Markoni uses AI to help hospitality, travel and experiences operators manage Guest touchpoints across voice, WhatsApp, email, web chat and connected operations systems.
This Responsible AI Policy explains the principles we use when designing, deploying and improving AI-assisted workflows.
1. Purpose
Markoni is designed to help operators respond faster, reduce missed inquiries, coordinate work and improve Guest experience. AI should assist teams, not remove accountability from operators.
2. Core principles
2.1 Grounded responses
AI should respond using approved property information, connected tools, configured workflows and Customer-provided knowledge sources.
2.2 No unsupported commitments
AI should avoid inventing prices, availability, policies, discounts, refunds, upgrades or commitments.
2.3 Human escalation
When AI is uncertain or a workflow requires human judgment, Markoni should escalate according to configured rules.
2.4 Guest clarity
Customer deployments should be configured so Guests are not misled about important workflows, pricing, policies or human handoff.
2.5 Auditability
Important AI interactions and actions should be logged where technically available.
2.6 Continuous improvement
AI quality should be monitored and improved through feedback, logs, testing and Customer configuration.
3. Appropriate use
Markoni may be appropriate for:
- Booking inquiries.
- Availability checks.
- Common Guest questions.
- Service requests.
- Concierge and travel desk requests.
- Upsell workflows.
- Review and feedback requests.
- Repeat Guest journeys.
- Staff handoff and task routing.
4. High-risk use cases
Markoni should not be used without appropriate human oversight for:
- Emergency response.
- Medical advice.
- Legal advice.
- Safety-critical instructions.
- High-value disputes.
- Sensitive complaints.
- Irreversible Guest-impacting decisions.
- Workflows where incorrect output could cause serious harm.
5. Customer responsibility
Customers are responsible for:
- Providing accurate property data.
- Maintaining updated policies.
- Reviewing AI configuration.
- Choosing automation boundaries.
- Monitoring AI performance.
- Responding to escalations.
- Training staff to supervise AI-assisted workflows.
- Ensuring legal compliance for Guest communications and outbound campaigns.
6. Data and training
We do not use Guest data to train public AI models.
Operational data may be used in aggregated or de-identified form to improve reliability, analytics and product quality, subject to applicable agreements and law.
7. Limitations
AI systems may make mistakes. Voice transcription, translation, intent detection, summarization and generated responses may be imperfect.
Customers should test and monitor Markoni before enabling high-impact automations.
8. Contact
For AI safety, product quality or responsible use questions, contact:
DeepNav Experiences Private Limited
Product: Markoni by GydeXP
Email: contact@gydexp.com
Address: Site 128 Kno 340/302/128, 2nd Floor, Tharabanahalli, Bettahalsur, Bangalore North, Karnataka, 562157