Customer support
An AI customer support chatbot that resolves, not deflects
Deflection is a word support software uses when it means the customer gave up. Paperchat is built to actually finish the conversation: it trains on your help content, connects to the systems where the answers live, and looks things up while the customer is still typing. When it can answer, it answers with specifics. When it cannot, it says so plainly and passes the conversation to a person with the full history attached, in the same thread the customer is already in. It works on your website and in Telegram, Instagram and Messenger, and every conversation lands in one searchable inbox.
- Answers from your own content
- Hands over with full history
- 58+ languages automatically
I was charged twice this month.
I can see two charges on 3 March, one of which failed and was reversed the same day.
Only one payment was taken
My bank still shows both.
That is a pending authorisation. I am passing this to the team to confirm with your bank reference.
Trained on what your team already knows
Support quality depends on the knowledge behind it. Paperchat trains on your help centre, website, uploaded documents and any plain text you paste in, which means the internal notes that never made it into a public article can still inform the answers customers get.
Because the knowledge base is yours, the agent's tone and detail match your business rather than a generic support voice. When a policy changes, retrain that one source and every channel starts giving the new answer immediately.
Website and help centre
Crawled from your address or sitemap
Documents
PDFs, guides, terms and specifications
Plain text
The answers that were never written up
Connected systems
Live data looked up during the conversation
It can look things up, which is what resolution requires
Most support questions cannot be answered from documentation alone, because they are about one specific customer: their order, their invoice, their account. Paperchat connects to the systems that hold those records and queries them mid conversation, showing the customer a short status pill while it works.
That is the difference between a chatbot that explains your refund policy and one that tells a customer their refund was issued on Monday. Connections include Shopify, WooCommerce and WordPress on the free plan, with Stripe, Zapier, HubSpot, Pipedrive, ClickUp and others on paid plans.
Shopify and WooCommerce
Orders, products, stock and policies
Stripe
Payments, refunds and subscriptions
HubSpot and Pipedrive
Contacts, deals and history
Zapier
Thousands of other apps through one connection
Put this on your site today
Train a chatbot on your own content, connect what it should be able to do, and have it answering in the time it takes to make coffee. The free plan needs no card.
Handover that does not restart the conversation
The worst moment in automated support is the handover, where the customer is asked to explain everything again to a person who cannot see what just happened. Paperchat moves the same thread to your team: the customer keeps typing in the same window, and your agent sees the entire history including anything the bot looked up.
You decide when handover is offered, including after a set number of messages, and your team is notified in Slack or Telegram so nobody has to sit watching a dashboard.
- 1The customer asks for a person, or the agent decides it cannot help
- 2The conversation switches to staff mode in the same thread
- 3Your team is notified in Slack or Telegram
- 4An agent replies with the full transcript in front of them
One inbox behind every channel
Support arrives wherever the customer happens to be: the website widget, an Instagram DM, a Messenger reply, a Telegram message. Running a separate tool for each is how conversations get lost and how the same customer gets three different answers.
Paperchat runs the same trained agent everywhere and puts every conversation in one place, with search, status and assignment. You can see what has been asked this week, what the agent handled and what needed a person.
Hi, is the studio open on bank holiday Monday?
We are open 10am to 4pm on bank holidays, with a reduced class timetable.
Can I move my Tuesday booking to then?
I will pass this to the team so they can check availability and move it for you.
Frequently asked questions
What to look for in an AI customer support chatbot
The market for AI customer support chatbots has filled up quickly, and most of the products in it do a version of the same thing: ingest your help centre, embed it, and answer questions from the resulting index. That is genuinely useful, and it is also where most tools stop. The limitation shows up the first time a customer asks something that depends on their own record rather than on your documentation, which in practice is most of the questions a support inbox receives.
So the first thing worth checking is whether the tool can look anything up. A support agent that can query your store, your payment processor or your CRM during the conversation can close a ticket that a documentation-only bot can only describe. Paperchat does this through connected systems, showing the customer a short status while it works, then answering with the specific detail: this order shipped on Thursday, this refund was issued on Monday, this plan includes that feature.
The second thing to check is the handover. Automation that cannot admit defeat gracefully damages more goodwill than it saves, and the classic failure is dumping a frustrated customer into a contact form after four unhelpful replies. A good handover keeps the same thread, notifies a human immediately and gives that human the entire history, including what the bot looked up. The customer should notice only that the answers got better.
Third is coverage. Support does not arrive on one channel any more. A customer who found you on Instagram will ask there, someone mid purchase will use the website widget, and a repeat buyer may message on Telegram. Running separate tools for each guarantees inconsistent answers and lost history. One trained agent across every channel, with one inbox behind it, is both cheaper and more coherent.
Finally, look closely at pricing, because the model shapes how you use the tool. Per seat pricing punishes you for involving your team, which is exactly backwards when the goal is smooth escalation. Paperchat charges by message credits, one per answer, so cost follows conversation volume and adding a colleague costs nothing. The free plan includes 100 credits and the store connections, which is enough to put a trained agent in front of real customers and judge it on the transcripts rather than on a sales demo.
More ways to use Paperchat
Platforms
AI chatbot for Shopify
Answer product, order and policy questions straight from your Shopify catalogue.
Read moreUse cases
Lead generation chatbot
Answer the question first, then capture the contact while interest is high.
Read moreIndustries
AI chatbot for SaaS
Answer product and billing questions from your docs, and cut repeat tickets.
Read morePricing
Start on the free plan with no card. One credit is used per answer, so you only pay for conversations that actually happen.
Free
$0per month
- 100 message credits (~20+ conversations)
- 1 AI chatbot
- 800 KB knowledge base
- 1 team member
- Shopify, WooCommerce & WordPress actions
Basic
Most popular$35per month
- 4,000 message credits (~1000+ conversations)
- 1 AI chatbot
- 25 MB knowledge base
- 3 team members
- Conversation history
Pro
$99per month
- 10,000 message credits (~2500+ conversations)
- 3 AI chatbots
- 40 MB knowledge base
- 4 team members
- Conversation history
Larger plans and a lifetime option are on the pricing page.