Clays gets a lot of questions. Availability, group sizes, food, drinks, packages, private events, changing a booking — the usual things you’d expect when you run busy competitive socialising venues.

Their phones used to be nonstop.

Today, a lot of those conversations start with Peggy instead. Peggy is Clays’ digital colleague, powered by Rookoo. She works across the website, WhatsApp and Instagram, answers questions, gathers the information the team actually needs and hands over when a human should take it from there.

She’s now handled more than 20,000 customer conversations.

And once you get to that kind of volume, you start learning quite a lot.

People really don’t like waiting

That part isn’t particularly surprising.

Someone lands on your website at 21:46 because they’re trying to organise a birthday for 18 people. They want to know if you’re free next Saturday, what it costs and whether they can add food. They’re probably not going to write the question down and patiently wait until your sales team gets back in the morning.

They’re going to ask someone else too.

For Clays, that used to create a lot of pressure on the people answering phones and enquiries. As Dan puts it:

“Our phones are nonstop — or they were nonstop.”

Peggy now handles a large part of that first contact. She can answer the straightforward questions immediately, collect the details the team needs and keep the conversation moving.

That doesn’t mean the team disappears from the process. Quite the opposite. They get involved when their attention actually matters.

One of those handovers turned into a £25k booking

When a conversation reaches the point where someone from Clays should step in, the team gets flagged and join while the customer is still there.

That sounds simple, but timing matters. A promising enquiry sitting in an inbox for three hours is very different from a promising enquiry where someone can jump in there and then.

Kimberleigh Lloyd, Revenue Director at Clays, explains:

“We have that flag up now so the team can jump in and in real time can convert bookings kind of straight away, which we’ve seen huge success with. We had an event of about 25 grand that booked right there and then.”

That distinction matters.

Peggy didn’t close a £25k event. The Clays team did.

What Peggy did was help make sure the right people were in the right conversation at the right moment.

That’s a much more useful role for AI in hospitality than trying to remove people from the process. The goal is not fewer humans. It’s fewer good people spending their day answering the same repetitive questions.

Then something more interesting happened

Handling 20,000 customer conversations creates an interesting side effect: you end up with 20,000 pieces of customer feedback.

Not survey feedback. Not something somebody remembers from a quarterly meeting. Actual questions from actual people while they are trying to make a booking.

Those conversations leave clues everywhere. The same question appearing forty times in a week. A package people keep misunderstanding. A date everyone wants but struggles to find. A step in the booking process where customers repeatedly need help. A request the team thought was unusual that turns out to be fairly common.

One conversation is anecdotal. Twenty thousand conversations that’s data you make business decisions on.

Your inbox has probably been trying to tell you things

This became one of the more useful parts of Peggy for Clays.

Dan Marshall, Product Owner at Clays, explains:

“Now we’ve got this data and we’re really able to drill into it and understand it more and optimise it. And so we can now basically build our processes on the back of seeing the pain points from the customers.”

That’s a much bigger idea than simply answering enquiries faster.

Every venue already has customer research happening in its inbox every day. Most of it just isn’t treated that way. Someone answers a question, closes the thread and moves on to the next email, WhatsApp message or phone call.

Once those conversations are structured, you can start seeing patterns in your reporting instead. And once you can see a pattern, you can usually do something about it.

Clays has already done exactly that

Earlier in the collaboration, conversations handled by Peggy started exposing friction in the booking journey.

Guests were running into issues while selecting timeslots and completing bookings. Before, some of those customers might simply have disappeared. No completed form, no booking and no neat little record explaining what had gone wrong — just someone quietly leaving the website.

Because customers could ask Peggy for help, repeated problems became visible. One of them was surprisingly simple: guests sometimes forgot to select a timeslot, which created an error.

So the flow was changed to select the first available timeslot by default.

Other recurring questions found their way into the FAQ and booking journey as well.

Nothing revolutionary happened. It was simply a series of small operational improvements based on what customers were actually doing rather than what everyone assumed they were doing.

20,000 conversations beat a hunch

Venue teams usually know their guests very well, but there’s still a difference between saying, “I feel like people often ask about this,” and being able to see that a specific question came up 137 times last month.

The second one is easier to act on.

That’s where conversation data becomes useful beyond customer service. You can start asking which questions come up before a booking, where people get stuck, which requests tend to convert, which enquiries need a human quickly and what customers keep asking for that is still unclear on the website.

Those aren’t chatbot questions.

They’re venue management questions.

And yes, it also saves the team a lot of time

There is also the obvious benefit: Peggy answers a lot of questions.

Kimberleigh puts it like this:

“Peggy, our AI bot, already made a real impact. She’s handling a huge volume of customer conversations for us, taking pressure off the team, improving response times and making everything much more efficient.”

That’s the immediate value. Fewer repetitive conversations for the team, faster answers for guests and better information by the time an enquiry reaches sales.

But the longer Peggy works, the more useful the second layer becomes.

You don’t just get conversations handled. You learn from them.

There still needs to be a human behind the digital colleague

We talk a lot about digital colleagues, but there’s another part of the equation we care about just as much: the human colleagues helping build and improve them.

Dan says:

“Rookoo’s head of customer success, Thomas, always made us feel like a priority. From the get-go, he was proactive and had our best interests at heart, like he was one of our colleagues. Rookoo definitely lives up to what they promise.”

That matters because every venue works differently. Different packages, booking rules, spaces, audiences and sales processes mean a generic chatbot will only get you so far.

Peggy had to learn Clays.

And she keeps learning.

So, what has actually changed?

Clays now has a digital colleague handling customer conversations across multiple channels and around the clock. Straightforward questions can be answered without someone on the team needing to stop what they’re doing. Enquiries arrive with more context, promising conversations can be escalated while the customer is still engaged, and recurring customer problems become visible instead of disappearing into individual inbox threads.

Most importantly, those 20,000+ conversations have become something the team can use to improve the business.

And that gives Clays’ people more time to focus on the things humans are actually good at: hosting people, selling the right experience, solving unusual requests and creating something memorable.

Peggy can have the parking question.

Your venue’s conversations are probably telling you things too

You don’t need 20,000 enquiries before patterns start to appear.

They’re already there in your inbox, your WhatsApp messages, your contact forms and the questions your sales team can answer before the customer has even finished asking them.

The interesting question is what happens when you stop treating those conversations as admin and start treating them as information.

That’s what we’d like to show you.

Bring us your venue. We’ll show you what your own digital colleague could look like.