Close

What it is

Close is a sales CRM that treats the phone as part of the record instead of an integration hanging off it. Calling, SMS, and email live on the lead: a rep opens a lead, dials from it, and the call, recording, transcript, and outcome land on that lead with no connector in between. The closest structural comparison is Pipedrive with a dialer welded on — except Close built the telephony first and the pipeline around it.

The company itself is unusual for the segment. Close was founded in 2013 by Steli Efti, Anthony Nemitz, and Thomas Steinacher, runs about 120 employees, serves 11,500+ customers, and describes itself as profitable and customer-funded since launch with no venture rounds. That matters more to a buying decision than it sounds: most of the AI-selling vendors Close now sits beside are spending growth capital and reprice on a funding clock.

On 9 June 2026 Close moved Chloe to general availability on every plan for U.S. and Canada accounts. Chloe calls leads, runs structured discovery against your own qualification criteria, handles objections, books meetings onto the calendar, and logs the conversation. The beta ran with 306 businesses placing more than 818,000 calls, reaching 111,915 prospects and customers across 6,400+ hours of conversation (Close press release, 9 June 2026).

Why it shows up in RevOps stacks

• The dialer is a plan feature, not a separate vendor. Power dialer at Growth, predictive dialer at Scale, calling metered at roughly $0.02 per minute, phone numbers from $1 per month. A team that would otherwise buy a CRM plus Orum or Nooks buys one product and reconciles one bill.
• The MCP server is first-party and scope-split. Close publishes its own MCP endpoint with three scopes — mcp.read, mcp.write_safe, and mcp.write_destructive — authenticating over OAuth 2.0 or an API-key header, with documented clients for Claude, ChatGPT, Cursor, VSCode, and n8n (read-only). The separated destructive scope is the detail that matters: most CRM MCP servers hand you one token and one blast radius, so a hallucinated delete is indistinguishable from an intended one.
• Chloe reads the same objects the reps do. Because the agent runs inside Close rather than syncing into it, there is no replication lag and no second record of what was said on a call. Bolting an external AI-SDR onto a CRM buys you a sync job and a reconciliation problem; this does not.

Pricing reality

Four plans on annual billing: Solo at $9 per seat (one seat, 10,000-lead ceiling), Essentials $35, Growth $99, Scale $139. Monthly billing runs $19 / $49 / $109 / $149.

The real decision is Essentials versus Growth, and it is a 2.8× jump. Essentials gives you the record, the inbox, built-in calling and SMS. Workflows — Close’s sequencing — plus the power dialer and bulk email all start at Growth. Any team running outbound lands on Growth. Eight reps there is $792 per month, $9,504 per year.

That is the seat line, not the invoice. Calling is metered on top: eight reps at two hours of talk time a day is roughly 20,000 minutes a month, about $400 per month at the published rate — half the seat cost again. AI runs on a separate credit meter. Growth includes 1,500 credits per user per month against a 15,000 org ceiling, and Chloe voice calls, per-field enrichment, and post-call auto-updates all draw it down; credits reset monthly and do not roll over. Extra packs start at $20 per month for 1,200 credits and improve to $0.01 per credit at the highest tier. Lead summaries, the notetaker, AI rewrite, Chloe Chat, and AI search are free and never consume credits.

Budget the whole number. For eight reps on Growth with genuine dial volume and moderate Chloe use, plan on $12,000-14,000 per year, not $9,500.

Best for

Founder-led and SMB sales teams of roughly 3-25 reps whose motion is the phone — outbound dialing, inbound speed-to-lead, or both — that want one bill instead of a CRM plus a dialer plus a sequencer. Close is a direct answer to “Attio is too raw and Sales Hub is too expensive for eight reps, what is the third option,” and it answers it by being the only one of the three where a rep never leaves the record to place a call. ROI is best between roughly $1M and $20M ARR, where call volume is the constraint and there is no CRM admin to build the alternative.

Do not buy Close if reporting is the deciding requirement: analytics stop at fundamental pipeline metrics, and custom calculations or forecasting models mean exporting the data. Do not buy it for a flexible data model — Custom Objects exist but are Scale-only, every instance is tied to a lead, and the limits are 50 fields per object type and 100 instances per lead. Do not buy it as a marketing-and-sales system of record; that is HubSpot’s job, and Close does not pretend otherwise.

Versus the alternatives

HubSpot is the incumbent to beat in this segment. Sales Hub Professional is $90 per seat per month annual with a five-seat minimum and a one-time $1,500 onboarding fee, which puts eight reps at $10,140 in year one before any calling — within a few hundred dollars of Close Growth. Pick HubSpot when marketing automation and CRM must share one bill and one contact record; pick Close when the phone is the job and you would rather not meter minutes through a third party.

Attio is the fastest-growing AI-native entrant and the one Close is most often weighed against. Plus is $35 per seat annual and Pro $79, but Sequences and Call Intelligence both start at Pro, so Plus is a records tier rather than a selling tier. Pick Attio when the object model is what you are buying and your team lives in email; pick Close when your team lives on the phone and wants the pipeline shipped rather than modeled.

Pipedrive undercuts both — Growth at $24 per seat annual, Premium at $49 — and remains the cheapest way to run a pipeline for a small team. Pick Pipedrive when budget is the binding constraint and calling is occasional; pick Close when dialing is daily, because Pipedrive’s dialer is an add-on and Close’s is the product.

If none of these fit, the honest fallback is Apollo at $49 per seat per month for prospecting and sequencing plus a free or near-free CRM underneath. You trade a real system of record for a lower bill, which is the right trade under about five reps.

Watch-outs

• Chloe is English-only and US/Canada-only today. Close says other languages are on the roadmap, which is not a shipped feature. Guard: if any meaningful share of your calling is outside North America or in another language, scope Chloe out of the business case entirely and buy Close on the dialer and the record — then re-evaluate at renewal against what actually shipped.
• The tier you price is rarely the tier you need. Teams model Essentials at $35 and discover that sequencing, the power dialer, and bulk email are all Growth features, and that Custom Objects and role-based permissions are Scale. Guard: before signing, write down the five things your reps do daily, map each to the tier that includes it, and price the highest one. If that lands on Scale, compare against HubSpot again — the arithmetic changes.
• Usage is the part of the bill nobody forecasts. Calling minutes, phone numbers, the $50-per-month Call Assistant, and AI credits are all separate meters, and credits expire monthly. Guard: run a 14-day trial with your actual dial volume, read the usage page at day 14, and multiply by two before you take the number to finance. Set a credit-pack tier deliberately rather than discovering overage.
• Reporting is the recurring complaint in reviews, not the interface. Reviewers consistently rate the UI well and the analytics poorly: pipeline reports are not fully customizable, and deals that skip stages or move backwards do not report cleanly. Guard: if you need board-grade forecasting, plan on a warehouse sync through Fivetran or Airbyte and a BI layer from day one, and price that in rather than discovering it in quarter two.
• AI agents on the phone carry disclosure obligations. Close requires the agent to identify itself as AI and disclose recording, and calling-consent law varies by state and province. Guard: have whoever owns compliance approve the Chloe script and the call-recording notice before the first unsupervised dial, and keep a dated copy of the approved script with the call logs.