Question 1

How long does building hotel guest AI in-house actually take?

Accepted Answer

A demo on an LLM API takes days; a production system typically takes 6–18 months. The long tail is integrations (PMS reads and writes, housekeeping ticketing), channel certification (WhatsApp Business API, OTA inbox), escalation and audit tooling, and the edge cases you only find in live traffic. D3x deployments go live in days to weeks because that layer already exists.

Question 2

What does the D3x Skills Engine include that a raw LLM does not?

Accepted Answer

Codified hospitality logic above the model: hundreds of PMS and OTA quirks (for example, Booking.com reservation IDs gaining a "_1" suffix through the channel manager), human-in-the-loop controls, escalation rules, and full audit logs with the agent's rationale for every action. This is what turns a capable model into 70–85% autonomous resolution in production.

Question 3

Will our engineers lose control if we buy instead of build?

Accepted Answer

No. Every D3x customer gets an MCP server, so your engineers can control and extend D3x programmatically from Claude or any LLM. You get the customization surface of an internal platform without owning connector maintenance, model upgrades, and channel certifications.

Question 4

What is the ongoing cost difference?

Accepted Answer

In-house means permanent engineering headcount plus inference and channel costs, with every LLM API change triggering re-testing. D3x starts at €5 per room per month with public SMB pricing and includes continuous LLM upgrades, the learning loop, and the AI knowledge base, with enterprise plans scoped on properties, channels, and volume.

Question 5

Is our data safe with a vendor instead of in-house?

Accepted Answer

D3x is EU-hosted and GDPR-aligned, with SOC 2 Type II in progress, role-based access control, and SSO for enterprise. Customer data is never used to train third-party foundation models, and every agent action is logged with its rationale for audit.

Capability	D3x	Building in-house
Time to first value	Days to weeks	6–18 months typical for production quality
Hotel edge cases	Hundreds codified in the Skills Engine	Discovered one incident at a time
Integrations	25+ live; new ones ~3 weeks from API access	Each connector built and maintained by your team
Channels	WhatsApp (official Meta partner), OTA inbox, web chat, SMS, email, social, voice	Each channel onboarded and certified separately
LLM upgrades	Continuous upgrades included	Your team re-tests every model and API change
Maintenance	Included — vendor-managed	Your team maintains it, indefinitely
Human-in-the-loop & audit logs	Built in, with agent rationale per decision	Built from scratch
Programmability	MCP server included — control D3x from any LLM	Full control of your own codebase
Knowledge base & learning loop	Included	Built and curated by your team
Compliance	EU-hosted, GDPR-aligned, SOC 2 Type II in progress	Your team owns audits and certifications
Cost model	From €5 per room / month, public	Engineering headcount plus inference and channel costs
Proof at group scale	250K+ messages/month, 4.58 AI CSAT, 100% logo retention	Unproven until you launch

D3x vs building in-house: live in weeks, not quarters.

The short version

Buy vs build, honestly

Choose D3x if you need

Building in-house may fit if

Production proof, not promises

D3x vs Building in-house: common questions

See how D3x compares.

Compare D3x and Building in-house on your own stack.