D3X VS CONDUIT
D3x vs Conduit: production depth vs an early-stage pitch.
Conduit is a newer hospitality AI company with a polished marketing site. D3x has spent 4+ years, with a team of 20, shipping an orchestration layer that hotels actually run: Skills Engine, 25+ live integrations, EU hosting, and 250K+ messages a month at named groups like Staycity.
WHERE WE SIT
AI that executes, not a wrapper that replies.
Left to right: does the AI only answer, or does it complete work in your PMS and ops stack? Bottom to top: bolt-on AI on a legacy product, or an AI-native runtime built for hotels.
Peers shown for context · positions are qualitative, not scored metrics
D3x in production
D3x production figures, not category averages.
- Years building hotel AI
- 4+
- People on the product
- 20
- Autonomous resolution
- 60–70%
- Messages / month
- 250K+
- AI CSAT (vs 4.62 human)
- 4.58
- Live integrations
- 25+
Founders spent 10 years as hotel operators, then built and scaled a PMS software company through exit, before founding D3x. Today D3x runs with large enterprise hotel groups across Europe.
What execution means
- Live PMS read & write-back
- Housekeeping / ops tickets created
- Human handoff with full context
THE VERDICT
The short version
Homepage claims are easy to copy. Production hotel AI is not. Conduit is early-stage; publicly, hotel-stack execution depth, named group-scale references, and published automation rates remain thin compared with vendors that have been live for years. D3x differentiates on what you can verify: founders who spent 10 years as hotel operators and exited a PMS software company before building D3x; 4+ years shipping only for hotels with a 20-person team; a Skills Engine with hundreds of codified PMS/OTA quirks; 25+ live integrations; official WhatsApp partnership; transparent per-room pricing; and large European enterprise groups: Staycity, Best Western, Amano, running 250K+ messages a month at 60–70% autonomous resolution with a 4.58 AI CSAT.
SKILLS VS FAQ AUTOMATION
Most “80% automation” is FAQ answers, not resolved work.
Vendors often count answered questions as automation. Guests still need towels, invoice fixes, reservation changes, and maintenance, and those threads land in the operator inbox unresolved. D3x Skills execute in your PMS and ops stack. That is the difference between deflection and resolution.
D3x
13/15 Resolves in stack
- 13 resolve in stack
- 1 FAQ-only
- 1 on roadmap
Conduit
0/15 Resolves in stack
- 6 route to inbox
- 5 FAQ-only
- 4 not offered
What automation claims usually count
| Skill | D3x | Conduit |
|---|---|---|
| Answer FAQs & property knowledge | FAQ / answer only | FAQ / answer only |
| Make a new booking | Resolves in stack | FAQ / answer only |
| Live rates & availability | Resolves in stack | FAQ / answer only |
| Upsell written to the reservation | Resolves in stack | Routes to inbox |
| Online / digital check-in | Resolves in stack | Routes to inbox |
Operational work that empties (or overflows) the inbox
| Skill | D3x | Conduit |
|---|---|---|
| Reservation changes in the PMS | Coming Q4 2026 | Routes to inbox |
| Housekeeping task in ops systems | Resolves in stack | Routes to inbox |
| Maintenance request in ops systems | Resolves in stack | Routes to inbox |
| Invoice request, correct & reissue | Resolves in stack | Not offered |
| Lost & found (structured) | Resolves in stack | Not offered |
| Ticketing pushed to CRM and prioritised | Resolves in stack | Routes to inbox |
| Restaurant integration / table reservation | Resolves in stack | Not offered |
| Spa reservation | Resolves in stack | Not offered |
| Agentic email that resolves threads | Resolves in stack | FAQ / answer only |
| Voice agent on the same skills | Resolves in stack | FAQ / answer only |
Coverage reflects publicly documented capabilities and typical production behaviour. Always verify write-backs on your own PMS and housekeeping stack in a paid pilot.
SIDE BY SIDE
Feature by feature
Based on public positioning and D3x production capability. We keep this honest, if something changes, we update it.
| Capability | D3x | Conduit |
|---|---|---|
| Time in market | 4+ years shipping hotel AI in production | Newer entrant, early-stage market presence |
| Team depth | 20 people focused on the product and customer success | Smaller / earlier-stage team (publicly) |
| Core architecture | AI orchestration layer above your hotel stack: Skills Engine + agents | Marketing-led AI-agent positioning; stack depth less evidenced |
| Executes actions in PMS | Yes, live reads/writes: Mews, Opera, Cloudbeds, Apaleo | Claims integrations; depth and production coverage not clearly evidenced |
| Autonomous resolution | 60–70% in production (published) | Not published with comparable production methodology |
| Channels | WhatsApp (official Meta partner), OTA inbox, web chat, SMS, email, social, voice | Chat and voice emphasized on site |
| Housekeeping & ops ticketing | Yes: Optii, Flexkeeping, Alkimii, Snapfix, UniFocus | Not clearly evidenced at hotel-ops depth |
| Agentic email workflows | Yes: FAQs, groups, invoices, booking modifications (Q4) | Not publicly detailed |
| Skills Engine (codified hotel logic) | Yes, e.g. Booking.com "_1" reservation-ID suffixes handled | Not a documented hospitality edge-case layer |
| MCP server (control from any LLM) | Included for every customer | Not publicly offered |
| Human-in-the-loop & audit logs | Full decision logs with agent rationale | Not clearly documented |
| Data residency | EU-hosted, GDPR-aligned, SOC 2 Type II in progress | Not a stated European focus |
| Advanced enterprise reporting | Portfolio AI CSAT, resolution rates, channel and property views | Not clearly documented |
| Role-based access control & enterprise security | RBAC, enterprise SSO, scoped agent permissions; SOC 2 Type II in progress | Not clearly documented |
| Pricing transparency | From €5 per room / month, public | Not publicly listed |
| Proof at group scale | Staycity, Best Western, Amano, 250K+ messages/month, 4.58 AI CSAT | Limited named hotel-group production proof publicly |
Choose D3x if you need
- A vendor that has been live for years, not a homepage from the last funding cycle
- Depth in the hotel stack: PMS, housekeeping, CRM, helpdesk, and restaurant systems
- Published production metrics: 60–70% autonomous resolution, 250K+ messages/month, AI CSAT
- A Skills Engine with hundreds of codified hotel edge cases, not just a capable model
- EU data residency, official WhatsApp partnership, and named European hotel-group references
Conduit may fit if
- You are evaluating early-stage AI vendors and are comfortable with a thin public proof record
- Your use case is closer to short-term rental messaging than multi-property hotel operations
- You want to pilot marketing claims before asking for live integration demos and audit trails
- Named European hotel-group references and published automation rates are not required for your RFP
IN PRODUCTION
Production proof, not a pitch deck
D3x founders spent 10 years as hotel operators, then built and scaled a PMS software company through exit, before founding D3x. Four-plus years and a team of 20 later, that shows up as 250K+ guest messages a month across large European enterprise groups, 60–70% autonomous resolution, a 4.58/5 AI CSAT against a 4.62 human benchmark, 25+ live integrations, and 100% logo retention. At Staycity Group, 75% of web chat and WhatsApp conversations are handled by AI, connected to live reservation data, not a scripted site demo.
FAQ
D3x vs Conduit: common questions
Maturity and evidenced depth. D3x has 4+ years and a 20-person team shipping hotel AI in production: Skills Engine, 25+ live integrations, named European enterprise groups, and published resolution rates. The founding team previously operated hotels for a decade and exited a PMS software company. Conduit is a newer entrant; publicly, hotel-stack execution and group-scale proof are much thinner than the marketing suggests. Evaluate both on live demos against your PMS, not on homepage copy.
SEE IT LIVE
Compare D3x and Conduit on your own stack.
30 minutes with the founder, your properties, your channels, and what phase-1 looks like.