B2B marketing operating system
Email campaigns, the lead pipeline, social posts, ad accounts, market signals and attribution in one workspace. AI drafts and recommends. People approve.
Last 30 days, compared with the previous 30 days.
1 warning · 5 for review
Autumn outreach paused itself. Check the sending domain.
Logistics follow-up, submitted by a colleague.
Scheduled for this week once approved.
Emails and replies per day (UTC)
Built-in connectors
See all integrationsFeatures
Each area has its own screens and shares the same contacts, permissions, approvals and audit log.
Campaigns that pass a safety check before they leave.
Write a post once, adapt it for each account, publish after approval.
Recommendations first. Changes only inside the limits you set.
State an objective, approve the plan, follow the run.
Collect what is happening, and keep facts apart from inference.
Know who you may contact, and on what basis.
Turn replies into leads, tasks and meetings.
Use the models you choose, within the budgets you set.
Numbers that say what they rest on.
Decide who may do what, and see who did.
How it works
The same five steps apply to an email campaign, a social post and a change to an ad account.
Add an email provider, an AI provider and your social and ad accounts. Credentials are stored encrypted and are not shown again.
Import contacts, read the quality report, record consent and group people into segments.
Write templates, campaigns and posts, with AI drafts where you want them. A second person approves the exact content.
Approved campaigns send inside their sending window. Approved posts publish at the time you scheduled.
Read replies, leads, meetings and attribution, then act on the recommendations.
Safeguards
AI provides drafts and recommendations. Fixed rules and signed-in people decide what happens. These safeguards are part of the product, not settings to switch on.
Campaigns, social posts and ad changes are approved by a person, by default not the one who asked. An approval covers the exact content it was given for.
Agents run without permissions. They cannot send, publish, approve, change an ad account or lift a suppression, whatever they are asked.
Whether a change to an ad account is applied is decided by fixed rules over limits your administrators set, never by a prompt.
A campaign passes a checklist on the server before it is submitted, approved, scheduled and launched. A blocking failure stops it.
A suppressed address is left out of every send. Only a signed-in person with the permission can lift a suppression.
Observations and insights are labelled External data, Internal performance or AI inference. An inference is never stored as an observed fact.
Sensitive changes are recorded in an append-only log. Each entry is chained to the one before it, so a changed or removed entry is detected.
Keys and tokens are stored encrypted, shown at most as a masked fragment, and never given to an AI model.
Every record belongs to one organization, and data is read and written within that organization only.
Every action checks a permission. A role is a set of permissions that an administrator can tailor.
Integrations
Connections are made by an administrator inside the product, with a login button or a key, depending on the service.
Choose the model each agent uses.
Publish approved posts.
Read accounts; change them within limits.
Send through your own provider.
Receive bookings and inbound replies.
Bring in contacts and observations.
Each connection uses your own account with the service, and is subject to that service's terms and to the access it grants.
Third-party trademarks: the names and logos of other products and services belong to their respective owners. They are shown only to identify the integrations, and do not imply endorsement.
Who it is for
People see the areas their role may open, and nothing else.
FAQ
A workspace for B2B marketing and client acquisition. It covers contacts and segments, email campaigns and automations, the reply inbox and lead pipeline, social publishing, advertising, market intelligence, analytics and attribution, with AI agents that draft and recommend.
No. Agents run without permissions: they cannot send, schedule, approve or launch a campaign, publish a post, change an ad account or lift a suppression. What an agent produces is a draft, a recommendation or an insight, and it stays inert until a signed-in person acts on it.
Several things. A campaign passes a safety check before it is submitted, approved, scheduled and launched. Campaigns, social posts and ad changes are approved by a person, by default not the one who asked, and an approval covers the exact content it was given for. Suppressed addresses are left out of every send. Every action checks a permission, and sensitive changes are written to the audit log.
OpenAI, Anthropic, Google Gemini, xAI and DeepSeek for AI; LinkedIn, Facebook Pages, Instagram, Threads, X, YouTube, Google Business Profile and TikTok for social; Meta Ads and Google Ads for advertising; Amazon SES or any SMTP server for email; Cal.com and Calendly for meetings; and Apify as a lead source. The connectors are built in, and each one uses your own account with the service.
Yes. An administrator connects one or more providers, chooses which model each agent uses and sets budgets that cap what AI may spend.
Each organization's data is kept apart from every other's. Access is checked by permission on every action, two-factor authentication can be required, credentials are stored encrypted and cannot be read back, and sensitive changes are written to an audit log whose integrity can be verified. Backups and the security of the server belong to whoever runs the installation.
Not yet. There is no public sign-up today: an administrator invites each person to an existing workspace. A hosted service with self-service sign-up and plans is planned and is not available yet.
A user guide is built into the product under Help, with a chapter for each screen, a quick start and a glossary. It opens after you sign in.
Accounts are created by invitation from an administrator. If you have one, sign in to continue.