Support without the queueAI customer service for businesses and individuals
An AI customer service agent that resolves routine questions across chat, email and your help desk, follows your policies to the letter and hands real problems to your team.
WHAT IT DOESLOG
“Where’s my order?”
→ Tracked and answered in seconds
Return request
→ Checked against policy, started
Email in Spanish
→ Answered in Spanish
Unhappy customer
→ Passed to a person, summarised
“Monday morning: forty emails, half of them asking where an order is. The customers with real problems wait behind them.”
SOUND FAMILIAR?
What is AI customer service?
AI customer service means an AI agent handling the routine part of your support queue: answering questions about orders, deliveries, bookings, returns and accounts across chat, email and messaging, any time of day. It works from your own policies and systems, so the answers are specific and correct, and it hands anything complicated, sensitive or emotional to a person, with the conversation already summarised. Customers get answers in seconds; your team gets time for the problems that need them.
Tamkoz sets up AI customer service for businesses and individuals. We start by reading your real tickets, write the playbook the agent follows, connect it to your help desk and order system, and launch it gently: drafting replies for your team first, then answering on its own once it’s consistently right. If your help desk’s built-in AI does the job, we configure that instead.
THE JOBS IT TAKES ON
AI CUSTOMER SERVICE
“Where’s my order?”
→ Tracked and answered in seconds
Return request
→ Checked against policy, started
Email in Spanish
→ Answered in Spanish
Unhappy customer
→ Passed to a person, summarised
Monday inbox
→ Sorted by urgency before you log in
What AI customer service fixes
Most support queues are the same handful of questions, asked at the worst possible times.
Mornings lost clearing overnight emails
→ Routine questions answered overnight. The inbox you open is the one that needs you.
“Where’s my order?” fifty times a week
→ Order and tracking looked up and answered in seconds, on every channel.
Customers waiting hours for a simple answer
→ A reply in seconds, day or night, in the customer’s own language.
Different answers from different team members
→ Every reply follows the same written policy and tone of voice.
Angry customers lost in the queue
→ Frustration spotted early and passed to a person, with the history summarised.
No idea what customers keep asking
→ A weekly list of top questions and gaps in your help articles.
See it handle a support chat
Order tracking, a returns question and an unhappy customer, in under a minute.
Fern & Field
SUPPORT · REPLIES INSTANTLY
What just happened
A customer wrote in on a Sunday evening. Here’s what the AI agent did, and when it brought in a person:
SAMPLE CONVERSATION · FICTIONAL ONLINE SHOP
Who uses AI customer service?
Any business that answers the same customer questions every day, whatever the channel.
Online shops
Order status, delivery, returns and product questions handled around the clock.
Subscription and software businesses
Account, billing and how-to questions answered from your help centre.
Travel, events and hospitality
Booking changes, opening times and policy questions handled in any language.
Clinics, gyms and salons
Appointment changes, membership and pricing questions answered, with bookings updated.
Property and letting
Tenant questions answered and maintenance requests logged with the right details.
Service businesses
Job updates, quotes and invoice questions handled without a phone call.
Solo founders and creators
Customer emails answered while you’re making the thing they bought.
A GOOD FIT IF…
- Most tickets are the same few questions
- Customers write in at all hours
- Your team spends mornings clearing the inbox
- You use a help desk or a shared inbox today
What we set up
A support playbook
Your policies, tone of voice, limits and hand-over rules, written down and agreed with you.
An AI agent on your channels
Answering chat, email and messaging from one set of rules, inside your help desk or inbox.
Connections to your systems
Order, booking or account look-ups, so answers are specific and real.
Hand-over that works
A person takes over with the full conversation and a short summary, so customers never repeat themselves.
Drafts before autopilot
The agent drafts replies for your team first, and answers on its own only once it’s consistently right.
A weekly report
Resolved, handed over, top questions and missing answers, in plain English.
How we set up AI customer service
We start with your real tickets, not a demo script.
- ORDER 001START
1×Read your inbox
We sort a few weeks of your real tickets and messages by type, and work out which ones an AI agent could resolve and which must always reach a person.
- ORDER 002BUILD
1×Write the playbook
Your policies, tone of voice, refund and returns rules, and the moments to hand over, written down so the agent follows them exactly.
- ORDER 003CHECK
1×Connect and test
The agent is connected to your help desk, order system and knowledge base, then tested on past tickets before a single customer sees it.
- ORDER 004GO LIVE
1×Launch gently
It starts by drafting replies for your team to send, then answers routine questions on its own once the drafts are consistently right.
What it works with
We work inside the tools your team already uses, so nobody has to learn a new inbox.
- HELP DESKS
- ZendeskFreshdeskGorgiasHelp ScoutIntercomHubSpot Service Hub
- INBOXES AND CHAT
- GmailMicrosoft OutlookWebsite chatWhatsAppInstagramFacebook Messenger
- ORDERS AND BOOKINGS
- ShopifyWooCommerceStripeBooking systems
- KNOWLEDGE
- Help-centre articlesPolicies and PDFsNotionGoogle Drive
- AI MODELS
- OpenAIAnthropic (Claude)Google (Gemini)
Many help desks now include their own AI agent, often priced per resolved conversation. We compare that with a custom set-up and recommend whichever costs you less for the same result.
How much does AI customer service cost?
Setting it up is a one-off job we quote in writing. Running it is a monthly cost you pay directly: either your help desk’s AI add-on, or the AI model and platform behind a custom agent.
WHAT AFFECTS THE PRICE
How many channels
Email alone is simplest. Chat, WhatsApp and social messages each add set-up and testing.
How many systems it looks things up in
Order, booking and account look-ups make it far more useful, and each takes connecting.
How much it can do on its own
Answering questions is simpler than starting returns or changing bookings under rules.
Languages
Each language your customers use is tested before launch.
Volume
Per-resolution pricing suits low volumes. At higher volumes, a custom agent can cost less.
FROM THE PROVIDERS THEMSELVES
- Intercom prices its Fin AI agent at $0.99 per outcome (a conversation it resolves), with a minimum monthly commitment when used with another help desk. Source: Intercom.
- A custom agent pays for the AI model by the token instead. Anthropic’s smallest current Claude model costs $0.10 per million input tokens for prompts up to 100,000 tokens. Source: Anthropic.
CHECKED 10 OCTOBER 2026. PROVIDERS CHANGE THEIR PRICES, SO CHECK BEFORE YOU BUY.
YOUR QUOTE · FIXED AND IN WRITING
We quote set-up in writing, in your currency, and show you the monthly running cost both ways before you choose.
How long does it take?
Most set-ups are drafting replies within a few weeks and answering routine questions on their own soon after.
Good help articles make a good agent. If yours are thin, we’ll show you the gaps and can write them with you.
TYPICAL TIMELINE
YOUR QUOTE HAS THE REAL DATES
01 TICKET REVIEWABOUT A WEEK
We sort your recent tickets and agree what the agent should and shouldn’t handle.
02 QUOTE AND PLAYBOOKA FEW DAYS
Scope, price and the written rules the agent follows.
03 BUILD AND TESTTWO TO THREE WEEKS
Connected to your systems and tested on past tickets.
04 DRAFTS, THEN AUTOPILOTTWO WEEKS OR MORE
Your team approves its drafts until they’re consistently right.
AI customer service questions
Something else? Ask in the chat. A person answers.
What is AI customer service?
AI customer service uses an AI agent to answer customers’ questions across chat, email and messaging, and to complete simple tasks like tracking an order or starting a return. It works from your policies and systems, and passes anything complicated, sensitive or emotional to a person with a summary.
How is this different from a website chatbot?
A website chatbot is usually the front door: it answers questions from new visitors and books appointments. AI customer service works on your support queue: it reads emails and tickets, looks up orders and accounts, follows your policies, and helps your team resolve existing customers’ problems across every channel.
Will AI replace my support team?
In a small business, it usually replaces the repetitive part of their day, not the people. The agent clears the routine questions so your team has time for the customers who need a human, and every handover arrives with the history already summarised.
Which help desks does it work with?
Common ones include Zendesk, Freshdesk, Gorgias, Help Scout, Intercom and HubSpot, plus a shared Gmail or Outlook inbox. Many help desks now include an AI agent of their own; if yours does the job, we’ll set that up rather than build something new.
Can it look up orders and accounts?
Yes, if your systems allow it. We connect it to your shop, order or booking system so it can give real answers, like where an order is, rather than generic ones. What it may change, such as cancelling an order, is limited by rules you set.
What about refunds and complaints?
You decide. A common set-up is that the agent can start returns that clearly fit your policy, while refunds above a set amount, complaints and anything unusual always go to a person.
Can it answer in other languages?
Yes. Modern AI models can read and reply in many languages, so customers can write in their own language while your policies stay in yours. We test the languages your customers actually use.
Should customers be told they’re talking to AI?
We recommend it, and in some places the rules require it. Customers mostly care about a fast, correct answer and an easy way to reach a person, so we make both obvious.
How do I know it’s working?
You get a simple report: how many conversations it resolved, how many it handed over and why, and the questions it couldn’t answer. That last list tells you exactly what to add to your help articles.
Further reading
GOES WELL WITH
- AI assistants & chatbotsA chat assistant on your website, WhatsApp, Instagram or Facebook that answers customers in seconds, in your words, day and night.
- RAG & AI knowledge basesA private AI assistant that answers from your own documents, policies and data, shows where each answer came from, and keeps your information yours.
- AI agentsCustom AI agents that read, decide and act across your tools: they research a lead, update your CRM, draft the reply and ask you before anything important goes out.
SOURCES · CHECKED 10 OCTOBER 2026
Answer every customer, first time.
Send us a rough idea of your support volume and the questions you get most. We’ll show you what an AI agent could take off your team, and quote in writing.