SearchHorwath HTL opens in a new tab

Tech · Hotel AI Readiness Check

Questions and improvements

All 18 questions across six areas, with what full readiness looks like and how to close a gap.

Assess your hotel

Contents

Scoring capability and coverage

How answers are scored

For each question, use the global 0 to 2 definitions. The Full readiness text describes 2; partial capability or coverage is 1. Example gaps are illustrations, not additional scored subquestions.

Improvement options are a menu. Complete what addresses the actual gap, then reassess the original question. Completing one training session or buying a licence does not automatically convert the answer to 2.

0 · Not in place
The capability is absent or unavailable within the assessed scope.No follow-up field is required.
1 · Partly in place
Some capability exists, but meaningful quality, availability, reliability or coverage gaps remain.For 1, identify the gap and who or what is affected.
2 · In place where needed
The capability works reliably for the people, activities, locations or systems that need it within the agreed scope.For 2, record a brief example or confirmation.
Not sure · Verify
The answer or its coverage has not been confirmed. It is not a numerical zero.Name who could confirm this, if known.

For 2, both capability and coverage must be established. Meaningful gaps are those that prevent the intended work, exclude a necessary group or weaken an important control, not every minor imperfection.

Keeping a short note

For a Partly in place answer, choose the gaps that apply and add a short note if useful. For an In place where needed answer, record a brief example or confirmation. For Not sure, name who could confirm it if known. No follow-up field is required for Not in place. Keep the input light; an essay or attachment is not required.

A staff demonstration can establish practical ability, but one person's demonstration does not prove all included teams are covered. An account list, manager confirmation or supplier response can help establish coverage. Record disagreement as partial or unresolved rather than averaging optimistic and pessimistic answers.

Coverage rules

One capable sales user does not establish sales-team readiness. Day-shift access does not establish night-shift access. One usable system does not compensate for another essential system being inaccessible.

Central or external provision counts when it works for the assessed scope. Head office owning a contract does not establish that a property has access. Nobody receives extra credit for a large team, paid Enterprise label, particular AI brand or greater autonomy.

Calculating results

When all three questions in a section are applicable and numerically answered, add them for a score out of 6. When all 18 are applicable and answered, the optional overall sum is out of 36.

When a section contains an unknown or skipped answer, it shows as Provisional with counts of what remains to verify. No completed numerical section score or overall total is shown.

The six section profiles, confirmed gaps and unknowns are the main result. Equal points simplify the summary; they do not imply equal risk. No maturity bands, pass mark, percentile or “AI-ready hotel” certificate is provided.

The online assessment asks all 18 questions for the hotel as a whole, so every answer counts towards its area.

Area A

People

The people involved can use, evaluate and work alongside AI appropriately.

P1. Can the people expected to use AI carry out their intended tasks with it?

Consider
Confirmed direct users/operators, selected activities, teams, shifts and working conditions.
Full readiness
Intended users can demonstrate the tasks with the tools available, across the included scope. Realistic supervised exercises count; production deployment is not required.
Example partial gap
Sales staff can draft proposals, but housekeeping supervisors have not practised preparing briefings.
How to improve P1

Start here

Choose an ordinary task for each intended user group. Demonstrate it with approved information, then ask participants to complete a similar task themselves. Include the actual devices, languages and shifts involved. Identify where the difficulty lies: understanding the task, operating the tool, supplying context or using the result. Practise the missing part rather than repeating generic AI training.

Other improvement options

  • Role-specific coaching
  • Peer demonstrations
  • Short supervised practice
  • Clearer task instructions
  • Simpler interfaces
  • Refresher training for new starters

Completion check

The intended user groups can complete the task without avoidable assistance, and any unsupported groups are explicitly recorded.

Likely owner

Department lead with a practical AI support person.

P2. Can the people responsible for checking AI-assisted work judge its quality and know when to intervene?

Consider
Confirmed reviewers/approvers for every relevant output, decision, language, shift, site and external team. Include domain or risk specialists where the consequences require them.
Full readiness
Reviewers know the relevant standards, can apply suitable checks, and have the authority and time to correct, reject or escalate the work.
Example partial gap
Staff check tone but do not verify prices, facts or service commitments.
How to improve P2

Start here

Collect a few representative outputs, including plausible mistakes. Agree what must be right: facts, calculations, completeness, tone, guest treatment and permitted commitments. Practise checking against trusted sources. Establish who is responsible for acceptance and what warrants escalation. For translated work, identify someone or a process capable of checking the required languages.

Other improvement options

  • Task-specific review guides
  • Approved examples
  • Additional review for consequential work
  • Sample checks
  • Examples of common failure modes
  • Retention of source references
  • Defined acceptance and escalation criteria
  • Protected review time and backup coverage
  • Access to source information and action records
  • Language, accessibility or specialist review where needed

Completion check

Reviewers can identify a plausible error, check the underlying information and act appropriately, not merely say that the output looks reasonable.

Likely owner

The accountable operational or professional lead.

P3. Do people using or affected by AI have the guidance, support and opportunity to raise concerns they need?

Consider
Direct users, reviewers and affected staff, including participating shifts and external teams.
Full readiness
Practical help is available, people understand changes affecting their work, and they can practise, ask questions and report concerns without the arrangement depending entirely on one unavailable individual.
Example partial gap
A help contact exists for office hours, but night staff have no workable support route.
How to improve P3

Start here

Name a help contact and backup or alternative route. Explain the intended changes to the people affected and invite their input before settling the workflow. Allow practice time and give a clear route for questions and problems. Check whether the support actually works outside management's usual working hours.

Other improvement options

  • Peer support
  • Supplier assistance
  • Group-level help
  • Short drop-in sessions
  • Onboarding guidance
  • A shared collection of useful examples
  • Staff feedback reviews

Completion check

People across the included scope know where to obtain help or raise concerns, and the responsible team has demonstrated that it responds.

Likely owner

Department lead, supported by HR/training or a group service.

Area B

AI tools & access

Intended users have suitable, approved and manageable access.

T1. Do the people expected to use AI have approved access in their normal working environment?

Consider
Direct users/operators, their devices, locations and shifts, not everyone who receives the resulting work.
Full readiness
Each intended user group has authorised access to a suitable tool where and when the work occurs.
Example partial gap
Accounts exist, but attendants or evening supervisors cannot use them on their work devices.
How to improve T1

Start here

List the intended direct users by role and shift. Confirm which approved capability they should use, then test sign-in and availability on their normal devices. Resolve missing accounts, permissions, connectivity or device restrictions. Do not assume a corporate purchase has reached the property or prescribe personal accounts as a workaround.

Other improvement options

  • Enable suitable features in existing software
  • Provide managed assistant accounts
  • Use an embedded operational tool
  • Provide managed shared devices with appropriate authorised access rather than shared personal passwords

Completion check

The intended user groups can access the required capability in realistic working conditions; remaining exclusions are named.

Likely owner

Account administrator or technology provider with department managers.

T2. Is that access sufficient for the intended work?

Consider
Required functions, language/accessibility needs, usage allowances, availability and how access is funded.
Full readiness
Actual task testing confirms the necessary capability and capacity are available, with subscriptions or other access costs covered where needed.
Example partial gap
Staff have access but regularly hit a limit, lack a required feature or cannot work in a necessary language.

Subscription note: Business, team or Enterprise subscriptions can qualify. The label and price earn no points. Staff-facing subscriptions do not establish API rights, integration permissions or included API usage.

How to improve T2

Start here

Run representative tasks with the existing access. Identify the exact constraint rather than assuming a more expensive plan is needed. Check current software entitlements, expected usage and service availability. Arrange the least complicated suitable access and test again. Assess information-handling approval separately under R1.

Other improvement options

  • Reallocate licences
  • Change a subscription
  • Add a task-specific tool
  • Simplify the workflow
  • Improve devices/connectivity
  • Use an existing AI-enabled platform

Completion check

Normal intended work can be completed without recurring access or capability barriers, and the cost arrangement is clear.

Likely owner

Tool owner with department managers and budget authority.

T3. Can the hotel or group manage that access?

Consider
All accounts and permissions used for the assessed work, including centrally provided services.
Full readiness
An identified administrator can grant, change and revoke access and retain necessary business continuity when people or suppliers change.
Example partial gap
Work depends on a departing employee's personal account, or nobody at the property knows the escalation route to the central administrator.
How to improve T3

Start here

Record the approved tools, account owners and administrators. Establish how access is requested, changed and removed. Test the removal or transfer process using a suitable test or inactive account, and confirm that required business information remains available through an approved arrangement.

Other improvement options

  • Backup administration
  • Role-based permissions
  • Central billing
  • Periodic access reviews
  • Integration with joiner/leaver procedures
  • Clearer responsibilities between property and head office

Completion check

Access changes can be carried out promptly by an authorised person without avoidable loss of necessary business records.

Likely owner

Account administrator or central IT/service provider.

Area C

Data & knowledge

Required information is accessible, trustworthy and maintained.

D1. Can authorised people find the information needed for the work in a usable digital form?

Consider
Confirmed information sources for the selected activities, not every record in the hotel.
Full readiness
Required documents and records can be found and used without depending on paper, repeated retyping or someone's memory.
Example partial gap
Property facts are accessible, but housekeeping procedures exist only in an old binder or a supervisor's knowledge.
How to improve D1

Start here

List the information required by the first activities and locate its current sources. Bring a manageable set into an approved digital form and a place users can access. Ask someone who did not organise it to find and use it. Fix discovery or permissions problems before collecting a much larger information set.

Other improvement options

  • Digitise priority records
  • Improve names/folders
  • Standardise recurring notes
  • Establish an accessible handbook
  • Use supported source connections when justified

Completion check

Intended users can find and use the required information without repeatedly asking a particular individual for it.

Likely owner

Information owner and department lead.

D2. Is that information reliable enough for its intended use?

Consider
Property knowledge and operating data, their currency, completeness, consistency and authoritative source.
Full readiness
The accepted source is clear, and representative checks show sufficient accuracy, completeness and consistent definitions for the work.
Example partial gap
Two documents give different cancellation terms, or reports use different revenue definitions.
How to improve D2

Start here

Identify the authoritative source for key facts or measures. Check a representative sample for wrong facts, omissions, duplicates, outdated versions and inconsistent meanings. Resolve discrepancies that would affect the output. For numerical work, confirm dates, currencies, units, categories and totals against source records.

Other improvement options

  • Common definitions
  • Version labels
  • Automated checks
  • Standard identifiers
  • Reconciliation to source totals
  • Retiring obsolete files
  • Source references in outputs

Completion check

The sampled information meets the agreed standard, material discrepancies are resolved and users know which source to trust.

Likely owner

The department or data owner who can validate the content.

D3. Is someone responsible for keeping that information reliable?

Consider
Important information sets across the assessed activities, including shared central material.
Full readiness
Owners and update arrangements are clear, and corrections reach the places where the information is used.
Example partial gap
Procedures have been uploaded once, but nobody updates the AI-accessible copy after a policy changes.
How to improve D3

Start here

Assign an owner to each important information set. Agree what triggers an update, how staff report an error and how old material is removed or labelled. Test a routine change from the approved source through to the information people or tools actually use.

Other improvement options

  • Review schedules
  • Automated refreshes
  • Deputy owners
  • Change notifications
  • Links to maintained sources rather than duplicate copies
  • Simple correction forms

Completion check

A change or correction reaches the working source, outdated information is withdrawn and ownership is not ambiguous.

Likely owner

Named information owners, coordinated by an operational lead.

Area D

Systems & openness

The hotel can retrieve its information and establish workable connections.

S1. Can we extract and reuse information from the important systems in scope?

Consider
The actual source systems and required information, including history and permission to reuse it.
Full readiness
Usable exports have been verified and provide the required detail and history under workable reuse conditions.
Example partial gap
A report can be viewed but not exported in usable detail, or only one of two required systems permits export.
How to improve S1

Start here

Obtain a sample export from each essential system. Check required fields, dates, identifiers, detail and historical coverage. Confirm that the format can be used and the intended reuse is permitted. Ask the supplier or technical partner to clarify omissions rather than accepting a screenshot as equivalent to usable records.

Other improvement options

  • Scheduled exports
  • Agreed history retention
  • Common identifiers
  • Improved export formats
  • Data-access and exit requirements in future supplier agreements

Completion check

The information can be extracted, interpreted and reused for the stated need without an unresolved technical or commercial blocker.

Likely owner

System owner with supplier or technical partner.

S2. Can we authorise access through APIs or equivalent supported interfaces?

Consider
Real account-level rights and partner access, not whether a supplier advertises an API.
Full readiness
Supported access is confirmed for the required information and permissions, with documentation and workable authorisation arrangements.
Example partial gap
An API exists, but the property's plan, supplier restrictions or partner approval prevents the intended access.

Openness note: A working closed connector may solve one task without providing broader API choice. Manual exports, API access and working connections are separate capabilities. Browser automation does not by itself demonstrate API openness; evaluate any such workaround separately.

How to improve S2

Start here

Ask the supplier to confirm account eligibility, accessible information, permissions, documentation, partner restrictions, charges and approval steps. Record confirmed answers separately from claims that remain untested or unverified. Start with the least access needed, such as read-only access for reporting.

Other improvement options

  • Enable included API access
  • Obtain partner approval
  • Negotiate restrictions
  • Use a supported connector while documenting its limits
  • Include openness in future procurement decisions

Completion check

The hotel or its authorised partner can obtain the supported access needed under confirmed, workable terms. Read-only access can fully meet a read-only requirement.

Likely owner

System owner, head office or authorised integration partner.

S3. Can the connections needed for the work operate reliably on workable terms?

Consider
Required transfers, operating frequency, costs, error handling, approvals and support.
Full readiness
Relevant transfers have been verified in a test or live use, and ongoing cost and operational responsibility meet the need.
Example partial gap
A connector works intermittently, delivers incomplete records or has no owner when it fails.
How to improve S3

Start here

Test a required transfer using approved access and information. Compare records with the source, check timing and observe a failure or interruption in a safe test. Confirm how a failure is detected, who responds, what support costs and how the work can continue. A working test can establish readiness without requiring production deployment.

Other improvement options

  • Scheduled transfers
  • Alerts
  • Reconciliation checks
  • Retries and duplicate-prevention controls where appropriate
  • Monitoring
  • Documented recovery
  • A simpler transfer frequency where real-time access is unnecessary

Completion check

Required information arrives correctly at the necessary frequency, problems are detectable and a named person can resolve them.

Likely owner

Integration owner or technical service provider.

Area E

Safe use & control

Information boundaries, authority, oversight and recovery are clear.

R1. Have we established which information each approved AI tool may use?

Consider
The actual intended information, tools, access settings and users across the assessed work.
Full readiness
A responsible person has reviewed information handling, obtained necessary approvals, set appropriate restrictions and explained them to users.
Example partial gap
A tool is approved for public content, but staff also use it with guest records without a separate check.

Important: A vendor saying that data is not used for training does not, by itself, resolve all information-handling requirements. This question is a screening aid, not a legal or compliance determination.

How to improve R1

Start here

List the approved tools and intended information categories. Review how the tool handles access, storage, retention, sharing and reuse of information, including training use where relevant. Obtain appropriate specialist advice or approvals for sensitive uses. Explain the permitted and prohibited uses with hotel examples, then check staff understanding.

Other improvement options

  • A short tool register
  • Information-sharing guidance
  • Restricted upload permissions
  • Minimised or appropriately de-identified inputs
  • Review when provider terms or uses change

Completion check

Intended information uses have an explicit decision and working restrictions, and the users involved can explain the boundaries.

Likely owner

Tool owner with the responsible information/privacy/security approver.

R2. Is it clear what AI may do and when a person must approve or intervene?

Consider
Each selected application's outputs, authority, permission limits and oversight arrangement.
Full readiness
Responsibilities and operating limits are defined, understood and reflected in the actual workflow or technical permissions as appropriate.
Example partial gap
Staff know AI may draft a proposal, but it is unclear whether it may send it or offer a discount.
How to improve R2

Start here

State whether the application may draft, recommend, communicate or act. Name the accountable person and define approval or intervention points based on consequences. Where the tool connects to systems, ensure permissions enforce the intended limits. Test realistic boundary cases instead of relying only on a written instruction to behave carefully.

Other improvement options

  • Read-only access
  • Limited action permissions
  • Approval steps
  • Transaction limits
  • Activity records
  • Periodic checks that oversight remains workable rather than becoming automatic acceptance

Completion check

The people and tools cannot casually bypass the agreed authority boundaries, and the responsible reviewer can intervene when required.

Likely owner

Accountable business owner with technical support for enforced controls.

R3. Can we detect and deal with mistakes or failures?

Consider
Reporting, monitoring, stopping, correction and continuity across the selected uses.
Full readiness
Problems can be detected and reported; responsible people can stop unsuitable use and arrange correction or an alternative process.
Example partial gap
Staff can report a bad answer, but nobody knows who can disable the service or correct an action it has taken.
How to improve R3

Start here

Walk through a plausible wrong answer, inappropriate action or outage. Confirm who detects it, who receives the report, who can stop the affected use, and how work continues. Test the response safely. Identify who communicates with affected staff or guests and what recovery is actually possible; not every action can be undone.

Other improvement options

  • Sample reviews
  • Alerts
  • Incident records
  • Tested fallback processes
  • Correction or reversal mechanisms where available
  • Supplier escalation arrangements
  • Post-incident learning

Completion check

A realistic problem can be reported, contained and handled through a known response rather than improvisation.

Likely owner

Operational owner with the tool administrator or supplier.

Area F

Purpose & follow-through

Useful work has ownership, resources and a way to test its value.

G1. Have we considered opportunities across the assessed functions and selected work worth improving?

Consider
The stated scope, identified problems and people doing or affected by the work.
Full readiness
Priority work and desired outcomes are clear, based on actual needs and staff input, with alternatives to AI considered.
Example partial gap
Several ideas have been selected, but nobody has established what problem they solve or whether staff find it worthwhile.

Scoring note: Selecting catalogue checkboxes supplies context; it does not establish business value or automatically earn 2.

How to improve G1

Start here

Review recurring friction, errors, delays or service issues with the functions in scope. Describe the current task and desired improvement. Compare a few opportunities by likely benefit, effort and possible consequences. Check whether changing the process or using existing software would solve the problem more simply. Choose a manageable first application.

Other improvement options

  • Staff suggestions
  • Observing a task
  • Reviewing recurring complaints or errors
  • Revisiting deferred ideas
  • Combining similar needs across properties

Completion check

The owner and the people doing the work can explain the problem, intended improvement and reason for trying the chosen approach.

Likely owner

Operational manager or sponsor with the participating teams.

G2. Does someone have responsibility and the means to move the work forward?

Consider
Ownership, management backing, time, budget and approvals for the in-scope initiatives.
Full readiness
Named accountable roles have the resources and authority needed, with a clear escalation route for decisions outside their control.
Example partial gap
A willing volunteer exists but has no allocated time, budget authority or route to head-office approval.
How to improve G2

Start here

Name the person or role accountable for each priority improvement. Agree the time and support available, any budget, management backing and decision authority. Identify what needs supplier or head-office approval and resolve the immediate dependencies before setting an unrealistic delivery expectation.

Other improvement options

  • A deputy
  • Phased scope
  • External implementation help
  • Shared central support
  • Reallocated existing resources
  • A regular short sponsor review

Completion check

The owner can begin the agreed work with practical resources and knows how required decisions will be made.

Likely owner

General manager, functional sponsor or authorised group lead.

G3. Do we have a way to test value and decide what happens next?

Consider
A bounded trial or continuing review of established uses, including the complete cost and quality of the work.
Full readiness
A starting measure, success criteria, review date and continue/change/stop decision process are defined and used.
Example partial gap
Staff report that AI feels faster, but checking effort, errors, costs and service quality have not been compared.
How to improve G3

Start here

Record the current process and choose practical measures. Define a limited trial with users, boundaries and duration. Measure total preparation, checking and correction effort alongside relevant cost, quality and service outcomes. Set a review date before starting. For established uses, review continuing value rather than requiring a new pilot.

Other improvement options

  • Tests across shifts or task types
  • A simple benefit log
  • Staff or guest feedback where appropriate
  • Monitoring error rates
  • Periodic review after changes to tools or processes

Completion check

At review, the owner can explain what changed, what it cost, what problems arose and why the hotel should continue, alter or stop the use.

Likely owner

Trial/application owner with the operational team.

From findings to next steps

Turning findings into action

The result should answer three questions: What can we reasonably explore next? What must be resolved first? Who will do it?

A section score summarises foundations; it is not permission to deploy an application. A gap affecting only a later integration should not automatically block an unrelated staff-reviewed task. Conversely, a single missing safety or access requirement can block the particular use that depends on it, regardless of the total.

How recommendations respond to answers

How recommendations respond to answers
FindingAppropriate advice
0 · Capability absentUse the question's Start here guidance to establish the needed capability.
1 · Coverage gapExtend the working arrangement to the excluded teams, shifts, sites or systems. Do not automatically rebuild it.
1 · Quality or reliability gapImprove the part that is failing, using the relevant guidance.
1 · Permission or authority gapResolve the approval, access, authority, escalation or responsibility issue with the appropriate owner.
Not sureVerify with the indicated person, evidence or supplier before recommending purchases or declaring readiness.
2 · Working across scopeMaintain the capability and review after changes; do not invent an improvement deficit.

Recommend up to three immediate actions. Do not pad an action plan with unnecessary work to reach three. Keep all unresolved critical requirements visible separately, even when there are more than three.

Each action should contain a specific title; the evidence/gap behind it; the activities and people it affects; Start here; other options; a completion check; an editable owner and date; and a link to the relevant question in the reference.

Prioritise potential unsafe current uses first, then missing prerequisites for the selected focus work, then other enabling improvements. Unknowns on essential requirements become verification priorities. Do not rank solely by the lowest score, presume commercial value from catalogue selection, or recommend buying technology before checking existing provision.

Example advice

Finding
The head housekeeper has approved access, but participating supervisors on other shifts do not.
Action
Extend approved access to the supervisors preparing briefings.
Start here
Confirm who prepares briefings across the included shifts. Extend the approved arrangement to missing users and test it on their normal work devices.
Other options
Use an appropriate existing operational tool, or deliberately organise a central preparation team that can meet all participating shifts' needs. The latter is a workflow change requiring scope confirmation, not a way to conceal missing users.
Completion check
Every included shift has a workable briefing arrangement; the people producing the briefings have suitable access and someone is responsible for checking them.

A generic instruction to buy more subscriptions would not adequately describe this gap.