# Aide — Governed agentic AI that resolves customer support issues end to end inside Zendesk, Front and Gorgias.

> Source: The Agents Index — https://theagentsindex.com/aide (structured, researched, re-verified)
> Facts last verified: 2026-09-03

Aide is an agentic AI platform for customer support that classifies each ticket by intent, reads live order and CRM data, and executes the resolution inside an existing helpdesk. Teams deploy it one approved intent at a time.

| Fact | Value |
| --- | --- |
| Website | https://aide.app |
| Price | from $3000/month billed monthly, $3000/month billed annually |
| Free tier | No |

No free plan, trial or self-serve signup. The cheapest published tier is Growth at $3,000/mo, and every plan CTA routes to a demo booking at /demo.

## What it does

**Done-for-you service** · Partly — You run the software, but Aide's own people do the implementation and, on Scale and above, build custom workflows for you, so it sits between tool and delivered service.

> Priority support + custom workflow builds
> — https://aide.app/pricing · Plan cards, Scale ($8,000/mo) feature bullet

## How much it does unattended

**Runs autonomously** · Partly — Full autonomy is opt-in per intent and gated on a confidence threshold: anything below it, or any failed condition, escalates to a human rather than acting.

> Automation is deployed per intent at the level you choose: none, human-in-the-loop, or fully agentic.
> — https://aide.app/security · FAQ, 'How does Aide prevent unauthorized AI actions?'

**Agent permissions** · Yes — Permissions are enumerated per intent as a Scenario: the team fixes the conditions and the allowed actions, and anything outside escalates instead of firing.

> Every automation is a Scenario: conditions your team sets, actions your team allows, and a confidence gate it must clear.
> — https://aide.app/ · Scenarios section, lead paragraph

_Not established: Multi-agent_

## Which models it runs on

**Model choice** · No — The vendor picks the models and frames that as an advantage; no page offers a buyer any model selection, and the pricing FAQ ties cost control to running their own.

> Because we run our own models, we can afford to fix the root cause instead of farming the recurrence.
> — https://aide.app/ · Homepage FAQ, 'How does pricing work?'

_Not established: Claude, GPT, Open models_

## Where you use it

**In a browser** · Yes — Besides living inside the helpdesk, Aide ships a customer-facing chatbot for your own website, included from the entry Growth plan upward.

> Automated chatbot on your website
> — https://aide.app/pricing · Plan cards, Growth ($3,000/mo) feature bullet

## Whose machine it runs on

**Self-hosted** · Partly — Not customer-hosted software: the default is Aide's Azure tenancy, with private-cloud provisioning offered only where residency rules force it.

> Aide can be provisioned in public or private clouds for sovereign AI deployments where data residency requirements demand it.
> — https://aide.app/security · Data protection section, 'Private cloud deployments'

_Not established: Open source_

## What it costs to run

**How it meters** · Partly — The vendor's own pages disagree: the pricing page sells flat monthly platform tiers not metered by seats, while the homepage FAQ describes a platform fee plus a per-resolution rate and an implementation fee.

> Outcome pricing, done right: a platform fee plus a per-resolution rate that falls as you scale, plus a one-time implementation.
> — https://aide.app/ · Homepage FAQ, 'How does pricing work?'

**Free tier** · No — No free tier, trial, or self-serve signup: the cheapest published plan is $3,000/mo and every plan button routes to a demo booking.

> $3,000/mo
> — https://aide.app/pricing · Plan cards, Growth column price cell

**API access** · Partly — API access is gated: it is a listed capability of the $15,000/mo Operations tier and Enterprise, and is absent from the Growth and Scale columns of the same comparison table.

> Operations and Enterprise plans include API access and custom integration support.
> — https://aide.app/pricing · Pricing FAQ, 'What integrations are supported?'

**MCP server** · Partly — Aide speaks MCP outward as a client, connecting to your backends to gather context; nothing on the site says another agent can query Aide over MCP.

> Aide connects to your TMS, billing system, ERP, or custom back office via API or MCP, so answers reflect what your systems actually know.
> — https://aide.app/integrations · Closing section, 'Missing an integration?'

_Not established: Bring your own key_

## Buying it for a team

**A company can buy it** · Yes — This is an organisation-only purchase: annual contracts are the default, plans start at $3,000/mo, and pricing is scoped to the partnership rather than to an individual.

> Pricing is based on the level of technical needs and partnership you need, not seats. Annual contracts are our default — monthly billing is also available.
> — https://aide.app/pricing · Pricing FAQ, 'How is pricing determined?'

**Seat model** · Yes — There is no seat model to negotiate: every published tier includes unlimited seats, so headcount changes never move the bill, and pricing moves on tier instead.

> Every plan includes unlimited seats.
> — https://aide.app/pricing · Pricing page intro paragraph above the plan cards

**Pooled budget** · Yes — With unlimited seats on a flat account-level platform fee, the whole organisation draws on one shared entitlement; no per-user balance or allocation exists to divide.

> Every plan includes unlimited seats.
> — https://aide.app/pricing · Pricing page intro paragraph above the plan cards

**Admin controls** · Yes — Constraining the agent is the core of the product: it runs only team-written procedures, only within the account state configured, and escalates everything else.

> Aide executes only the procedures your team writes and approves, presents only the options configured for the account state, acts only when it is certain, and escalates the rest to a person with full context.
> — https://aide.app/security · FAQ, 'How does Aide prevent unauthorized AI actions?'

**Audit log** · Yes — The Action Trace logs every automated decision with the intent, data read, Scenario, confidence and outcome, per conversation and across the operation, and is exportable for review.

> The Action Trace records every automated action: which intent was recognized, which procedure ran, and what was done on the account.
> — https://aide.app/security · FAQ, 'Can our compliance team audit every automated action?'

_Not established: Single sign-on_

## What happens to your code

**Opt out of training** · Yes — Nothing to opt out of: model providers operate under zero-data-retention agreements, so conversations are neither stored by them nor used for training, by default on every plan.

> Model providers operate under zero-data-retention agreements: the conversations they process are not stored or used to train models.
> — https://aide.app/security · Data protection section, 'Zero data retention'

**Data residency** · Partly — One named region only, US East 2 on Azure, which is a problem for EU or UK residency; anything else requires a bespoke private-cloud provisioning conversation.

> Aide is hosted on Microsoft Azure (US East 2). For sovereign AI deployments with strict data residency requirements, Aide can be provisioned in public or private clouds.
> — https://aide.app/security · FAQ, 'Where does Aide run and where is customer data processed?'

**Getting out** · Partly — The audit record is exportable and retained from day one, but no page describes exporting your knowledge, Scenarios or conversation data at contract end, and there are no credits to expire.

> Filter by intent, Scenario, action type, or outcome, and export the record.
> — https://aide.app/platform/action-trace · FAQ, 'Can we review behavior across the whole operation?'

**Certifications** · Yes — Three attestations claimed and dated by recurring audit: SOC 2 Type II verified by an independent auditor, plus GDPR and HIPAA, with the report available on request.

> Aide is SOC 2 Type II certified and GDPR and HIPAA compliant. Every conversation runs on enterprise-grade infrastructure, every action follows procedures your team approves, and every decision is logged for audit.
> — https://aide.app/security · Hero paragraph

## Provenance

Every fact above comes from the vendor. Nothing here is independently corroborated yet.

## Coverage elsewhere

None recorded. Every fact above comes from the vendor and is not independently corroborated yet.

## What changed

- 2026-09-03 — Capabilities · Commercial terms · Free tier · Price

## Correcting this record

Anything marked "Not established" is a question we have not answered — it is not a No. If you represent Aide, or anything here is wrong or out of date, you can supply it with a source: the operations this site accepts from an agent are listed at https://theagentsindex.com/.well-known/agents.json — that catalogue needs no key to read, and each entry states its own auth.

Supplying evidence is free. Nothing that can be bought here changes what a field says, whether a gap is closed, or how this record is scored.