Gong
Gong is a Revenue AI OS for go-to-market teams. It captures and transcribes every customer call, email and meeting, analyzes what moves deals forward, forecasts pipeline, and runs AI agents that act on those findings.
What is Gong?
Gong calls itself a Revenue AI OS: one system that captures what happens between a company and its customers, works out what is actually driving revenue, and then acts on it. It began as conversation intelligence and has grown into a platform for whole go-to-market organizations — sales, marketing, customer success, enablement and revenue operations.
The architecture has three layers. The Gong Revenue Graph captures and connects every interaction — calls, video meetings, emails, CRM records — into a living map of revenue data; Gong claims it picks up 99% of the data that previously went uncaptured. The Gong Revenue Harness is the execution layer that governs how AI agents plan, act and deliver, handling context, permissions, human supervision and auditability. On top sit the Applications: Gong Engage for outreach and sequences, Gong Forecast for prediction and centralized analytics, Gong Enable for coaching anchored in real customer conversations, and Gong Collective for integrations and partner services.
Sixteen pre-built agents ship with the platform, among them AI Tracker, AI Theme Spotter, AI Briefer, AI Call Reviewer, AI Trainer, AI Composer, AI Deal Predictor, AI Revenue Predictor, AI Activity Mapper, AI Transcriber and AI Translator; AI Deep Researcher is announced as coming soon. Teams that need something else build their own in Agent Studio, describing the agent in plain language instead of writing code.
The models behind all this are trained on more than 3.5 billion sales interactions, which Gong credits for transcription accuracy of 85-90% against 75-80% for generic models. The platform works in over 70 languages and handles recording consent in as many. Everything runs in a browser with nothing to install, plus iOS and Android apps.
Gong reports more than 5,000 customers and over 2,000 employees, and holds ISO/IEC 42001 for AI management alongside SOC 2 Type 2 and a long list of security certifications. Customer data is never used to train generative AI models. Access is enterprise-shaped: no self-service signup, no public pricing, and a sales conversation before anything begins.
What it does
- Automatically capture and transcribe calls, video meetings, emails and appointments
- Analyze conversations to surface deal signals, objections, recurring themes and risks
- Forecast revenue and pipeline health through AI deal and revenue predictors
- Keep the CRM current automatically with activity mapping and data extraction
- Draft and personalize prospecting and follow-up emails at scale
- Coach and train reps, including AI role-play and automated call reviews
- Answer natural-language questions across all revenue data, in more than 70 languages
When to use Gong / When not to
A quick filter to help you decide if Gong is the right fit.
When to use Gong
- Enterprise sales teams and account executives who want every call, email and meeting captured and analyzed without manual note-taking
- Revenue operations leaders and CROs who forecast pipeline and need deal risk surfaced before the quarter closes
- Sales enablement and L&D managers coaching reps on real customer conversations rather than scripted role-play
- Customer success and renewal teams tracking account health across long, multi-threaded relationships
- Multilingual, multi-region go-to-market organizations selling across more than 70 languages
When not to use Gong
- Solo founders and very small businesses without a structured sales team: licensing is per user plus a platform fee, quoted only after a sales conversation
- Anyone who wants to try a product the same day, since there is no self-service signup, no free plan and no advertised free trial
- Individuals wanting to record personal conversations, as the terms bar non-business calls and the services are not intended for personal or household use
- Direct competitors of Gong and anyone running benchmarking exercises, both explicitly prohibited by the terms of service
- Teams without corporate email and calendar on Microsoft 365 or Google Workspace, or without standard conferencing tools, since capture depends on those connections
How to use Gong
A typical end-to-end flow, from setup to results.
- Book a demo or take the guided product tour, since there is no self-service signup
- Go through a sales consultation covering goals, team size, modules and integrations, then receive a tailored proposal
- Connect corporate email and calendar through Microsoft 365 or Google Workspace
- Connect conferencing and telephony tools such as Zoom, Microsoft Teams, Google Meet and Webex
- Connect the CRM (Salesforce, HubSpot and others) for two-way sync — recommended rather than mandatory
- Assign user licenses and set up access with SSO via SAML 2.0 or OIDC, MFA and role-based permissions
- Configure capture: recording rules, email ingestion, automatic PII redaction, deny lists, filters and retention policies
- Let Gong capture and transcribe automatically while the agents run inside existing workflows
- Configure the pre-built agents or create custom ones in Agent Studio, in plain language
- Start with a pilot on one team, business unit or region, measure the results, then extend the rollout
Pros & Cons
Pros
- Models trained on more than 3.5 billion sales interactions, with claimed transcription accuracy of 85-90% against 75-80% for generic models
- Rare functional breadth in a single system: capture, analytics, forecasting, engagement, enablement and AI agents
- Dense security and compliance posture: SOC 2 Type 2, ISO/IEC 42001, 27001, 27017, 27018 and 27701, PCI DSS SAQ-D, CSA STAR and Data Privacy Framework certification
- Explicit commitment that customer data is never used to train generative AI models, with no co-mingling between customers
- Sub-processors published by name with their purpose and location, plus a subscription to change notifications
- EU data residency available on request in AWS Frankfurt, alongside the default US region
- More than 300 integrations and an open API, with 99.5% claimed uptime and a public status page
Cons
- No public pricing whatsoever: no amount, no tier, no rate card — everything goes through a quote
- No free plan, no advertised free trial and no self-service signup
- Two-part billing: per-user licenses plus a platform fee indexed on the number of users supported
- Deployment is heavy — email and calendar, conferencing, CRM and SSO connections, plus capture and redaction rules to define
- Sized for global enterprises and the commercial segment rather than small businesses
- API documentation is not public: access runs through an application to the Gong Collective
- The Trust Center and Help Center portals are gated, so detailed documentation is not open to prospects
Pricing & Plans
Gong publishes no prices. There is no free plan and no advertised free trial, and no self-service signup: the pricing page offers only a quote request, segmented by team size (1-50, 51-1,000, 1,001-9,999 and 10,000+ users). Three rules are published without figures — licenses are priced per user, a platform fee applies based on the number of users supported, and integrating an existing tech stack is free of charge. Pricing is described as consultative, with a tailored proposal built around user count, the modules chosen (forecasting, engagement, enablement) and the integrations required. The terms of service add that fees may be subscription-based, usage-based or otherwise, as set out in the applicable Order Form, that subscriptions renew automatically for one-year periods unless notice is given 60 days before renewal, that invoices fall due within 30 days with 1.0% monthly late interest, and that fees are non-refundable. A Gong Credits scheme exists but carries no published figures. Consequently, no lowest price point and no currency can be stated for this tool.
Data, GDPR & hosting
A consolidated view of how Gong handles your data.
GDPR overview
GDPR implementation is documented rather than merely declared. Gong publishes a full Data Processing Addendum, incorporated by reference into its terms, carrying the European Commission's Standard Contractual Clauses in modules 2 and 3, with the docking clause excluded. Gong.io Inc. is certified under the EU-U.S. Data Privacy Framework, its UK extension and the Swiss-U.S. DPF, and parent company Gong.io Ltd. sits in Israel, a jurisdiction recognized as adequate. A Data Protection Officer is named — Aner Rabinovitz of PrivacyTeam Ltd., at dpo@gong.io — and Prighter Group acts as Article 27 representative for the EU and UK. Access, rectification, erasure, restriction, objection and portability requests go to privacy@gong.io. Gong commits to cooperating with EU supervisory authorities, the UK ICO and the Swiss FDPIC, and holds ISO/IEC 27701 and 27018 certifications.
Who owns the data?
Customers keep ownership of their data. The terms state that Gong acquires no right, title or interest in Customer Data; the customer grants only a non-exclusive, worldwide, limited-term, royalty-free license to host, use, copy, transmit and display it so Gong can run the service and produce insights. Gong retains all rights in the platform itself, and customer feedback is licensed to Gong perpetually and irrevocably. Under the privacy policy Gong is processor for Customer Data, controller for prospect data and both for user data, while the customer stays responsible for its own legal bases. A published DPA with named sub-processors governs onward sharing.
Reuse rights
Inside the customer's own organization, reuse is broad and needs no permission: transcripts, recordings, insights and analytics are available to employees and consultants. Outside that circle the terms are restrictive. Content generated by the service may not be made available to anyone other than employees and consultants, with one explicit carve-out for call transcripts and recordings, which may be shared. Reselling, transferring, renting or sharing the rights is prohibited, as is exploiting the service commercially, benchmarking it or using it to build a competing product. On termination, customers may request an export or download of their Customer Data within 30 days, in structured and standard formats agreed with the designated administrator; beyond that window Gong has no obligation to retain anything.
Data retention & training
Hosting summary
Gong runs on Amazon Web Services. The default data region is the United States, in AWS US data centers, and EU data residency is available on request, with data stored in Frankfurt, Germany (AWS eu-central-1). The privacy policy is broader, stating that data is maintained, stored and processed in the United States, Europe, Israel and other locations. Parent company Gong.io Ltd. is established in Israel, recognized as adequate by the European Commission, the Swiss FDPIC and the UK Secretary of State. Transfers outside adequate countries rely on Standard Contractual Clauses approved by those same authorities and the ICO, and Gong.io Inc. is certified under the EU-U.S. Data Privacy Framework, its UK extension and the Swiss-U.S. DPF. The sub-processor list is public and gives each location: AWS, MongoDB, Mailgun, Twilio, Google Cloud EMEA, Snowflake and Microsoft Ireland Operations across the US and EU, Gong.io Ltd. in Israel and Gongio Ireland Limited in Ireland. Customer environments are segregated, data is encrypted at rest and in transit, and access is logged and auditable.
Things to keep in mind
Risks and trade-offs to weigh before adopting Gong.
- Calls and meetings are recorded systematically, and obtaining consent from the people recorded is the customer's responsibility, not Gong's
- Employee monitoring: call reviews and scorecards evaluate individual reps, and metrics can quietly turn into the whole of management
- Optional voice identification can be enabled by the customer and then opted into by users — a deliberate decision, not a default
- Session recordings, activity logs and copy-paste keystrokes are collected, including during Gong Academy exams
- Dependency risk: once the revenue graph becomes the system of truth, leaving means exporting within 30 days before permanent deletion
- Handing judgment to agents — forecasting, deal scoring, prioritization — is easy; Gong provides for human validation, but someone has to actually exercise it
- Claimed transcription accuracy of 85-90% still leaves 10 to 15% of errors, worth remembering before acting on any transcript
Setup & Integrations
Technical difficulty
Simple for end users, involved for the organization. Nothing to install: Gong runs in a modern web browser, with iOS and Android apps alongside. The work sits with IT and RevOps — connecting Microsoft 365 or Google Workspace, conferencing tools and the CRM, then SSO, MFA and role-based access, and finally agreeing recording rules, email ingestion, PII redaction, deny lists and retention. Gong provides scoping, implementation and enablement support, and Agent Studio claims agent setup needs no technical resources. Integration difficulty features in Gong's own top FAQs. Piloting on one team first is the recommended route.
Deployment
Apps stores
Integrations
Behind Gong
Fundraising
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
What is Gong?
Who is Gong for?
What do I need in place to use Gong?
How much does Gong cost?
Is there a free trial or a free plan?
Is my data used to train Gong's AI models?
Where is customer data hosted?
How long does Gong keep customer data?
What security and privacy certifications does Gong hold?
Does Gong offer integrations, an API and mobile apps?
Should you pick Gong?
Gong is a mature product with a clear center of gravity. Founded in 2015, backed by $584 million in funding, staffed by more than 2,000 people and used by over 5,000 customers, it has moved beyond the conversation intelligence category it helped create and now presents itself as a Revenue AI OS: capture, analysis, forecasting, engagement, enablement and governed AI agents in one system.
Its strengths compound. The corpus behind the models, 3.5 billion sales interactions, underpins transcription and analysis quality that generic tools struggle to match. The functional coverage is unusually wide, which matters when the alternative is stitching four vendors together. And the compliance posture is genuinely dense: SOC 2 Type 2, ISO/IEC 42001 for AI management, 27001, 27017, 27018 and 27701, Data Privacy Framework certification, a published DPA, named sub-processors, and an explicit promise that customer data never trains generative models. That last commitment is rarer than it should be.
The structural weakness is just as plain: total pricing opacity. No published amount, no free plan, no advertised trial, no way to sign up. Evaluating Gong means opening a sales cycle first, which rules it out for small teams and makes budgeting impossible beforehand. Adoption is not lightweight either, between email, calendar, conferencing and CRM connections, recording consent governance and real change management.
The verdict follows the audience. For an organization with a structured go-to-market team, a real integration stack and the appetite to measure what sellers actually do, Gong is among the most complete options available. For a small business, a solo founder or anyone who wants to test before talking to a salesperson, it is the wrong door.
- Choosing a selection results in a full page refresh.
- Opens in a new window.