Sybill

What it is

Sybill records sales calls, writes the summary, and fills the CRM fields the rep would have filled that evening and did not. The category leader it most resembles is Gong — same recording-to-deal-context arc, same ambition to be the place a manager checks before a pipeline review. The difference is the price tag and how you get it: Sybill publishes a per-seat number on its website, sells single seats, and charges no platform fee.

The 2026 product is four things. An invisible recorder joins Zoom, Google Meet, Microsoft Teams, and Webex without a visible bot tile. Magic Summaries return the call’s intent, outcome, next steps, and each participant’s stated pain points. Ask Sybill answers questions across calls, emails, CRM records, and Slack. CRM Autofill discovers the fields in your Salesforce, HubSpot, Zoho, or Dynamics 365 instance — custom fields included — and writes per-field AI output into them: budget signals, decision-maker identification, objections, agreed next steps, MEDDPICC and BANT criteria, competitor mentions.

One thing to name plainly, because it is why many people have heard of Sybill. The non-verbal behavioral analysis that anchored the original pitch — reading engagement from video — is no longer on the marketing site. The 2026 positioning is deal context, autofill, and agent access.

Why it shows up in RevOps stacks

• It is the published-price pole of a quote-only category. Gong, Chorus, and Attention all route you to a demo form. Sybill’s Business tier — the one that writes to the CRM — is $90/user/month on annual billing with no seat minimum and no platform fee. A 6-rep team can buy it on a card this afternoon.
• Autofill maps itself. Sybill auto-discovers CRM fields rather than asking an admin to map them, which is the step that turns a two-week CI rollout into a two-month one. The vendor puts first-value at roughly 15 minutes, tuned against your last 30 deals; treat that as a vendor claim and verify it in the trial.
• The MCP server is first-party and per-user scoped. The endpoint is https://mcp.sybill.ai/mcp, reached through mcp-remote from Claude Desktop, Claude Code, Cursor, or VS Code, with a browser OAuth sign-in and no API key to rotate. Reads are scoped to what the signed-in rep can already see, so a manager’s agent and a rep’s agent return different answers from the same query. MCP access starts on Business.

Pricing

Per user per month, checked 2026-09-03 against sybill.ai/pricing and Sybill’s credits guide. Annual billing runs 20% under monthly.

• Free — $0. 500 credits/week. Unlimited call recording, the invisible recorder, 20 AI meeting summaries and briefs per month across the team, AI follow-up emails, mobile app. A 14-day Business trial runs without a card.
• Pro — $30 annual, $36 monthly. 2,500 credits/week. Unlimited summaries and briefs, customizable summary formats, unlimited storage, CRM push.
• Business — $90 annual, $108 monthly. 5,000 credits/week per user. CRM context inside Ask Sybill, CRM Autofill capped at 10 fields, API and MCP access, custom email templates, two-way Slack and Teams, the deal workspace.
• Enterprise — quote-only. Unlimited credits, unlimited autofill fields, SSO, data migration support, dedicated account manager.

The tier you need is almost always Business, because autofill and MCP are the reason to buy this rather than a notetaker. Ten reps on Business annual is $10,800/year, all-in, with no separate platform charge. The same ten reps on Pro is $3,600/year and writes nothing meaningful into the CRM.

Against the leader: Gong’s licenses run roughly $1,400-$2,000 per seat/year on standard-tier deployments plus a mandatory platform fee reported between $5,000 and $50,000 — and Gong’s practical floor is 15-25 seats, so a ten-rep team is not a customer at any price. Credit overages have their own line: admins buy 5,000-credit bundles at $100 each, and bundle credits are shared across the team while plan credits are per seat.

Best for

AEs and full-cycle sellers on teams of 3-30 who need CRM fields populated after every call and cannot clear a six-figure revenue-intelligence contract — plus the RevOps operator who wants an agent reading deal context over MCP without standing up middleware first.

Not for

Teams whose deliverable is the forecast. Sybill has a deal workspace and deal inspection, not the weighted-pipeline modeling and forecast-submission workflow that Clari and Gong sell. It is also wrong for a coaching program run at scale across 100+ reps, where scorecard libraries and call-review queues are the product you are actually buying. And check your CRM first: Salesforce, HubSpot, Zoho, and Dynamics 365 are live, while Pipedrive and NetSuite are listed on the integrations page as not yet available — treat them as absent when scoping.

Versus the alternatives

• Gong — the segment leader. Pick Gong when 50+ reps, a coaching program with scorecards, and forecast rollups are the reason for the purchase, and when a $55,000 median annual contract is a normal line item. Pick Sybill when the deciding question is whether the CRM is current on Friday.
• Chorus — sold as part of a ZoomInfo platform contract, median $33,500/year, not as a standalone seat. Pick Chorus when you are already buying the contact database and want conversation intelligence on one paper. Pick Sybill when you want the recorder without the data subscription attached.
• Fathom — the low-cost pole at $25/seat/month annual on Business, which includes CRM field sync. Pick Fathom when four or five people need notes and one or two need writeback. The cap is the catch: CRM sync on Team and Business is limited to 3 users per domain, so a 12-rep rollout lands on an Enterprise quote and Sybill’s flat $90 becomes the cheaper predictable number.
• Attention — the fastest-growing entrant, selling agentic CRM write-back and quote-only, estimated in the $70-140/seat/month band. Pick Attention when you want agent actions as the unit of value and will negotiate for them. Pick Sybill when you want a price you can read on a website and a credit meter you can count.

Watch-outs

• Credits meter the agent surface, and MCP is the expensive part. Summaries and pre-meeting briefs cost 0 credits, but a deal analysis run through MCP costs roughly 300-400 credits, pipeline or win/loss analysis 50-100, and a generated deck or dashboard 100-200. Business gives 5,000 credits/week per seat, which is about 12-16 MCP deal analyses per rep per week before features pause. Guard: multiply intended agent runs per rep per week by 350 before signing; if the product exceeds 5,000, price the $100 credit bundles into the annual number or open the Enterprise conversation now rather than in month three.
• The 10-field autofill cap binds faster than it sounds. MEDDPICC alone is seven or eight fields. Add next steps and competitor mentions and you are at the Business ceiling with nothing left for the two custom fields your ops team actually reports on. Guard: write the exact field list before the demo, count it, and if it clears 10, get Enterprise pricing in the same cycle instead of buying Business and upgrading under pressure.
• Sybill’s own marketing pages disagree with its pricing page. Several blog and feature pages still cite $79/user/month for the Business tier, while the pricing page and the credits guide both state $90 annual and $108 monthly as of 2026-09-03. Guard: negotiate against the pricing page and the credits guide by URL and date, and get the per-seat number and the weekly credit allowance written into the order form.
• Call-type classification is where reviewers report it slipping. G2 reviewers describe Sybill misreading internal meetings and coaching calls as customer calls, and autofill inherits that mistake by writing deal fields from the wrong conversation. Guard: restrict recording to calendar events that have an external attendee, then audit the first 30 autofilled opportunities against their source calls before letting it write unattended.
• The behavioral-signal capability that put Sybill on your shortlist is not what is being sold in 2026. Guard: if the evaluation started from a demo of engagement scoring, confirm in your own trial call that the output still appears, and keep it out of the business case either way — build the case on autofill coverage and rep hours returned, both of which you can measure in the 14-day trial.