Actively

What it is

Actively is an account-intelligence platform that assigns one persistent AI agent to every account in your book. Each agent — Actively calls them Per-Account Agents — runs in the background around the clock: it reads the CRM record, call transcripts, intent data, and public signals like funding rounds and executive changes, then works out which accounts deserve attention now, who to contact there, and what to say. Finished work lands in an Agent Inbox, prioritized for the rep to act on; Watchtower handles monitoring, an Assistant answers questions about accounts, and an API Platform lets internal AI teams build on the same agents. The company pitches this as “intelligence-led revenue”: agents decide where effort goes, humans execute the conversations that matter.

Actively was founded in 2022 by Anshul Gupta and Mihir Garimella, two former Stanford AI researchers, and is based in New York. It raised a $5M seed from First Round Capital and a $17.5M Series A led by Bain Capital Ventures (April 2025), then a $45M Series B co-led by TCV and First Harmonic, announced 2026-04-28, at a reported $250M valuation — roughly $68M raised in total. Named customers include Ramp, Samsara, Attentive, Ironclad, and Verkada.

Why it shows up in RevOps stacks

• It is explicitly not an AI SDR. At the Series A, Gupta argued that AI SDRs failed because they optimized for volume — contacting as many prospects as possible. Actively’s bet is the opposite: fewer, better-targeted actions on the accounts that matter, with the human rep still sending. That makes it a fit for enterprise AE teams where a bad email to a top-200 account costs more than the email saves.
• Account memory that survives turnover. Because the agent belongs to the account rather than the rep, territory changes and rep churn don’t wipe out context. For a team reshuffling territories every year, that is the concrete value.
• It sits on the stack you already run. Output flows into Salesforce, Outreach, and Salesloft; it ingests Gong transcripts and third-party intent from sources like 6sense and HubSpot, plus internal enablement content. Actively uses a mix of in-house models and reasoning models from OpenAI and Anthropic.
• The results claim is specific. Samsara runs it across a 1,000+ person GTM team and reports 2x conversion on Actively-driven outreach; Ramp credits it with tens of millions in added revenue. Both are vendor-published, so treat them as a ceiling, but they are named and scoped rather than anonymous percentages.

Pricing reality

Actively does not publish prices. The site offers a demo request only — no free tier, no trial, no monthly plan. Contracts are annual and scoped by account volume, seats, and which modules you license. Third-party reviewers put deals from the mid-five figures per year upward, with large deployments well into six figures; one industry write-up brackets the category at roughly $30,000 a year to the high six figures, with Actively at the premium end. Budget accordingly: this is a line item that needs a CRO sponsor, not a team-lead credit card, and the ROI case has to rest on pipeline from a defined account set.

Best for

Enterprise and upper-mid-market sales orgs — CROs, RevOps leaders, and AE managers — running named-account or territory models of a few hundred to a few thousand target accounts, who want every account continuously researched and ranked without adding SDR headcount. The best-fit use case is expansion and strategic new-logo coverage on a finite list, where each conversation is worth a lot.

Skip it if you sell transactionally into a large SMB TAM, if you need an agent that sends autonomously at volume, or if your CRM data is too thin for an agent to reason over.

Versus the alternatives

• Salesforce with Agentforce — the market-share default. Pick it when you would rather consolidate agents inside the CRM you already pay for and can accept generic account research over a purpose-built per-account model.
• 6sense — the incumbent account-intelligence and intent platform. Pick it when your problem is identifying in-market accounts across a large TAM and orchestrating ads and marketing; pick Actively when the list is known and the problem is what reps should do next on each account.
• Rox — the fastest-growing entrant in agent-per-account selling, at a $1.2B valuation in March 2026. Pick Rox when you want agents that also write back to the CRM and handle admin, with a published free tier to pilot on; pick Actively when prioritization and account strategy matter more than CRM hygiene.
• Unify or 11x — pick these when the goal is outbound volume triggered by signals, which Actively deliberately does not chase.

Watch-outs

• No public price, annual lock-in. You cannot pilot it cheaply, and reviewers name cost and the annual commitment as the main complaint. Guard: negotiate a scoped first-year contract on one segment (a single region or 200–300 accounts) with a pipeline-sourced exit metric written into the order form.
• The agents are only as good as your CRM and call data. Reviewers flag occasional contact-data inaccuracies, and an agent reasoning over stale opportunity records will prioritize the wrong accounts. Guard: audit account ownership, contact roles, and Gong coverage on the pilot segment before go-live, and have reps flag bad recommendations in the Agent Inbox for the first 60 days.
• Vendor attribution is not your attribution. “2x conversion” and “tens of millions” are customer quotes in vendor releases. Guard: hold out a matched control segment during the pilot and compare meetings and pipeline created, not agent-activity counts.
• Build-vs-buy pressure is real. Some enterprises are building internal account agents on their own data. Guard: if you already have an internal AI team with warehouse access, run a two-week internal prototype before signing — Actively’s API Platform also lets you extend rather than replace that work.