Artha Fintech — Support System Overview

Section 01

What is a Support System?

A support system is the complete infrastructure that connects customers who have a problem with the people and processes that can solve it — reliably, quickly, and without losing track.

🎫

Ticket Management

Every customer problem becomes a ticket — a tracked record with a unique ID, a full conversation thread, attachments, priority, and status. Nothing falls through the cracks. Every interaction is logged, searchable, and auditable.

⏱

Service Level Agreements (SLAs)

SLAs define how fast a team must respond and resolve. A critical banking issue might demand a response within 1 hour. The system automatically tracks these deadlines, warns agents before a breach, and alerts supervisors when one occurs.

🔀

Routing and Escalation

Tickets are automatically assigned to the right agent based on category, load, or skill. When a junior agent can't resolve something, the system escalates it up the chain — with a clear reason, a timestamped record, and the right people notified.

📊

Visibility and Reporting

Managers see the big picture: how many tickets are open, which agents are overloaded, where SLAs are failing, and what customers are most frustrated about. Reports are generated automatically and delivered on schedule.

💡

Why fintech needs a specialised support system

Banking support is not like e-commerce support. A failed UPI transfer or a frozen account is urgent — sometimes legally time-bound. Agents need to see account balances, transaction IDs, and KYC status alongside the conversation. Data must stay isolated per client. Responses may need to be white-labeled. Regulators require 7-year audit trails. A general-purpose help desk doesn't handle any of this well.

Without a support system	With a support system
Customer emails support@ and it gets lost in a shared inbox	Email is parsed into a ticket automatically, assigned to the right agent, SLA clock starts
No one knows which agent is handling what	Every ticket has one owner, visible to the whole team in real time
Agent replies from personal email — no record kept	Every reply is logged in the ticket thread, timestamped, auditable
A critical issue sits unresolved for 3 days — discovered by accident	SLA watcher fires an alert at 80% elapsed time; supervisor notified at breach
Client asks "what happened with ticket X?" — nobody knows	Full audit trail: who did what, when, with what reason
Monthly report takes 2 days to compile manually in Excel	Reports generated in seconds, scheduled and emailed automatically

Section 02

About the Artha Support System

Artha Fintech is two things at once: a neo-bank serving its own customers, and a white-label platform powering other fintech companies. The support system had to work for both — without mixing data, without separate logins, without rebuilding things twice.

Stream A — Artha Direct

Artha's own neo-bank customers

When an Artha customer has a problem — a UPI transfer didn't go through, their card was declined, KYC is stuck — they reach out through the mobile app, email, WhatsApp, or phone. The ticket lands in the Artha workspace.

Artha's own support agents handle these tickets with full access: they can see the customer's account balance and transaction history, reply, reassign, escalate internally, merge duplicates, and close. The full lifecycle is in one place.

Stream B — White-Label Clients

FinCo Payments, SwiftBank, and others

Artha's platform clients run their own support teams. Their customers use a branded portal at support.[client].io and log in with their existing Auth0 account — no new signup. Client agents handle tickets in their own workspace.

Artha watches these tickets in read-only mode. If a client can't resolve something, they escalate it to Artha. Artha responds — but always in the client's brand. The end customer never sees "Artha Fintech".

🔐

How Auth0 portal authentication works

Each client configures their own Auth0 tenant. When a customer opens support.finco.io and logs in, the system reads their Auth0 token. If the token contains a support_agent role, they get the agent view (all workspace tickets). If not, they get the customer view (their own tickets only). Internal notes are never visible to customers — the API enforces this regardless of what token is presented. Agents are auto-created in the system on their first login — no separate invitation process needed.

Who uses it

🛡

Artha Admin

Manages everything: all workspaces, all clients, all agents. Configures Auth0 per client. Full system access.

👁

Artha Supervisor

Oversees all Artha and client tickets. Sees cross-client health dashboard. Handles escalated tickets. Configures SLAs.

✏

Artha L2 Agent

Handles complex Artha tickets and escalated client tickets. Can see all client workspaces in read-only mode.

🏢

Client Admin

Full control of their own workspace. Can escalate tickets to Artha. Manages their own agents and SLA config.

🎧

Client Agent

Handles tickets in their own workspace. Authenticates via Auth0. Auto-provisioned on first login. No Artha access.

👤

Customer

Logs in at support.[client].io with their existing account. Sees only their own tickets. Cannot see internal notes.

How the key features work together

Ticket Lifecycle

1Customer raises issue via app, email, WhatsApp, phone, or API
2Ticket created automatically with unique number, SLA clock starts
3AI classifies category, priority, and sentiment instantly
4Auto-assigned to agent by category/load/skill routing rules
5Agent replies using AI draft, canned responses, or their own words
6Status changes: Open → Pending → On Hold → Resolved → Closed
7CSAT survey sent 5 minutes after resolution

Client Escalation to Artha

1Client agent cannot resolve — too complex or needs Artha's access
2Client Admin clicks "Escalate to Artha" with reason + note
3All Artha supervisors notified instantly (push + email)
4Ticket appears in Artha's escalation queue in real time
5Artha agent takes ownership and works the ticket
6Reply sent to customer in client's brand — not Artha's
7Client Admin notified when Artha resolves it

SLA Engine (Automated)

1SLA policy applied at ticket creation (priority-specific targets)
2Background service evaluates every open ticket every 5 minutes
3At 80% of time elapsed: agent receives "At Risk" push alert
4At 100%: breach alert fires to agent + supervisor (immediate)
5On Hold status: SLA clock pauses automatically
6When On Hold clears: clock resumes from where it paused
7SLA compliance tracked per workspace in reports

Artha Monitoring Client Workspaces

1Artha supervisors see all client workspaces in one dashboard
2Per-client: open count, SLA compliance %, CSAT, escalations
3Artha can read all client tickets — cannot reply unless escalated
4Artha can add internal notes visible to client admin only
5AI anomaly detection: volume spikes, SLA degradation, surges
6Artha generates client health reports and shares with client admin
7Client data strictly isolated — FinCo never sees SwiftBank data

🤖

AI features (powered by Claude)

The system uses Claude to draft contextual replies for agents (in the client's brand voice if needed), auto-classify incoming tickets for category and priority, suggest relevant knowledge base articles, generate KB articles from resolved tickets, write SLA risk briefings for supervisors, and detect anomalies across all client workspaces. Agents review and approve AI output — AI assists, it does not act alone.