AI Voice for Kennels and Catteries: Boarding Bookings

A boarding kennels or cattery lives and dies by the phone in its peak weeks: the run-up to the summer holidays and the days before Christmas. That is exactly when the owner is on site with both hands full, hosing down a run or settling a nervous new arrival, and cannot pick up. Every missed call in that window is a multi-night booking that walks straight to the site down the road. The problem is not that the calls are complicated. It is that they arrive in a flood, out of hours, and each one carries checks a diary slot never does.

Those checks are what make boarding different from a haircut or a table for two. Before an animal can stay, its vaccination record has to be current, its feeding and medication needs have to be written down, and someone has to decide whether it can share space or needs an isolated pen. Get any of those wrong and you have a welfare incident, a licence problem, or a very unhappy owner at collection. This guide is about how a voice AI agent handles that enquiry properly: what it should capture, what the law actually requires, and the line it must never cross.

The wider context is that most businesses are now automating something. In its November 2025 report The State of AI, McKinsey found that the large majority of organisations now use AI in at least one function, yet only about a third have moved it into production. Boarding is a place where the move to production is unusually clean, because the enquiry is structured and repetitive, and the risky judgements can be fenced off and left to a person.

This guide is shipped by the team behind Dilr Voice, enterprise voice AI built for regulated deployments. Or see DATS, our five-stage AI consulting system for placing agents where they actually pay back.

What does an AI voice agent do for a boarding kennels or cattery?

A Dilr Voice agent answers the boarding line when the owner cannot, day or night. It takes the enquiry, confirms the dates and the drop-off and collection windows, records each animal's feeding, medication and temperament details, states the vaccination-record requirement, and books the stay against pen or run capacity. Welfare worries, sick animals and complaints are routed to a named person. It captures, it does not judge.

Think of it as the front-of-house that never sleeps and never forgets a question. The value is not that it replaces the owner's judgement about which dogs can share a paddock. It is that by the time a human looks at the booking, every fact needed to make that judgement is already on the record: the breed and size, the neutering status, the feeding routine, the medication times, and whether the owner has flagged reactivity around other animals. For a single-site cattery that runs on a whiteboard and a mobile, this is often the first time enquiries stop leaking, and it is the sort of narrow, well-fenced deployment Dilr.ai was built to ship.

Why is a boarding enquiry harder than a single appointment?

A boarding enquiry is a multi-night reservation, not a slot. The agent has to hold dates across drop-off and collection windows, capture per-animal feeding and medication schedules, note temperament and whether the pet can share or needs isolation, and flag the vaccination-record requirement before anything is confirmed. A dog groomer's appointment reserves one grooming slot. A boarding booking reserves a pen for a week and takes on a duty of care for a living animal the whole time.

That duty is why the checks matter. The UK has a large base to serve: UK Pet Food estimated in 2026 that there are around 15.5 million pet dogs and 13 million pet cats in the country, with 62 per cent of households owning a pet of some kind. Kennels board the dogs and catteries board the cats, and every one of those stays needs the same structured intake. A voice agent that treats a boarding call like a restaurant reservation will miss the things that keep the site licensed and the animals safe.

Because the intake is structured, it maps cleanly onto a scripted flow. That is the same reason our approach to placing AI tends to score booking-heavy trades highly: the work is repetitive, the exceptions are nameable, and the handoff to a person is easy to define.

Why must vaccination proof be checked before a stay?

In England, a licensed boarding site must be able to show an inspector an up-to-date vaccination record for every dog or cat it boards, under the statutory guidance attached to its licence. The vaccination record is the gate that protects every other animal on site. A Dilr Voice agent states that requirement on the booking call and logs it, so no owner arrives at drop-off unaware that it will be checked.

The requirement is worth quoting, because it is precise. The statutory dog kennel boarding guidance, which tells licensing inspectors what an operator's records must show, states that "An up-to-date veterinary vaccination record must be seen to show that dogs, including resident dogs, have current vaccinations against" the core diseases named by the Department for Environment, Food and Rural Affairs. For dogs the guidance names canine parvovirus, canine distemper, infectious canine hepatitis (adenovirus) and leptospirosis, and notes that protection against kennel cough may also be required. For cats, the equivalent cat boarding guidance names feline panleukopenia, feline infectious enteritis and the feline respiratory viruses. A recent titre test certificate may be accepted in place of a booster in some cases, which is a decision for the site and its vet, not the phone line.

So the agent does one precise thing here: it tells the owner, on the booking call, that an up-to-date vaccination record will need to be seen at drop-off, and it logs that this was stated. It does not accept a photo, it does not judge whether a lapsed booster still counts, and it does not admit the animal. That check happens in person, by a member of staff, against the physical record. Naming the requirement early is the single biggest thing a voice line can do to stop a doorstep argument on the busiest morning of the year.

How does the agent capture feeding, medication and temperament for each pet?

A Dilr Voice agent takes each animal separately and builds a per-pet record, because a two-dog booking is two care plans, not one. For each animal it captures the name, breed and size, the feeding routine, any medication with doses and times, neutering status, and temperament flags such as reactivity to other dogs or anxiety when kennelled. It reads the details back to confirm, then attaches the record to the booking for staff.

Two of those fields do more work than the rest. Medication times drive the daily routine on site, so a voice AI agent captures them as structured data rather than a free-text note, the same discipline our voice AI address capture work applies to postcodes: a field that will be acted on has to be captured cleanly, not transcribed loosely. Temperament and neutering status feed the next decision, which is whether the animal can share space at all.

How does the agent handle pen or run compatibility and the welfare handoff?

It does not decide compatibility; it gathers the facts that let a person decide, and hands over the moment a welfare question appears. When an owner says a dog is reactive to others, that two household dogs should share, or that a cat needs a quiet block, a Dilr Voice agent records the flag and marks the booking for staff review rather than arranging a pen itself. A sick or distressed animal goes straight to a person.

This is where the "capture, do not judge" rule earns its keep. Deciding whether a nervous rescue dog can be kennelled next to a boisterous spaniel is a welfare judgement with a duty of care behind it, and it belongs to trained staff. The agent's contribution is that the staff member making the call has the neutering status, the temperament history and the household grouping in front of them before they pick up, instead of ringing the owner back to ask. For genuinely clinical questions, such as whether a pet is well enough to travel or board, the right destination is a vet, which is the boundary our AI voice for veterinary practices guide draws in detail.

The same handoff discipline underpins our AI operating model consulting, which is mostly about defining, in advance, exactly which calls a machine finishes and which a human must own.

What licence and rules apply to a UK boarding kennels or cattery?

In England, boarding is a licensable activity under the Animal Welfare (Licensing of Activities Involving Animals) (England) Regulations 2018. Schedule 1 covers boarding for cats, boarding in kennels for dogs, home boarding and dog day care, and Schedule 4 sets separate specific conditions for each. The local authority issues the licence and inspects the premises, and the duty to meet the conditions sits with the licence holder, meaning the operator, not the voice line.

The Animal Welfare (Licensing of Activities Involving Animals) (England) Regulations 2018 also attach specific welfare conditions. Schedule 1 sets out the licensable activities, and under regulation 4 the authority appoints a suitably qualified inspector, must take account of whether the applicant is a fit and proper person, and attaches both general and activity-specific conditions to any licence it grants. Schedule 4 requires each boarding site to run a preventative healthcare plan agreed with a registered veterinarian and to keep vaccination, worming and treatment records. A voice agent does not change any of this; it simply makes sure the enquiry collects the information those conditions assume the site holds.

One scope note matters, because the slug says UK and the law does not. These regulations are England-only. Wales, Scotland and Northern Ireland run their own animal-activity licensing regimes, so an operator over a border should check the rules that apply to their nation. The data the agent collects is governed UK-wide: the owner's contact details and the pet's records are personal data under UK GDPR, and a booking confirmation sent afterwards is a service message rather than marketing, a distinction the ICO draws under PECR. A pet's medical details are not, by themselves, special category data about a person, so there is no need to over-engineer the consent flow, though the usual security and retention discipline still applies.

What must the agent never decide on a boarding call?

It must never admit an animal, waive the vaccination requirement, or make a clinical or welfare ruling. A Dilr Voice agent states that a current vaccination record will be checked at drop-off; it does not accept proof by phone or judge a lapsed booster. It records temperament and grouping flags but does not confirm two animals can share, and it quotes the deposit and cancellation terms the site has set without ruling on whether a charge applies.

Drawing that line clearly is the whole safety case. Everything the agent finishes is a fact-gathering or booking step with a clear right answer: dates, contact details, per-animal care notes, the stated vaccination requirement, and a provisional booking against capacity. Everything with judgement or duty of care attached, an admission decision, a compatibility decision, a clinical decision, or a disputed charge, is escalated to a person with the record already prepared. That is the difference between an agent that reduces risk and one that quietly creates it, and it is the first boundary we set when we stand up an AI execution office for an operator.

How does it handle deposits, cancellations and peak-holiday demand?

It takes the booking and quotes the terms; it does not invent them. A Dilr Voice agent applies the deposit and cancellation policy the site already publishes, states it plainly to the caller, and records that it was stated. What counts as a fair charge, and how consumer cancellation rights apply to a dated booking, are set by the site's own terms and by consumer law, not by the phone line.

The mechanics of deposits and seasonal cancellation policy are covered in depth in our AI voice for holiday parks guide, and the consumer cancellation-rights position in our distance selling and cooling-off rights explainer. Peak demand is where a voice line pays for itself. School holidays and bank holidays produce a spike of out-of-hours enquiries that no single owner can answer while running the site, and the same spike creates a wave of no-shows and last-minute changes on established bookings. An agent absorbs both: it books new stays around the clock and processes changes and reminders as service messages, the same pattern our AI voice for restaurant table bookings guide describes for covers and deposits. The result is that the peak weeks that used to overwhelm the phone become the weeks the site captures the most.

Want to see this in production? Try Dilr Voice live, book an AI placement diagnostic, see our DATS methodology, or read about our approach to placing AI inside real operations.

What is the best voice AI setup for a UK kennels or cattery in 2026?

The best setup depends on call volume and how much the site cares about the welfare handoff. For a single-site cattery with a genuinely quiet line, a simple call-back service or a shared inbox may be enough, and an agent is over-buying. For a busy kennels, a multi-site group, or any operator losing peak-season bookings to voicemail, a purpose-built voice agent that captures the full boarding intake and escalates cleanly is the stronger choice.

Among the platforms, the self-serve builders such as Vapi, Retell AI, Bland AI and Synthflow let a technical operator assemble a booking bot quickly and cheaply, and for a low-stakes line that can be the right call. Where the boarding case gets harder, the escalation logic, the per-animal records, the licence-aware scripting and the integration with the kennel-management software the site already runs, a managed platform such as PolyAI or Dilr Voice tends to fit better, because the vendor builds and tunes the deployment with you rather than handing you a toolkit. The systems worth connecting to are the ones the site already lives in: its booking or kennel-management software, its calendar, and its customer records, wired up so the agent writes into the same place staff read from. That integration work is the core of our DATS methodology.

Frequently asked questions

Can the agent tell an owner whether their pet is well enough to board?

No. Whether an animal is fit to board or needs to see a vet is a clinical judgement, and a Dilr Voice agent is built not to make it. If an owner describes a sick, injured or distressed animal, the agent routes the call to a person and, where appropriate, points the owner towards veterinary advice. Clinical triage on the phone belongs to a veterinary practice, a boundary set out in our veterinary voice AI guide.

Does the agent work for both catteries and dog kennels?

Yes, with the right details for each. A Dilr Voice agent runs the same structured intake for a cattery and a kennels but uses the correct vaccination list for the species and the correct pen or run language. Home boarding for dogs and dog day care are covered by their own specific conditions, so an operator offering those should have the agent scripted for each rather than reusing one flow.

Written by the Dilr.ai engineering team, practitioners who ship enterprise AI in production. Follow us on LinkedIn for shipping notes, or subscribe via the RSS feed. Explore more in the industries collection.