Salesforce Edition Native AI platform, in your org

AgenTorQ for Salesforce — the platform, native in your org.

Install from AppExchange and run the same AgenTorQ AI platform natively inside Salesforce — LWC, Apex, Flows, live records, governed write-back. Record-aware AI Workmates ground on your CRM, route across 14+ models, and act under USER_MODE with a full audit trail.

Same engine as AgenTorQ Cloud — the only difference is where it runs.

Book Demo Explore Platform Explore Deployments
Native AppExchange managed package USER_MODE — CRUD/FLS/sharing enforced Apex, LWC & Flow native Immutable audit trail
Why the native edition

Why Salesforce teams choose the native edition

AgenTorQ for Salesforce is a real AppExchange managed package. It runs on-platform, honors your security model, and is managed by admins — so security teams can say yes and your data never leaves Salesforce.

Native managed package

Distributed on AppExchange and installed into sandbox or production. No middleware to stand up, no data lake to sync — your data stays in Salesforce.

Governed in USER_MODE

Every read and write runs in USER_MODE, so object- and field-level security and sharing are enforced per running user. Write-back is gated behind approval.

Record-aware by default

Workmates read the record in view — standard and custom objects, related lists and fields — so answers are grounded in your CRM, not generic.

Admin-managed rollout

Install, run a one-click setup, and assign permission sets. Admins control features, models and access with deny-by-default RBAC.

Multi-model, BYOK

Bring your own keys across 14+ providers and route each workflow to the right model with cost control and fallback — never locked to one vendor.

One engine, portable

The same AI engine, Workmates, governance, connectors and model routing as AgenTorQ Cloud. Adopt natively now; the platform travels with you.

Connect anything AI at cost / BYOK Deploy anywhere Governed & auditable Cited answers Multi-model routing AI Workmates DevSpace Studio
Salesforce-native architecture

One governed path, from record to model and back

Every request flows through a single governed path — from your Salesforce users and records, through orchestration and model routing, to the right AI model and back — logged end to end, on-platform.

Salesforce Org

Sales, Service and Experience Cloud, custom objects, and the Apex, LWC and Flow your team builds with.

Sales Cloud
Service Cloud
Experience Cloud
Custom objects

AgenTorQ Managed Package

Record-aware context, invocable Apex actions and LWC surfaces on record pages, app pages and the utility bar — running inside your org, no data lake to sync.

Orchestration + Multi-Model Routing

An Intent Router grades every turn and plans the work — data lookup, cited knowledge, reasoning or a governed action — selecting the right model per task with cost control and fallback.

Governance + Audit + Security

USER_MODE CRUD/FLS and sharing, SOQL and DML guards, named PII masking, RBAC access sets and a complete audit trail — enforced on every request.

AI Models — your keys, 14+ providers

OpenAIAnthropic ClaudeGoogle GeminixAI GrokDeepSeekMistralGroqCerebrasPerplexitySambaNovaMiniMaxAI21HuggingFaceOllama · self-hostedOpenAI-compatible
Live CRM grounding

AI grounded in the record in front of the user

AgenTorQ grounds every response in live Salesforce context — the current record, its related data and your schema — then reasons over it and proposes governed actions the user can approve in place. Answers cite the records they came from, so there is no guessing and no hallucinated data.

  • Standard & custom objects
  • Related lists & lookups
  • Field-level context
  • SOQL under USER_MODE
  • Case & opportunity summaries
  • Risk & sentiment scoring
  • Cited, grounded answers
  • Approval-gated write-back

Beyond records, add RAG over your knowledge — folders, files and chunks with embeddings and citations — plus document and OCR intelligence for PDF, Word, Excel, CSV and scanned images.

Book Demo
Supported surfaces

Built for the clouds and surfaces your teams already use

One governed platform across Sales, Service and Experience Cloud, custom objects and every native UI surface — with the same access controls and audit everywhere.

Sales Cloud

Record-aware selling on leads, opportunities and accounts.

  • Opportunity & renewal risk
  • Account briefs & next-best action
  • Quote & proposal drafting

Service Cloud

Faster, grounded case handling for agents and managers.

  • Case summarization & triage
  • Suggested replies from history
  • Knowledge-grounded answers

Experience Cloud

Governed self-service AI for portals and communities.

  • Member-facing assistants
  • Guest & authenticated access rules
  • Cited, on-brand responses

Custom objects

Ground on any standard or custom object, with your fields, relationships and validation rules respected.

Record & app pages

Drop AgenTorQ Lightning Web Components onto record and app pages via Lightning App Builder — no tab-switching.

Utility bar

Dock a persistent Workmate in the utility bar so it follows the user across every page in the app.

Flows & Apex

Call governed AI from screen and autolaunched Flows, and from invocable or callable Apex, with mappable outputs.

Profiles & permission sets Sharing rules respected Sandbox & production Object-level include/exclude
Meet your AI Workmates

Role-aware AI Workmates on every record

A Workmate is a role-aware AI agent that reads your data, reasons over your knowledge, and takes approved actions. Inside Salesforce, Workmates live on record pages and the utility bar — the same governed engine behind every role.

Sales Workmate

Grounds on your live pipeline to brief reps, surface risk, and log the next step.

Helps
AEs, sales managers, RevOps
Does
Summarizes opps, flags stalled deals, drafts follow-ups & creates tasks — on approval.

Service Workmate

Answers case questions from records and knowledge, then acts on the resolution.

Helps
Support agents, service managers
Does
Reads cases & KB with citations, drafts replies, and updates records on approval.

Ops Workmate

Connects systems and runs the busywork across your apps and workflows.

Helps
Operations, enablement, shared services
Does
Pulls data across connectors, runs Flows, and automates multi-step tasks — audited.

Admin Workmate

A configuration copilot grounded in your org's metadata and access model.

Helps
Salesforce & platform admins
Does
Explains schema & validation rules, drafts config, and manages access with guardrails.

Developer Workmate

A developer-grade copilot for Apex, LWC, Flows and GitHub inside DevSpace Studio.

Helps
Developers, technical architects
Does
Browses code, proposes diffs, runs tests, and ships — you approve before deploy.
See all Workmates on Solutions Browse the full feature catalog
Technical layers

Native to Apex, LWC and Flow — governed at every layer

Invoke governed AI from the platform primitives your admins and developers use every day — no external runtime, no glue code — with SOQL/DML guards, USER_MODE, FLS and sharing enforced end to end.

Apex

Call agents from invocable and callable Apex, pass record IDs and context, and receive structured results. AgenTorQ can also read, generate and review Apex under governance.

Lightning Web Components

Drop AgenTorQ LWCs onto record pages, app pages and the utility bar via Lightning App Builder. Components inherit your theme and honor the running user's access.

Flow

Add an AgenTorQ action to any screen or autolaunched Flow to summarize, classify, draft or route — with approvals and outputs mapped back to your variables.

Invoke an agent from Apex or Flow

The invocable action takes a record ID, an agent name and a governance profile, and returns a grounded summary plus approval-gated actions. Wire it into Flow with no code, or call it directly from Apex.

  • Invocable from Flow & Apex
  • SOQL reads under USER_MODE
  • DML writes behind approval
  • SOQL guard & DML guard
  • FLS & sharing enforced
  • Structured, mappable outputs
OpportunityRiskInvocable.cls
// Invoke a governed AgenTorQ agent from Apex or Flow.
public with sharing class OpportunityRiskInvocable {
  @InvocableMethod(label='Analyze Opportunity Risk')
  public static List<Result> run(List<Id> opportunityIds) {
    return AgenTorQ.invoke(new AgenTorQ.Request{
      agent      = 'opportunity-risk',
      recordIds  = opportunityIds,
      model      = 'auto',        // routing + fallback
      accessMode = AccessLevel.USER_MODE // CRUD/FLS honored
    });
    // Result.summary + Result.actions (approval-gated)
  }
}
DevSpace Studio

AI-native development for Apex, LWC, Flow and GitHub

A developer-grade copilot inside the platform — browse, edit and ship Apex, LWC and Flows in your org, and drive real git on GitHub, all with diffs you approve before anything deploys. Ask, plan, edit, diff, approve, deploy, test, fix.

  • VS Code-style 3-pane workspace
  • Apex, LWC & Flow
  • GitHub branch / PR / merge
  • AI diffs before approval
  • Test-coverage gutter
  • Persistent file context
Use cases

Governed AI applied to Salesforce work

Concrete, record-aware workflows enterprise teams put into production — every one under role-based access and audit, from grounded answers to real write-back and Flow execution.

Account briefings

User
Account manager, AE
Pain
Signals scattered across objects
Outcome
Consolidated account health, whitespace and next steps

Opportunity risk

User
Sales leader, AE
Pain
Surprise slipped and churned deals
Outcome
Risk scoring with reasons on every open opportunity

Case summaries

User
Support rep, manager
Pain
Long case threads slow to triage
Outcome
One-paragraph summaries and suggested next steps

Knowledge answers

User
Support, self-service members
Pain
Answers buried in articles and records
Outcome
Grounded, cited answers from KB and CRM

Email drafts

User
Sales, service, success
Pain
Hours writing follow-ups and replies
Outcome
On-brand drafts from CRM data, sent on approval

Field updates

User
RevOps, admins, reps
Pain
Manual data entry and stale records
Outcome
Approval-gated write-back under USER_MODE

Flow execution

User
RevOps, operations
Pain
Manual multi-step processes
Outcome
Agent-run Flows with human approval gates

Developer assistant

User
Salesforce developer
Pain
Context-switching to build and review
Outcome
AI that reads, edits and deploys Apex, LWC and Flow

Quote & proposal drafting

User
Sales, deal desk
Pain
Hours per quote and proposal
Outcome
Drafts from CRM data and approved templates
Setup

From AppExchange install to live in a single pass

Install the managed package, run the one-click setup, and go live. Admins stay in control at every step — models, Workmates, knowledge and access.

1

Install from AppExchange

Add the AgenTorQ managed package to a sandbox or production org and assign the permission sets.

2

Connect model providers

Add your own API keys across 14+ providers. Set a default engine and test latency — BYOK, no per-message metering.

3

Configure Workmates

Set purpose, base CRM and intent, and guardrails per role — or start from a ready-made Workmate template.

4

Add knowledge

Point Workmates at folders, files and objects for cited RAG, with object-level include/exclude and PII masking.

5

Assign access

Build SFDC-style profiles from the feature-toggle grid, deny-by-default, with per-user overrides — server-enforced.

6

Go live

Embed LWCs on record and app pages, dock a Workmate in the utility bar, and wire actions into Flow. Live in minutes.

Enterprise connectors

Reach beyond Salesforce data, too

The native edition grounds first on your CRM — and reaches 600+ apps via the connector layer, plus REST, GraphQL, databases and MCP, each with per-user OAuth and full call logging. Long-tail and enterprise systems are reachable through the connector layer, REST or MCP — not bespoke certified integrations.

Productivity

Microsoft 365, Google Workspace, Slack, Teams

Storage

SharePoint, OneDrive, Google Drive, Box

Development

GitHub, GitLab, Jira, Bitbucket, Azure DevOps

Databases & APIs

Postgres, Snowflake, BigQuery, REST, GraphQL, MCP

Explore Connectors
One platform, two editions

Same platform, also available standalone

Prefer a no-install start, or need AI across Microsoft 365, Google Workspace, Slack, GitHub and databases without a CRM? AgenTorQ Cloud runs the same engine, Workmates, governance, connectors and model routing as a standalone web workspace.

Same AI engine · Same Workmates · Same governance · Same connectors · Same model routing. Different deployment.

Launch Workspace Explore Deployments
FAQ

AgenTorQ for Salesforce, answered

The native edition, its governance, its models and how it relates to the standalone platform.

Is AgenTorQ a Salesforce-only product?

No. AgenTorQ is an enterprise AI platform with two deployment editions of one engine: AgenTorQ Cloud, a standalone web workspace that needs no CRM, and AgenTorQ for Salesforce, a native AppExchange managed package that runs inside your org. Salesforce is one supported enterprise system, and this page covers the native edition.

How does AgenTorQ install into Salesforce?

AgenTorQ for Salesforce is distributed as a managed package on AppExchange. An admin installs it into a sandbox or production org, runs a one-click setup that wires up model providers, Workmates, knowledge and access, and assigns permission sets. There is no middleware to stand up and no separate data lake to sync — your data stays in Salesforce.

Does the Salesforce edition support Apex, LWC and Flow?

Yes. AgenTorQ ships invocable Apex actions callable from Flow and Apex, Lightning Web Components you drop onto record pages, app pages and the utility bar via Lightning App Builder, and Flow actions to summarize, classify, draft or route. Inside DevSpace Studio it can also read, generate, review and deploy Apex, LWC and Flow metadata under governance.

How is Salesforce data kept governed?

Every read runs as SOQL and every write as DML in USER_MODE, so object- and field-level security and sharing are honored per running user. A SOQL guard and a DML guard wrap data access, named PII is masked before any prompt leaves the org, write-back is gated behind explicit approval cards, and every prompt, action and model call is recorded to an immutable audit trail.

Which AI models can the Salesforce edition use?

14+ providers on your own keys — OpenAI, Anthropic Claude, Google Gemini, xAI Grok, DeepSeek, Mistral, Groq, Cerebras, Perplexity, SambaNova, MiniMax, AI21, HuggingFace and self-hosted Ollama — plus any OpenAI-compatible endpoint. Each workflow routes to the right model with cost control and fallback, so you are never locked to a single vendor model, making it an open alternative to a native single-vendor CRM agent.

Is it the same platform as AgenTorQ Cloud?

Yes. The Salesforce edition and the standalone AgenTorQ Cloud edition share the same AI engine, AI Workmates, governance, connectors and model routing. Only the deployment differs — one runs natively inside Salesforce, the other as a standalone web workspace with no install.

Book your Salesforce architecture session

Bring your org model, security requirements and top use cases. We will map AgenTorQ into your Salesforce architecture — governance, USER_MODE access, Apex/LWC/Flow touchpoints and model routing — and share the AppExchange managed-package install path.

Book Demo View AppExchange Edition Explore Platform
reachout@agentorq.com