SuprHost

Industry

Hospitality SaaS · events and venue operations

Engagement

Ongoing technology and AI automation partnership

Scope

Enquiries, events and customer communications

Services

Strategy, social management, enquiry and event workflow, automation

Stack

AI automation, social APIs, CRM

Operating model

One operational layer, automation added last

The challenge

SuprHost grew the way most hospitality businesses grow: one channel at a time. Enquiries arrived through social messages, email, web forms and phone calls, and each channel was worked by whoever happened to see it first. That is manageable at a handful of enquiries a day. It stops being manageable the moment event season arrives.

Four problems compounded as volume rose:

• No single inbox. The same enquiry could sit unanswered in one channel while a colleague answered it in another, and nobody could tell which had happened.
• No owner per enquiry. Without an assigned owner, follow-up depended on memory. Quiet enquiries went cold rather than being lost outright, which made the problem hard to see.
• Event context lived in people. What a client had asked for, what had been promised and what had changed existed in message threads rather than in a record anyone else could pick up.
• Reporting was archaeology. Answering “how many enquiries did we get and how many converted” meant reading back through channels by hand.

The brief was not to add another tool. It was to give the team one place where an enquiry becomes a tracked piece of work with an owner and a history.

Our approach

We sequenced the work so the team got relief early rather than waiting for a finished platform. Automation came last, deliberately: automating a workflow nobody agrees on just makes the disagreement faster.

1. Map how work actually runs

We followed real enquiries from first contact to booked event and wrote down what actually happened, not what the process document said. That map is what the rest of the engagement was built against.

2. Consolidate the channels

Social, email and web enquiries were routed into one operational view, so the first question a team member asks is “what needs me next” rather than “which app should I check”.

3. Give every enquiry an owner and a state

Each enquiry carries an owner, a current state and its own history. Nothing advances because someone remembered it; it advances because it is visibly waiting.

4. Automate only the settled parts

With the workflow agreed, we layered AI where it removed real work rather than where it demonstrated well: drafting routine replies, classifying and routing incoming enquiries, and chasing the follow-ups that a human would otherwise track by hand. Anything requiring judgement stays with a person.

What we delivered

• A unified social and enquiry management layer across the channels the team already used
• Enquiry routing with an owner, a state and an audit trail per enquiry
• AI-assisted reply drafting and classification, with human approval retained on anything client-facing that carries commitment
• Event management workflows so what was promised is recorded against the event rather than held in a thread
• Communication automation for confirmations and follow-ups
• Reporting on enquiry volume, response time and conversion, replacing the manual read-back

What changed

The engagement moved SuprHost from channel-by-channel firefighting to a single operational workflow. What follows describes how the work runs now compared with how it ran before — not a claim about SuprHost's commercial results, which are theirs to publish.