Aura — the autonomous AI
customer support agent for SaaS
Resolves up to 60% of conversations autonomously using your help center and product knowledge. Executes multi-step actions through AI Missions. Escalates to humans with full context. Included on every Customerly paid plan from €19/seat/mo — no per-resolution charge.
What is an AI customer support agent?
An AI customer support agent is an autonomous AI system that resolves customer conversations end-to-end — not a chatbot, not an agent-assist layer, not a deflection widget. It uses large language models grounded in your help center and product knowledge, reasons through unscripted questions, executes multi-step actions on your customer's behalf, and escalates to humans only when it's uncertain.
The category emerged in 2023-2024 alongside Intercom Fin, Ada, and Decagon. By 2026 it's the default first line of customer support resolution for any SaaS team handling more than 500 conversations per month.
Aura is Customerly's AI customer support agent. Built into the Customerly platform from day one, Aura is included on every paid plan — no per-resolution charge, no enterprise contract, no professional services engagement required to deploy.
How Aura works
Trains on your knowledge
Connect your help center, product docs, and past conversation history. Aura is live and answering questions in hours, not weeks.
Resolves end-to-end
Aura answers questions, executes multi-step AI Missions (refund, lookup, escalation), and confirms resolution with the customer.
Escalates with full context
When Aura's confidence drops below threshold, it hands off to a human with the full conversation, customer account data, and reasoning trail.
How Aura compares to other AI customer support agents
| Capability | Customerly Aura | Intercom Fin | Zendesk AI Agents | Ada |
|---|---|---|---|---|
| Autonomous resolution | Yes | Yes | Yes | Yes |
| Pricing model | Per seat (included) | Per resolution (~$0.99) | Per resolution | Custom contract |
| Time to deploy | Hours | Days | Days-weeks | Weeks-months |
| Multi-step actions | AI Missions | Fin Actions | Actions | Yes |
| Help center grounding | Yes | Yes | Yes | Yes |
| Self-serve setup | Yes | Yes | Partial | No (PS-led) |
| Bundled platform (chat + email + surveys) | Yes | Add-ons | Add-ons | No |
See detailed head-to-head comparisons at /compare.
AI customer support agent FAQ
The questions every SaaS support leader asks before deploying an AI agent.
An AI customer support agent is an autonomous AI system that resolves customer questions end-to-end — not just suggesting replies for human agents, but handling the full conversation, executing multi-step actions (lookup an order, trigger a refund, schedule a callback), and escalating to humans when it's uncertain. Aura is Customerly's AI customer support agent, included on every paid plan.
A chatbot follows scripted decision trees and falls back to humans the moment a question goes off-script. An AI customer support agent uses large language models grounded in your help center and product knowledge to handle unscripted questions, reason through multi-step problems, and execute actions on your customer's behalf. The user experience is conversational; the back-end is autonomous reasoning, not rule trees.
Aura typically resolves 40-60% of incoming conversations autonomously for SaaS teams. TeamSystem, a Customerly customer with 600,000 users and 15 agents, runs at a 60% resolution rate in production. Real-world resolution rate depends on knowledge base coverage, conversation volume, and prompt training more than the underlying model.
Pricing models vary widely. Per-resolution pricing (Intercom Fin at ~$0.99/resolution, Zendesk AI Agents) scales linearly with volume. Per-seat pricing with AI included (Customerly Aura at €19+/seat/mo) is flat regardless of volume. Custom enterprise contracts (Ada, Decagon) usually start in the high five figures annually. For SaaS teams scaling resolution volume, per-seat with included AI is typically the most predictable.
Yes. Modern AI customer support agents support multi-step actions: looking up an order, triggering a refund, updating a subscription, escalating a complaint, scheduling a callback. In Customerly this is called AI Missions — you define the workflow in natural language and Aura executes it autonomously when the right intent is detected.
For self-serve platforms like Customerly, Aura is live in hours: connect your help center, configure tone of voice, and Aura starts handling conversations the same day. Enterprise-grade vendors (Ada, Decagon) typically require multi-week professional services engagements before the AI agent goes live.
Aura escalates whenever it's uncertain about an answer, when a customer requests a human, or when an issue requires actions outside its configured AI Missions. The full conversation context, the customer's account data, and Aura's confidence reasoning are handed off to the human agent so there's no repeat-yourself moment.
Aura is grounded in your help center and product knowledge using retrieval-augmented generation, which means every response cites source content rather than inventing facts. You can review every AI response in the admin panel, and Aura escalates to humans whenever its confidence drops below a configurable threshold rather than guessing.
