Open Questions

These are genuine uncertainties in agentic hospitality—problems the industry is working through, not marketing questions with obvious answers. We use "How Might We?" (HMW) framing to keep them productive.

HMW know which evidence to trust?

Evidence is only useful if it's reliable. But sources vary in quality and can conflict.

A venue says "dog-friendly." A review says dogs aren't allowed at breakfast. A DMO verified dogs welcome throughout. Which is current?
How should different sources be weighted? Is a DMO inspection worth more than 50 reviews?
What happens when evidence ages? A 2023 inspection might not reflect 2026 reality.
How can gaming be detected—venues manufacturing evidence or buying reviews?

Confidence scoring helps, but the thresholds are judgment calls that affect real recommendations.

HMW capture vibe in structured data?

Vibe is subjective. Can the feeling of being somewhere really be encoded in a schema?

Two people visit the same hotel—one finds it romantic, another finds it stuffy. Both are right.
Vibe descriptors (intimate, buzzy, refined) mean different things to different people.
Some venues defy categorisation. They're sui generis.
How much gets lost in translation from experience to structured fields?

Structured vibe is probably better than nothing, but it's not clear how much better.

HMW handle the cold start problem?

New venues have no evidence. How do they get discovered?

A venue joins a catalog with no reviews, no DMO verification, no booking history.
Agents prefer venues with evidence. New venues have none.
Does this disadvantage newcomers? Create a two-tier system?
How long until a venue has enough signal to compete fairly?

Curator pathways and provisional listings help, but the fundamental tension remains.

HMW resolve no-show disputes fairly?

When a guest says they arrived and a venue says they didn't, who's right?

PMS logs can be edited. Keycard systems can malfunction.
Guests sometimes arrive late or to the wrong property.
Front desk staff declarations are subjective.
The financial stakes (full charge vs full refund) make disputes heated.

Evidence requirements help, but edge cases are genuinely hard to adjudicate.

HMW handle curator conflicts of interest?

Curators verify venues. But they may also have commercial relationships with them.

A DMO's members pay membership fees. Does that compromise verification?
A portfolio curator earns commissions from listed venues. Can they be objective?
Editorial authorities accept advertising. Does that affect their ratings?

Transparency disclosure is necessary, but disclosure isn't the same as independence.

HMW know when agents should step back?

Some situations need humans. But which ones?

A guest with complex accessibility needs might need a direct conversation.
A complaint during a stay can't wait for an agent to process it.
Special requests ("can you arrange a surprise for my partner?") need human judgment.
When does agent assistance become agent interference?

Human-in-the-loop patterns help, but the boundaries are context-dependent.

HMW balance personalisation with privacy?

Better recommendations require knowing more about the guest. But how much is too much?

An agent that knows dietary restrictions, mobility needs, and past preferences can recommend better.
That same data could be used to discriminate or price-gouge.
"Personalisation" can become "profiling" quickly.
Different cultures have different expectations about privacy.

Privacy-preserving approaches exist, but the trade-offs are real and contested.

HMW prevent infrastructure from becoming gatekeepers?

Agentic booking needs shared infrastructure. Infrastructure can become bottlenecks.

If one catalog becomes the way to be discovered by agents, it's a gatekeeper.
Open specs help, but network effects could still concentrate power.
Venues dependent on agent traffic are vulnerable to platform changes.
How does infrastructure stay accountable to the ecosystem it serves?

Open protocols are part of the answer. But history shows how power concentrates even around open standards.

These aren't solved problems. They're active tensions that shape how agentic hospitality develops. Different players will make different choices, and it's not obvious who's right.

If you're thinking about these questions—as a venue, a builder, a curator, or just someone interested—perspectives are welcome.

selfe.ai/contact