AgenTorQ
Experience WorkspaceWorkspace Book Demo Start free trial
30-day free trial Every plan · both editions

One enterprise AI platform. In your browser or inside Salesforce.

AI Workmates that answer from your data with their sources, and take action only after you approve — across 400+ connected apps and 14+ AI providers on your own keys.

Start your 30-day free trial Book a demo

Every plan starts with a 30-day free trial. Cancel anytime before it ends at no cost.

AgenTorQ Cloud

In your browser

No install
Explore the browser edition
AgenTorQ for Salesforce

Inside your org

Native
Explore the Salesforce edition
Early accessAuto-Pilot crews for Salesforce Routine record work, with approvals, receipts and undo. Browser editionOpen a browser and start No install and no Salesforce required. Every plan30 days free Cancel anytime before it ends at no cost.

The life of one request.

Scroll to follow a single question from the apps it reads to the line it leaves in the audit trail. The same five steps run in the browser and inside Salesforce.

Illustrative product screen

  1. 1

    Connected app data

    The Workmate reads the records, messages and files it needs from the apps you connected — 400+ through the connector layer — under the asking person's own access.

  2. 2

    An answer with its sources

    Named personal data is masked before the prompt leaves. The answer comes back from the provider you chose, on your own key, and cites every source it used.

  3. 3

    An approval

    A proposed change never runs on its own. It waits on an approval card that previews exactly what will change, until a person says yes.

  4. 4

    The action

    Once approved, the change is written back to the system it belongs to. Auto-Pilot crews, in early access, also leave a receipt you can undo.

  5. 5

    The audit trail

    Who asked, what was masked, which model ran, what was read, who approved and what was written: every step lands in an immutable trail you can export via API.

AgenTorQ Cloud · browser edition

Open a browser. Your AI workspace is ready.

No install and no Salesforce required. Sign in from your browser, connect the apps you already use, and put governed AI Workmates to work.

400+connected apps, each linked with the user's own sign-in
14+AI providers on your own keys, with no markup
Maskednamed personal data, before any prompt leaves the platform
Auditedactions wait for your approval, and every step is logged
  1. 1

    Sign up

    Open the workspace, sign in, and pick Personal or Team under Subscription. The first 30 days are free.

  2. 2

    Connect your apps

    Link Salesforce, Microsoft 365, Google Workspace, Slack, GitHub and more, each with your own sign-in.

  3. 3

    Ask a Workmate

    Get answers with their sources, and approve any action before it runs.

30 days free, every plan.

Every plan — Personal and Team — starts with a 30-day free trial. Cancel anytime before it ends at no cost.

Start your 30-day free trial See the browser edition
How it helps

Outcomes by role — what gets easier, the risk it removes, and the control you keep.

Pick a role. Every point below maps to a capability you can see on this site: answers with sources from 400+ apps, approvals before actions, PII masking, an audit trail, 14+ providers on your own keys, the browser or Salesforce, and Auto-Pilot crews in early access with undo.

For the CEO

What gets easier

Leaders get one governed AI layer across the business instead of a different assistant in every tool, in the browser or inside Salesforce.

The risk it removes

Staff pasting company data into unmanaged consumer chatbots, with no masking and no record of what happened.

The control you keep

Choice of 14+ AI providers on your own keys, so the business is never locked to one vendor.

400+ connected apps Browser or Salesforce 14+ providers, your keys

For the CISO / Security

What gets easier

One control plane for how AI is used: role-based access, model allowlists and data scope set centrally and enforced on the server.

The risk it removes

Personal data reaching a model — named PII is masked before any prompt leaves, and only the masked prompt reaches the model you chose.

The control you keep

An immutable audit trail of every prompt, model call, approval and write, exportable via API.

PII masking RBAC, deny-by-default Audit trail

For the Sales / RevOps leader

What gets easier

Reps get answers about accounts and pipeline with the sources attached — records, email and chat together — without switching tools.

The risk it removes

Unreviewed changes to pipeline data: every update is proposed on an approval card first.

The control you keep

Auto-Pilot crews (early access) can ask first, respect never-touch fields and daily limits, and every change leaves a receipt you can undo.

Answers with sources Approvals before actions Auto-Pilot (early access)

For the Service leader

What gets easier

Agents get case answers grounded in records and knowledge, with citations, and drafted replies they can edit.

The risk it removes

Answers invented from nothing — Workmates answer from approved sources scoped to the user's access, and show them.

The control you keep

Record updates wait for approval, and destructive steps always confirm.

Cited knowledge Edit before send Approval cards

For the Salesforce Admin / Architect

What gets easier

A native managed package on record pages and the utility bar, admin-managed, with the same engine as the browser edition.

The risk it removes

AI seeing more than the user can — the Salesforce edition runs in USER_MODE, so CRUD, field-level security and sharing are always enforced.

The control you keep

SOQL and DML guards on every read and write, and permission sets you assign.

USER_MODE SOQL / DML guards Admin-managed

For the IT / Ops

What gets easier

Connect Salesforce, Microsoft 365, Google Workspace, Slack, GitHub, databases, REST APIs and MCP servers once, through one connector layer.

The risk it removes

Shared service accounts and secrets in prompts — each user connects with their own OAuth, and credentials live in an external vault.

The control you keep

Per-user and per-tenant isolation, with call logging on every connector.

Per-user OAuth Credential vault Call logging

Without AgenTorQ

  1. Someone copies customer details into a consumer chatbot to get a summary.
  2. The answer arrives with no sources, so nobody can check it.
  3. They switch to the CRM, email and chat to find what actually happened.
  4. The record is updated by hand, or not at all, with no review.
  5. Nobody can say later what the AI saw or what changed.

With AgenTorQ

  1. The person asks a Workmate in the browser or on the Salesforce record.
  2. Personal data is masked and the request runs on the provider you chose, on your own key.
  3. The answer cites the records, messages and files it used, across 400+ connected apps.
  4. The proposed update waits on an approval card with a preview of the change.
  5. Every step lands in the audit trail — and Auto-Pilot crews (early access) leave a receipt you can undo.
How a request flows

From 400+ apps to a Workmate, an approval and an action — and every step logged.

Answers are grounded in the systems you connect. Personal data is masked before the prompt leaves, the request runs on the provider you chose, and any change waits for a person.

400+ CONNECTED APPS SalesforceMicrosoft 365Google WorkspaceSlack · GitHubDatabases · APIs · MCP Workmateanswers with sources Mask + routePII masked first14+ providerson your own keysfallback ready Approvalpreview the changeApprove Actionwrite-backreceipt · undo* Auditevery steplogged * receipts and undo are part of Auto-Pilot crews (early access)
  1. Connected apps — 400+ through the connector layer — supply the context, under the user's own access.
  2. A Workmate prepares the answer and cites its sources.
  3. Personal data is masked and the request is routed to one of 14+ providers, on your own key, with fallback.
  4. Any change waits on an approval card with a preview.
  5. The approved action is written back — Auto-Pilot crews (early access) add a receipt and undo — and every step is logged.
Grounding and routing Changes held for approval Live request
Two ways to deploy · one platform

Deploy AgenTorQ your way.

Choose the experience that fits your organization. Both editions use the same AI platform, governance, and orchestration engine — the only difference is where it runs.

Edition · Standalone web

AgenTorQ Cloud

A standalone AI workspace connecting Salesforce, enterprise applications, documents, APIs, databases and multiple AI providers — no CRM required.

Best for Cross-functional teams Enterprise AI workspace Multiple business systems Centralized governance Fast rollout
  • AI Workmates
  • RAG
  • Multi-model AI
  • Connectors
  • Enterprise Search
  • Workflow Automation
  • Audit
  • Approvals
  • APIs
Launch Workspace
Edition · Salesforce-native

AgenTorQ for Salesforce

Install AgenTorQ inside Salesforce via AppExchange and bring AI Workmates to every record and workflow — enforced in USER_MODE.

Best for Salesforce-first orgs Native UX Record-aware AI Admin-managed deployments
  • LWC
  • Apex
  • Flow
  • Record Pages
  • Utility Bar
  • Salesforce Security
  • Live CRM Grounding
  • AI Actions
View AppExchange Edition

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

Explore Platform Book Demo
The problem

Enterprise AI is fragmented. AgenTorQ brings it together.

Point tools, locked-in copilots and brittle scripts leave data siloed, actions ungoverned, and models out of your control. One governed platform closes the gaps.

Shadow AI risk

Staff paste company data into consumer chatbots with no masking, no policy, and no audit trail.

Data locked in silos

Answers live across Salesforce and dozens of apps, files and databases — but nothing reads them together.

No model flexibility

Vendor-locked copilots pin you to one model family and meter you per conversation.

No audit trail

When AI acts on real systems, teams can't see what it did, on whose behalf, or why.

No workflow execution

Most assistants only chat. They can't update a record, run a Flow, or send an approved email.

Brittle integrations

Hand-rolled point-to-point scripts break on every API change and never reach the long tail of apps.

Meet your AI Workmates

AI Workmates that answer, reason, and act.

A Workmate is a role-aware AI agent that reads your data, reasons over your knowledge, and takes approved actions — the same governed engine behind every role, in the web workspace or inside Salesforce.

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.

New · Early access Auto-Pilot for Salesforce

AI crews that watch the records you choose and do the routine work on a schedule — applying changes automatically or asking first, with a receipt and an undo for every change.

See Auto-Pilot
Under the hood · the read → reason → act loop

Understand intent

The Intent Router grades every request and routes it to data lookup, knowledge, reasoning, or a real action — nothing hardcoded.

Ground on enterprise data

Grounds on Salesforce records, uploaded files, your RAG knowledge base and 400+ connected apps — cited every time.

Choose the right model

Route by task across OpenAI, Claude, Gemini, Grok, DeepSeek, Mistral, Groq, Ollama and more — with fallback and cost control.

Act with governance

Create and update records, run Flows, and send email behind approval cards — every action logged to an immutable audit trail.

Explore Solutions See all Features
Enterprise connectors

Connect your whole enterprise.

Salesforce is one supported system among many. Reach 400+ apps via the connector layer, plus REST, GraphQL, databases and MCP — each with per-user OAuth and full call logging.

SalesforceHubSpotDynamics 365Microsoft 365Google WorkspaceSlackTeamsSharePointOneDriveGoogle DriveBoxGitHubGitLabJiraAzure DevOpsServiceNowWorkdaySAPNetSuitePostgresMySQLSnowflakeBigQueryRESTGraphQLMCP serversWebhooks Workmate400+ apps

Per-user OAuth on every connection, call logging on every connector. Names shown are examples from the connector layer, not a complete list.

CRM

Salesforce, HubSpot, Dynamics 365

Productivity

Microsoft 365, Google Workspace, Slack, Teams

Storage

SharePoint, OneDrive, Google Drive, Box

Development

GitHub, GitLab, Jira, Bitbucket, Azure DevOps

Enterprise

ServiceNow, Workday, SAP, NetSuite

Databases

Postgres, MySQL, Snowflake, BigQuery

AI providers

14+ engines, your keys — OpenAI, Claude, Gemini…

Generic

REST, GraphQL, MCP servers, Webhooks, custom

400+ apps via the connector layer, plus REST, databases, and MCP — not bespoke certified integrations.

Explore Connectors
Multi-model AI platform

Use the right model for every task.

Bring your own keys and route by task, cost, latency, reasoning quality, or policy — across every major provider, plus self-hosted and any OpenAI-compatible endpoint.

OpenAI Anthropic Claude Google Gemini xAI Grok DeepSeek Mistral Groq SambaNova Cerebras Perplexity MiniMax AI21 Hugging Face Ollama · self-hosted OpenAI-compatible
Bring your own keys Route by task Cost & latency control Policy & fallback Live model fetch — hundreds of models

All product names, logos, and trademarks shown are the property of their respective owners and are used for identification purposes only. AgenTorQ is model-agnostic and is not affiliated with, endorsed by, or sponsored by any AI provider; logos indicate compatibility only.

Governed enterprise AI

Governed by default. Auditable by design.

Zero-trust controls wrap every request — so security teams can say yes to AI, and you keep a complete record of every action.

RBAC

Feature-access control by role and profile, deny-by-default, with per-user overrides, server-enforced.

PII masking

Named personal data is masked before any prompt leaves — only the masked prompt reaches the model you chose.

Salesforce sharing & FLS

The Salesforce edition runs in USER_MODE — sharing, CRUD and field-level security are always enforced.

SOQL / DML guards

Every read passes a SOQL guard and every write a DML guard before it ever touches your data.

Per-user OAuth

Credentials live in an external vault with per-user OAuth and automatic token refresh — never shared.

Audit logs

Every prompt, action and model call is recorded to an immutable audit trail for review and compliance.

Action approvals

Write actions are gated behind explicit approval cards, and destructive steps always require confirmation.

Model access controls

Allow or restrict specific providers and models per role, team or workflow.

Data isolation

Per-user isolation keeps personal chats and OAuth tokens private; each team and tenant sees only its own data.

View Security

Compliance-ready architecture — your data stays in your systems, and only the masked prompt reaches the model you chose. Read our Privacy Policy.

Nothing changes without your say. Try it free for 30 days.

Every plan starts with a 30-day free trial. Cancel anytime before it ends at no cost.

Start your 30-day free trial See how it is governed
Developers

DevSpace Studio: AI-native development for Salesforce, GitHub, GitLab and Bitbucket.

A developer-grade copilot inside the platform — browse, edit and ship Apex, LWC and Flows, and drive real git on GitHub, all with diffs you approve before anything deploys.

VS Code-style 3-pane workspace Apex, LWC & Flow GitHub branch / PR / merge AI diffs before approval Test, deploy & fix Persistent context
Explore Developers
How it compares

An open AI platform vs. locked-in CRM AI.

A native, single-vendor CRM agent is convenient — but locked to one model, metered per conversation, and limited in what it reads and does. The honest side-by-side.

Capability AgenTorQOPEN AI PLATFORM Native CRM AIVENDOR-LOCKED
Model choice14+ providers + self-host, per taskLocked to one vendor model family
Pricing modelBYOK — pay providers directlyMetered per conversation / credits
Grounding across appsSalesforce + 400+ apps, files & RAG, citedGrounding within one vendor's framework
Document / OCRPDF, Word, XLSX, CSV & scanned-image OCRLimited file understanding; OCR rare
Connectors400+ apps + REST, GraphQL, DBs & MCPClosed ecosystem; approved integrations only
Workflow actionsCreate/update, run Flows, send email — approvedLimited to predefined topics/actions
ObservabilityLogs, audit trail, call logs & health consoleLimited visibility into calls and failures
Deployment optionsStandalone web and Salesforce-nativeOne place only
Developer workspaceDevSpace: Apex/LWC/Flow + GitHub/GitLab/Bitbucket gitNo developer IDE
BYOKYour keys across every providerNo bring-your-own-key option

AgenTorQ is an independent product and is not affiliated with, endorsed by, or sponsored by Salesforce or any AI model provider. A general, good-faith comparison against a typical native CRM AI agent; capabilities vary by edition, license and release.

See the full atomic comparison
Use cases

Use cases by team, in the browser or inside Salesforce.

Twelve use cases across sales, service, deal desk, IT and security, operations, Salesforce teams and leadership — each with the capabilities it uses.

Sales / RevOpsCloud (browser)Salesforce

Account and pipeline briefings before every call

The problem
Reps piece together account context from the CRM, their inbox and team chat, and still walk into calls missing the latest escalation.
What the Workmate does
The Sales Workmate summarizes open opportunities, flags stalled deals and drafts the follow-up, citing the records, emails and messages it used. Tasks and record updates are proposed on an approval card.
Capabilities it uses
Answers with sources400+ connected appsPer-user OAuthApproval cards
Sales / RevOpsSalesforceAuto-Pilot · Early access

Auto-Pilot renewals crew that keeps opportunities current

The problem
Next steps and follow-up fields drift out of date between pipeline reviews, and nobody has time to sweep every record.
What the crew does
Describe the job in one sentence and Auto-Pilot suggests a Sales crew you can edit. The crew watches the opportunities you choose, finds its own work on a schedule, and either applies updates automatically or asks first.
Capabilities it uses
Auto-Pilot (early access)Named rules and never-touch fieldsDaily limits and working hoursReceipts and undo
IT / SecurityCloud (browser)Salesforce

A governed AI rollout across the company

The problem
Teams adopt AI tools faster than security can review them, with no shared policy and no record of use.
What the Workmate does
Security sets profile-style roles from a feature-toggle grid, allowlists models per role and workflow, and relies on PII masking and the audit trail on every request.
Capabilities it uses
RBAC, deny-by-defaultModel allowlistsPII maskingImmutable audit trail
See all use cases
Insights

Read how governed AI works in practice.

Illustrative scenarios

How a Workmate or crew works, step by step

Four clearly labelled walkthroughs — challenge, setup, steps and the controls in play.

Read the scenarios White paper

Governed enterprise AI: approvals, audit trails and PII masking in practice

How identity, access control, guards, approvals and an immutable audit trail fit together — and what to check before rollout.

Read the paper White paper

Multi-model AI with bring-your-own-keys: choosing and routing across 14+ providers

Why a stable model abstraction, routing and fallback matter, and how BYOK keeps the AI bill with the provider, at cost.

Read the paper

Put a governed AI platform over your whole enterprise.

Start in the web workspace or book a guided architecture walkthrough. Same platform, same governance — you choose where it runs.

Book Demo Launch Workspace
reachout@agentorq.com

About the name

Is AgenTorQ the same as AgentRQ?

No. AgenTorQ — spelled A-g-e-n-T-o-r-Q, from agent and torque — is an enterprise AI agent platform from Opnyx Solutions, available as a standalone cloud workspace and as a native Salesforce AppExchange managed package. AgentRQ is an unrelated product from a different company. If you are looking for the governed, multi-model AI agent platform for Salesforce and enterprise apps, you are in the right place: agentorq.com.

How do you spell and pronounce AgenTorQ?

One word, with a capital A, T and Q: AgenTorQ. Pronounced agent-tork. It is also written Agentorq or AgenTorq in plain text. The company behind it is Opnyx Solutions, and the platform ships in two editions: AgenTorQ Cloud and AgenTorQ for Salesforce.