Rox

• ▸Sales Intelligence agents that answer account questions (e.g. hiring changes, tech stack) on demand
• ▸Conversation Intelligence: call recording and analysis
• ▸Revenue Intelligence: pipeline/opportunity dashboards by rep
• ▸Outbound Agent rebuilt around a four-stage flow (Prospect, Configure, Execute, Monitor) with live counts of leads found, contacts sequenced, actions taken, meetings booked
• ▸Dialer with agent-prepped call context and talking points, plus native call steps in sequences
• ▸Governance layer applying role-based, field-level access rules to both human users and agents
• ▸Configurable scheduled research columns on Accounts that refresh on a set cadence
• ▸Auto-prospecting: agents build their own prospect lists from a written buyer description, score each contact for fit, and can add good-fit prospects to a sequence automatically
• ▸Agentflows: natural-language instructions plus scoped tools, with the agent working out and running the steps (e.g. lead routing, renewal prep, list building)

Machine-readable spec

No OpenAPI link

Platform Tour: How Rox runs one AI agent per account to manage your sales book of business — Rox

Checked 2026-09-24 against their pricing page. Vendors change pricing without notice.

Rox positions itself as a vertically integrated revenue system built around AI "Agent Swarms" operating on a proprietary data layer (Rox System of Record / Revenue Data Fabric), rather than a thin AI layer bolted onto an existing CRM. Its own docs describe integrations with Salesforce, HubSpot, Gmail, Outlook, and Slack as data sources feeding that layer, and recent release notes show active development across outbound agents, apps, governance, and a dialer.

Enterprise revenue teams (the vendor names "Global 2000") with a top-heavy sales org that wants to make its best reps materially more productive with AI agents handling research, prioritization, and outreach prep, rather than automating junior reps on small accounts.

Both are metered per agent action, which can get expensive fast for teams running research-heavy workflows at volume. Teams wanting a simple, low-cost CRM without an enterprise governance/permissions layer are outside what Rox is built for.

Tools that let a revenue team compose several data providers into one workflow, run it conditionally across a list of accounts or people, and push the result into a CRM or sequencer. The unit of work is a table or a pipeline, not a single lookup.

Category: GTM data orchestration. Membership is decided by 4 stated tests, not by category tags — see them and the full comparison.

CargoPipeline-shaped rather than table-shaped, and aimed at the engineer who would otherwise build the orchestration in-house.ClayTable-first workflow for orchestrating enrichment providers, AI research and activation steps.DeeplineAPIMCPAgent-first: the workflow is expressed as instructions to a research agent rather than as columns.FloqerAPIMCPLeads with CRM enrichment and buying-signal monitoring rather than with list building.FreckleRuns GTM workflows from a CLI instead of a grid, aimed at teams that prefer scripted workflows.OctaveFocused on turning enriched context into positioning and messaging rather than on the enrichment itself.UnifyBundles intent signals and outbound sending with the enrichment, so it competes as a whole motion rather than as a data layer.ZoomInfo GTM AIThe incumbent data vendor's own orchestration layer, with ZoomInfo's database as the default source rather than a bring-your-own waterfall.

• Exact per-task-type agent-action costs (the pricing FAQ explains the metering principle but publishes no per-task numbers)
• A machine-readable OpenAPI specification was not included in the reviewed source set.