Skip to main content
View Categories

AI Concierge: FAQ & Troubleshooting

11 min read

The questions we hear most often, plus quick fixes for common first-day issues.


Product & Capability #

Does AI Concierge work if I only have a few listings? Yes. As of v1.54.0 there is no minimum catalog size. The chat assistant works for any number of listings and answers property-specific questions regardless of how big your catalog is. (Smaller catalogs still benefit from property chat, recovery, and photo-aware answers.) See Pricing & Plans.

Can the Concierge handle more than one question in a single message? Yes. As of v1.55.0 the Concierge is multi-intent aware. If a guest writes “somewhere pet-friendly with a pool, and what’s your cancellation policy?” it decomposes the separate intents and handles them together rather than answering only the first.

Will the AI hallucinate features my properties don’t have? No. The Concierge is constrained to your actual property data, knowledge-base entries, and brand-voice rules. When a guest asks about something the property data doesn’t cover (for example “does the hot tub work in winter?”), the Concierge says it doesn’t know and offers to connect them with your team. Confidence-aware responses are part of the core design.

Can I review what the AI says before it goes live? Yes. The Test Search tab lets you dry-run any query against your live catalog without affecting real guest sessions. The Chat Sessions log shows you every real conversation after the fact.

What languages are supported? The underlying model handles 50+ languages well. Quality is strongest in English, Spanish, French, German, Italian, and Portuguese. Brand voice carries across languages; the Concierge will sound like your brand in Spanish as much as it does in English.

Do guests know they’re talking to AI? By default, yes. The chat tab is labeled “Concierge” and the opening greeting indicates it’s an automated assistant. You can customize all customer-facing copy if you want a different framing.

Can the Concierge book a property for the guest? No. The Concierge surfaces the right property and hands off to your existing HomeRunner booking widget. We don’t replicate the checkout flow; your booking engine remains the source of truth for availability, pricing, and reservations.

How does AI Concierge compare to other vacation-rental AI tools? Most other AI tools in the space focus on operator-facing automation (auto-responses to inquiries, dynamic-pricing copilot, listing-description writers). AI Concierge is guest-facing; its job is to improve the discovery and booking experience for the people visiting your site. The two categories complement each other; we don’t compete with auto-response tools, and we work alongside them when our customers use both.

Can I customize the Concierge’s appearance? Yes. The chat drawer respects your site’s theme colors by default. Advanced operators can override CSS via the standard WordPress child-theme mechanism. We don’t recommend major visual customization for the drawer itself; the default UX is the result of significant testing.

Does it work on mobile? Yes. The chat drawer is responsive and tested on iOS Safari, Android Chrome, and desktop Chrome, Safari, Firefox, and Edge. Voice input works on mobile browsers that support the standard SpeechRecognition API (most do).

Can the chatbot describe specific rooms in my properties? Yes. Photo Analysis extracts per-room observations from your property gallery: living, kitchen, bedrooms, bathrooms, outdoor space, and amenities. When a guest asks “tell me about the outdoor space at [property],” the Concierge cites specific features the model saw in the photos (for example “above-ground pool, wooden deck, screened-in porch”), not a generic summary. It will never claim a feature is absent based on photo absence; if a room isn’t in the visual data, the Concierge falls back to your property knowledge base or recommends contacting your team. See Photo Analysis.

Pricing & Subscription #

Is there really no credit card required for the trial? Correct. The 30-day trial activates with just a license key from your welcome email. We’ll prompt you to subscribe before day 30, but the Concierge keeps working through the grace period even if you don’t.

What happens at day 30 of the trial? You’ll have received subscription emails earlier in the trial. If you’ve subscribed, nothing changes. If you haven’t, the Concierge enters a 30-day grace period (so you have time to decide), then gracefully disables. Your HomeRunner Core site continues to run unchanged.

Can I extend the trial? For active prospects working through evaluation, yes. Contact sales@homerunner.io. We’ve extended trials for operators running multi-property analysis or coordinating with internal stakeholders.

Is there an annual discount? Contact sales for annual pricing on subscriptions over $295/month.

Can I downgrade if my catalog shrinks? Yes. Pricing adjusts at your next renewal cycle based on current listing count.

What if my catalog grows mid-cycle? We don’t true-up mid-cycle. Pricing adjusts at your next renewal. You’re not penalized for growth.

Can I cancel any time? Yes. Cancellation takes effect at the end of your current billing period. No early-termination fees.

See Pricing & Plans for the full pricing table.

Setup & Configuration #

How long does setup actually take? 20 to 30 minutes for the configuration steps. Photo Analysis runs in the background after that, between 5 minutes (small catalog) and 60 minutes (very large catalog). The Concierge tab also has a setup checklist that tracks your progress. See Getting Started.

How many photos does Photo Analysis use per property? Up to 10 photos from your real property gallery, typically the first 10 of however many you have in your PMS (Guesty, Hostaway, Hostfully, OwnerRez, and others). Current versions discover the full gallery automatically; older setups sometimes fell back to the single featured image and produced thinner analyses. If you upgraded from an older version, your existing properties will use the richer 10-photo analysis the next time they’re re-analyzed. See Photo Analysis.

Do I need to write content for every property’s knowledge base? No. Start with your top 5 to 10 most-booked properties and expand over time. The Concierge works fine on properties without a custom KB; it falls back to the listing description and Photo Analysis output.

What if I don’t have a “brand voice” yet? You can launch with a generic but specific brand voice and refine over time. Example starting brand voice for an operator without strong existing voice guidelines: “Professional and helpful. We’re a regional vacation-rental company in [REGION]. Stay concise. Default to bullet points for amenity questions. Never use ‘literally’ or ‘just’ as filler.” That’s enough to launch.

Can I import knowledge base content from a spreadsheet? Currently no batch import. Use the Property KB tab to edit inline, or write to the _hr_ai_enhancer_local_kb post meta field via the WordPress REST API if you have technical resources. Batch import is on our roadmap.

Can I use AI Concierge on a staging site? Yes. License activations are per-domain, so your staging and production sites each consume an activation slot. Contact support if you need additional activation slots beyond your subscription default (most subscriptions include 2: production plus staging).

Integrations #

Does it work with my PMS? AI Concierge reads from your HomeRunner property records, which are kept in sync by your PMS connector. Whatever your PMS already syncs to HR Core, the Concierge can use. If you have a custom field for house rules or local tips, point the Concierge at it via the Chatbot -> Knowledge Base custom field dropdown.

Does it work with Conversion Kit? Yes, better than yes. AI Concierge and Conversion Kit are designed to integrate:

  • CK wishlist contents become readable in chat (“you saved Casa del Mar, Mountain Hideaway, want me to compare them?”)
  • AI-assisted bookings feed back into CK’s attribution layer
  • See the Direct Booking Platform bundle for paired pricing in Pricing & Plans

Does it work with my email marketing tool? Indirectly. AI Concierge’s lead-capture flow writes to your existing HomeRunner contact form, which integrates with most major email marketing tools (Mailchimp, ActiveCampaign, and others) via the existing HR Core webhooks. No separate Concierge-to-email integration to configure.

Can it integrate with my CRM? Same as above, via the HomeRunner contact form. Direct CRM integration is on our roadmap.

Does it work with my booking widget? Yes. The Concierge integrates with the HomeRunner booking widget out of the box and hands guests off cleanly when they’re ready to book.

Troubleshooting #

“Chat tab isn’t appearing on the live site”

Check, in order:

  • General tab -> “Enable Enhancer” is ON
  • Chatbot tab visibility rules. If “only_listed” is set, the drawer only appears on listed URLs. Set to “everywhere” to display site-wide.
  • Incognito browser test: a stale logged-in cookie can cause cached HTML to be served without the chat tab. Test in incognito to confirm. If the tab appears in incognito but not in your normal browser, clear the site’s cookies in your normal browser.

“License invalid” message

  • Verify you pasted the exact key from your welcome email (no leading or trailing spaces)
  • Check that your trial hasn’t already expired (the License tab shows the end date)
  • Use the Check License button to manually re-validate
  • If validation continues to fail, contact support. There’s a 30-day grace period so the Concierge keeps working while you reconcile.

“AI responses sound generic”

This is almost always a brand-voice issue:

  • Open the Concierge tab and rewrite your brand voice to be more specific
  • “Friendly tone” produces generic output; “warm, drops coffee-shop references, never uses ‘literally’ as filler” produces specific output
  • Aim for 3 to 5 sentences with concrete guidance and concrete don’ts

“Slow response times”

  • Photo Analysis may still be running. Let it finish, then re-test (typically 5 to 60 minutes depending on catalog).
  • Very long brand-voice or KB content (over 3,000 characters per property) slows responses; tighten where possible
  • Knowledge base entries above 3,000 characters trigger a soft warning in the editor. See the Property KB tab to spot and trim long entries.

“Search returns the wrong properties”

  • Check Search Prompt tab -> Prompt Fields table. Verify the right meta fields are toggled ON and configured as hard filters versus soft preferences.
  • Run the failing query in Test Search to see exactly how the AI parsed it
  • If a recurring pattern emerges, the Self-Assessment loop will likely surface a Concierge Lesson you can accept into the prompt

“Cost is higher than expected”

  • Open the Performance tab and review the per-feature cost breakdown (see Performance)
  • The most common cause is very long brand-voice or KB content; tighten where possible
  • Photo Analysis cost is one-time per property (about $0.20 typical). If you see ongoing photo costs, check whether your PMS is constantly re-syncing photos.

“The Concierge said something it shouldn’t have”

Pull the conversation from Chat Sessions and email it to support@homerunner.io. We can usually identify the cause (brand-voice gap, missing KB entry, edge-case prompt behavior) and recommend the right correction. For repeat patterns, the Self-Assessment loop is designed exactly for this.

“Photo Analysis Regen didn’t change anything”

If you click Regen on a property and the analysis looks identical to before, it’s almost certainly because the photo set hasn’t changed. Regen short-circuits to the cached analysis to save API costs.

To force a fresh vision call (for example after upgrading to a new version, or if you suspect the prior analysis was thin), click Clear first, then Regen. Clear wipes the cache; the next Regen has to call the API.

“Some rooms aren’t showing in the Rooms table”

The Rooms table only displays rooms the vision model actually observed in your photo gallery. If “Bedrooms” or “Bathrooms” is missing from a property’s table, it means your photo gallery didn’t include a clear bedroom or bathroom shot, not that the property lacks one.

This is intentional. The Concierge treats photo data as supportive observations, never as a complete inventory. The chatbot will never tell a guest “no bathroom is visible.” When you want a missing room to appear, add a representative photo of that room to your PMS gallery and click Clear plus Regen on the property. The next analysis will pick it up.

“I want to start over”

In the Concierge tab there’s a Flush Prompt Cache button that clears any cached prompts so your next change takes effect immediately. For a complete reset, deactivate and reactivate the plugin. Your settings are preserved across deactivation, so this is safe.

Still stuck? #

  • Documentation hub: homerunner.io/docs
  • Email support: support@homerunner.io. Under 24 hours on business days; same-day for outages.
  • Trial customers: same support coverage as paying customers
  • Strategy sessions: 30-minute calls available for active subscribers at no additional charge. Book via your customer portal.

Last updated: 2026-05-25 · AI Concierge v1.55.0

Leave a Reply

Your email address will not be published. Required fields are marked *