MAKIN AI copilot for customer support, drafting from your own tickets
What’s included
01- ✓Trained on your resolved tickets
- ✓Suggested replies with sources
- ✓Tone and policy guardrails
- ✓Agent feedback loop
- ✓Deflection and CSAT reporting
- ✓Helpdesk integration
The details
02Does it reply to customers on its own?
No. The copilot drafts; your agent reads, edits and sends. A person stays accountable for every reply, while the copilot does the searching and the first draft — finding a similar past answer and the relevant policy line. Each suggested reply shows the sources it drew on, so an agent can check a returns rule or a fee before sending rather than trusting the draft blindly.
If you want routine questions answered automatically, with no agent involved, that is a different package: WhatsApp & Support Automation, from ₹12,000, answers and routes customer messages on WhatsApp and hands the tricky ones to a person. Which of the two you need first is a good question for the scope call.
What does it learn from?
Your resolved tickets — the answers your team has already given — and your written policies. Tone and policy guardrails sit on top, to stop a draft offering something your policy does not allow or drifting into a voice your brand does not use. The agent feedback loop lets agents flag drafts that miss, which shows where the copilot needs better sources or tighter rules.
How will we know it is working?
Deflection and CSAT reporting is part of the package, so the judgement rests on customer satisfaction scores and ticket outcomes, not on how the team feels about a new tool. Read alongside the agents’ feedback on individual drafts, the reports show where the copilot is helping and where agents are still writing replies from scratch.
What is not included?
Helpdesk integration is included; the helpdesk itself and its licences are not. Language model usage is billed by the model provider, on top of this price. The copilot can only be as good as your ticket history and written policies, so if resolved tickets are thin or the rules live in people’s heads, raise it on the scope call before anything is priced.
If your team mainly needs answers from manuals, product documents and internal policies rather than from past tickets — or the same knowledge should be open to staff outside support — the Enterprise RAG Knowledge Assistant (from ₹1,75,000) fits better, with citations and permission filtering.
How it works
03- 0115-min callTell us what’s breaking. We ask the questions that change the price.
- 02Scope in writingA fixed quote and a written scope: what’s in, what isn’t.
- 03Build & reviewWorking software early, with review rounds on real screens.
- 04LaunchWe go live with you, train your team and hand over access.
- 05SupportFixes, changes and growth on a retainer, if you want one.
FAQ
04What does an AI support copilot cost?
The AI Support Copilot starts at ₹45,000 and is delivered in up to 30 days, including helpdesk integration, suggested replies with sources, tone and policy guardrails, the agent feedback loop and deflection and CSAT reporting. The language model’s usage charges are not in this price. Your fixed price arrives in writing after a 15-minute scope call.
Which helpdesk does it work with?
Helpdesk integration is one of the six items in the package. Which helpdesk you use, and whether its API gives the copilot access to your ticket history and a place to show drafts to agents, is checked on the scope call and written into the proposal before the build starts.
Will it make things up?
It can get things wrong, which is why it drafts rather than sends. Each suggestion lists the past tickets or policy lines behind it, guardrails are there to keep drafts inside your policies, and an agent reviews every draft before it goes out, so a wrong suggestion reaches a person before it can reach a customer.
How much ticket history do we need?
Enough resolved tickets to cover the questions you get most often. How much that is depends on how varied your queries are and how consistently they were answered, so it is assessed on the scope call. Where history is thin, written policies carry more of the load.
What changes for our support agents?
Less starting from a blank reply. A suggested draft appears with its sources; the agent checks it, edits what needs changing and sends it as their own. When a draft misses, they flag it rather than quietly rewriting it, so the misses are recorded instead of lost.
More in AI
0515 minutes.
One quote.
Tell us what you need. We reply on WhatsApp or email with a time, and a written quote after the call.
Re: AI Support Copilot