A dashboard displays data by month

B. A campaign has a low open rate

C. One guest has three separate profiles because of inconsistent email entries

D. A forecast includes a confidence interval

Correct Answer: C. One guest has three separate profiles because of inconsistent email entries

Question 4

Why should a hotel monitor net contribution rather than only completed bookings when evaluating a campaign?

A. Net contribution accounts for discounts, commissions and related costs

B. Completed bookings cannot be recorded in a PMS

C. Net contribution eliminates the need for experimentation

D. Bookings are never useful as marketing measures

Correct Answer: A. Net contribution accounts for discounts, commissions and related costs

Question 5

Which statement best describes predictive analytics?

A. It records only historical transactions without interpretation

B. It estimates likely future outcomes using data and models

C. It guarantees that every individual will behave as predicted

D. It replaces the need for managerial judgement

Correct Answer: B. It estimates likely future outcomes using data and models

Question 6

A hotel sends an upselling message immediately after a guest submits a serious complaint. Which control would most directly prevent this failure?

A. Increasing advertising impressions

B. Creating a complaint-related communication suppression rule

C. Using a last-click attribution model

D. Removing all guest data from the CRM

Correct Answer: B. Creating a complaint-related communication suppression rule

Question 7

Which is the strongest reason to audit data quality before implementing an advanced AI personalization product?

A. AI systems never require customer data

B. Poorly standardized data can produce confidently incorrect recommendations

C. Data quality affects only the visual appearance of dashboards

D. Personalization is unrelated to identity resolution

Correct Answer: B. Poorly standardized data can produce confidently incorrect recommendations

Question 8

A dashboard showing website visits, booking-engine searches, initiated bookings and completed reservations is primarily being used to analyse what?

A. A marketing and booking funnel

B. A property maintenance schedule

C. A staff payroll cycle

D. A payment-security audit

Correct Answer: A. A marketing and booking funnel

Question 9

Which practice best reflects responsible use of generative AI by a hotel employee?

A. Uploading confidential guest records to any public AI tool

B. Publishing every AI-generated claim without review

C. Using AI to draft content while applying human review and approved-data controls

D. Allowing AI to make all complaint decisions without escalation

Correct Answer: C. Using AI to draft content while applying human review and approved-data controls

Question 10

Why can a high platform-attributed return on advertising spend be misleading?

A. Attribution always ignores revenue

B. The platform may claim credit for demand that would have occurred without the advertisement

C. Advertising platforms cannot record clicks

D. Return on advertising spend measures only employee satisfaction

Correct Answer: B. The platform may claim credit for demand that would have occurred without the advertisement