# Maven AGI — Enterprise AI agent platform that resolves customer conversations across chat, email, voice and internal tools.

> Source: The Agents Index — https://theagentsindex.com/maven-agi (structured, researched, re-verified)
> Facts last verified: 2026-09-25

Maven AGI is an enterprise AI agent platform for customer experience that resolves support conversations across chat, email, voice and web on one reasoning engine. It installs on top of an existing help desk, CRM and telephony stack and executes actions in those systems.

| Fact | Value |
| --- | --- |
| Website | https://www.mavenagi.com |
| Price | Maven AGI uses a clear, predictable enterprise pricing structure — no per-resolution ambiguity, no surprise invoices at scale. |
| Free tier | No |

No standing free plan and no self-serve trial is published. Every entry point is 'Book a demo', and onboarding is an implementation team rather than self-serve setup. One customer quote refers to a trial period, but no trial terms are offered publicly.

## What it does

**Writes code** · No — Maven's only contact with a customer repository is read-only knowledge ingestion; its own GitHub FAQ disclaims writing to repositories, and the agent authors support answers and system actions, not source files.

> Repository data is encrypted in transit and at rest, and Maven never writes to your repositories.
> — https://www.mavenagi.com/integrations/github · Frequently Asked Questions, answer to “Is the GitHub integration secure?”

**Done-for-you service** · Partly — You buy and operate the platform yourself, but Maven staffs embedded Forward Deployed Engineers and a dedicated implementation team to build and land the initial deployment, so the launch phase is a service rather than self-serve software.

> Forward Deployed Engineers (FDEs) are embedded engineers who work alongside customer teams to carry AI capability across the final distance into production, where workflows are messy, data is imperfect, and trust is non-negotiable.
> — https://www.mavenagi.com/resources/forward-deployed-engineers · Blog body, section 'The Solution: Forward Deployed Engineers'

## How much it does unattended

**Runs autonomously** · Yes — The platform is sold on unattended resolution, claiming up to 93% of incoming queries closed without a human, with escalation triggered by configurable confidence and risk signals rather than by default.

> Maven AGI is an enterprise AI agent platform that automates customer support across chat, email, voice, and web, resolving up to 93% of incoming queries without human intervention.
> — https://www.mavenagi.com/product/agent-platform · Hero paragraph, 'The AI Agent Platform Built for Enterprise CX'

**Multi-agent** · Partly — Maven describes several outcome-specific agents on one platform and per-situation Charters inside a single reasoning engine, but it explicitly avoids channel-specific or peer agents coordinating with each other, so this is one governed agent with many roles rather than a multi-agent orchestration graph.

> One platform working across the entire customer journey. Each agent owns a specific outcome and resolves it end to end, from the first question someone asks to the conversation that keeps an account renewing every year.
> — https://www.mavenagi.com · Homepage section 'What we resolve', intro paragraph

**Agent permissions** · Yes — Permissions are configured centrally: one control layer sets guardrails, escalation patterns and which systems an agent may act in, and Charters scope behaviour and access per situation.

> One control layer governs tone, guardrails, escalation patterns, and system permissions.
> — https://www.mavenagi.com/product/agent-designer · Section 'Centralized AI Agent Behavior Controls and Guardrails'

## Which models it runs on

**Claude** · Yes — The Master Services Agreement names Anthropic, alongside OpenAI, as one of exactly two LLM providers generating output today, and lists it as a contracted sub-processor rather than an optional integration.

> At this time, Maven AGI uses OpenAI and Anthropic large language models (“LLMs”) to generate Outputs.
> — https://www.mavenagi.com/legal/msa · Section “AI/Machine Learning”, bullet “AI Tools; Inputs and Outputs”

**GPT** · Yes — OpenAI models are named as engines Maven runs on, and the vendor publishes its own enterprise phone-call evaluation of OpenAI's newest realtime voice model, so GPT-family support is asserted rather than implied.

> Maven Voice listens, interprets intent, and adjusts in real time using best in class engines like OpenAI, Phonic, and Cartesia.
> — https://www.mavenagi.com/product/maven-voice · Section 'Multilingual AI Voice Agent with Real-Time Speech Understanding'

**Open models** · No — The agreement enumerates its LLM providers as OpenAI and Anthropic and promises written notice if any other is added, a closed list with no open-weight or self-hosted model in it.

> If, at any time following the Effective Date, Maven AGI uses other LLM sub processors or AI Tools to process customer content, it will provide you with notice either in this Agreement
> — https://www.mavenagi.com/legal/msa · Section “AI/Machine Learning”, bullet “AI Tools; Inputs and Outputs”, sentence after the provider list

**Model choice** · Partly — Maven claims its own architecture is LLM-portable on the platform page, but the contract shows the vendor picks the models and may change them on notice; no buyer-facing model selector is published anywhere.

> Vendor agnostic and LLM portable
> — https://www.mavenagi.com/product/agent-platform · Section “Integration-First Architecture”, third capability bullet

## Where you use it

**In your editor** · No — The vendor enumerates its surfaces as voice, chat, messaging, email and internal tools; no editor or IDE extension appears in that list or anywhere else on the site.

> Maven brings a single AI agent to voice, chat, messaging, email, and internal tools — all powered by one reasoning engine and one set of policies.
> — https://www.mavenagi.com/features/agent-channels · Hero paragraph under the H1 “One AI Agent for Every Customer and Employee Channel”

**On the command line** · No — No terminal surface is claimed: the same enumeration of channels covers voice, chat, messaging, email and internal tools, and the docs that would describe a CLI are behind a login.

> Maven brings a single AI agent to voice, chat, messaging, email, and internal tools — all powered by one reasoning engine and one set of policies.
> — https://www.mavenagi.com/features/agent-channels · Hero paragraph under the H1 “One AI Agent for Every Customer and Employee Channel”

**In your pipeline** · No — The GitHub connector reads documentation and re-indexes on push; the vendor states outright that it never writes to a repository, so acting on a pull request or in CI is ruled out.

> Maven focuses on the documentation layer: README files, wiki pages, markdown docs, and code comments or docstrings. It does not index raw source code line by line
> — https://www.mavenagi.com/integrations/github · Frequently Asked Questions, answer to “Does Maven index source code or just documentation?”

**In a browser** · Yes — There is a hosted web console: Agent Designer is described as a unified browser workspace for analytics, testing and guardrails, reached through the vendor's app login.

> Agent Designer gives CX, operations, and product teams a unified workspace to analyze performance, refine knowledge, tune behavior, and validate agent changes before they reach customers, all without waiting on engineering.
> — https://www.mavenagi.com/product/agent-designer · Hero paragraph, 'The AI Agent Management Platform for CX and Operations Teams'

## Whose machine it runs on

**Self-hosted** · No — The vendor's own marketplace listing classes delivery as SaaS and states the software is not deployed in the buyer's AWS account; no on-premise, VPC or single-tenant option is published on any tier.

> Delivery method Software as a Service (SaaS)
> — https://aws.amazon.com/marketplace/pp/prodview-l2djyrnxnfffq · Details panel, “Delivery method” row (adjacent “Deployed on AWS” row reads “No”)

**Open source** · No — The licence bars any attempt to reach the platform's source code; the only public repositories are generated API client libraries, and they carry no open-source licence file at all.

> reverse engineer, disassemble, decompile, decode, or otherwise attempt to derive or gain access to the source code of the Application or any part thereof
> — https://www.mavenagi.com/legal/msa · Section “Restrictions”, clause (c)

## What it costs to run

**How it meters** · Partly — The vendor names resolution volume as what the price scales with, but publishes no per-resolution rate; the contract meters against a subscription usage allowance with overage priced privately in the order form.

> Starting prices listed here - contact us for custom pricing based on your expected scope, complexity, and volume of resolutions.
> — https://aws.amazon.com/marketplace/pp/prodview-l2djyrnxnfffq · Pricing, “12-month contract (1)” table, Description cell of the “Enterprise Custom Pricing - Minimum” row

**Free tier** · No — Every call to action on the site is a booked demo, and onboarding is described as a dedicated implementation team explicitly instead of self-serve setup, so there is no way in without a sales conversation.

> A dedicated implementation team — not a self-serve setup guide — gets you to measurable deflection within the first week.
> — https://www.mavenagi.com/explore/ai-agents-roi-calculator · Section 'White-glove onboarding'

**API access** · Yes — Maven publishes SDKs and platform APIs, and Maven Apps are written in TypeScript against a named SDK package with a real deploy pipeline, so a customer team can drive and extend the platform from its own stack.

> Maven offers comprehensive SDKs and APIs to empower developers.
> — https://www.mavenagi.com/resources/platform-launch · Platform launch announcement, developer paragraph

**MCP server** · Yes — Maven supports Model Context Protocol as a client: a team pastes an MCP server URL and configures access in the UI, though the vendor routes activation through your assigned CX manager rather than open self-serve setup.

> That's why Maven now supports Model Context Protocol (MCP), an open standard for connecting AI agents to software systems. If a system has an MCP server, your team can configure access through Maven's UI in just a few steps; no additional engineering required.
> — https://www.mavenagi.com/resources/model-context-protocol · Blog body, paragraph introducing MCP support

**Bring your own key** · No — Inference is Maven's to arrange: the contract names the two LLM providers it uses and reserves the right to change them on notice, with no provision for customer-supplied keys, accounts or cloud.

> At this time, Maven AGI uses OpenAI and Anthropic large language models (“LLMs”) to generate Outputs.
> — https://www.mavenagi.com/legal/msa · Section “AI/Machine Learning”, bullet “AI Tools; Inputs and Outputs”

## Buying it for a team

**A company can buy it** · Yes — This is sold only to organisations: pricing is an enterprise structure reached through a demo and implementation team, with named enterprise customers and no individual sign-up path anywhere on the site.

> Maven's enterprise AI agent platform is built for the customer journey. Drive action across your existing help desk, CRM, and telephony while keeping IP and data protected with certified, auditable security.
> — https://www.mavenagi.com · Homepage hero paragraph

**Pooled budget** · Yes — Usage is metered against one account-level subscription allowance rather than per user, and the vendor must warn the organisation at ninety percent of that shared limit before overage begins.

> Company will notify Customer in writing when Customer’s usage reaches 90% of the agreed subscription usage limit.
> — https://www.mavenagi.com/legal/msa · Section “Payment”, bullet “Fees and Overages”

**Admin controls** · Yes — Administrators constrain the agent directly: routing rules, workflow policies, identity controls and deterministic logic gate every action, and Agent Designer controls which systems the agent may act in.

> Routing rules, workflow policies, identity controls, and deterministic logic ensure every AI action follows your internal standards. Nothing operates outside governance.
> — https://www.mavenagi.com/features/trust-compliance · Section 'Policy-aligned workflows you can actually enforce'

**Audit log** · Yes — Agent decisions and actions are logged, traceable and exportable with SIEM integration, and the vendor separately lists comprehensive audit logging among its platform security controls.

> Every AI decision is logged, traceable, and exportable. SIEM integrations provide line of sight into inputs, outputs, system references, and policy checks across the full agent lifecycle.
> — https://www.mavenagi.com/features/trust-compliance · Section 'Full transparency and auditability'

**Single sign-on** · Partly — SSO is asserted as a platform capability beside RBAC and MFA, and Maven says it federates to the identity system a customer already runs, but no protocol, no SAML or OIDC, and no IdP is named publicly.

> RBAC, SSO, MFA, secure SDLC practices
> — https://www.mavenagi.com/features/trust-compliance · Section “Security, Trust, Compliance, and AI Ethics – Built In, Not Bolted On”, first accordion panel, feature bullet

_Not established: Seat model_

## What happens to your code

**Opt out of training** · Yes — No opt-out is needed because the vendor states customer data is never used to train or fine-tune foundational models and is not aggregated for model improvement; sensitive data is also redacted before any external model call.

> Maven AGI does not use customer data to train or fine-tune foundational models. Customer data is not aggregated or reused for model improvement.
> — https://www.mavenagi.com/resources/security-and-governance-at-maven · Section 'Product Security: The AI-Specific Layer', paragraph 'Customer data is not used to train models.'

**Getting out** · Partly — Agent decisions and logs are described as exportable with SIEM integration, which covers operational records, but nothing read states how knowledge, conversation history or configuration leave the platform at contract end.

> Every AI decision is logged, traceable, and exportable.
> — https://www.mavenagi.com/features/trust-compliance · Section 'Full transparency and auditability'

**Certifications** · Yes — A deep certification stack is claimed and partly independently audited: SOC 2 Type II, ISO 27001:2022, PCI DSS 4.0 Level 1 and HIPAA/HITECH are stated as achieved, while ISO 27701/42001/27017/27018 and the GDPR, CCPA and CPRA privacy attestations are dated as expected Q1 2026.

> Maven AGI is certified for SOC 2 Type II, ISO 27001:2022, PCI-DSS v4.0 Level 1, and HIPAA/HITECH, with additional certifications (ISO 27701, 42001, 27017, 27018) and privacy attestations (GDPR, CCPA, CPRA) expected in Q1 2026.
> — https://www.mavenagi.com/legal-privacy · Section 'Security & Compliance'

_Not established: Data residency_

## Provenance

1 of 18 sources behind the facts above are independent of the vendor.

## Coverage elsewhere

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

## What changed

- 2026-09-25 — Capabilities · Commercial terms · Free tier · Price
- 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 Maven AGI, 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.