Everything we've shipped.

Scheduling now runs against your connected calendar rather than a pasted link. Availability is computed in real time across Google and Outlook, double-booking is prevented at the slot level, and every confirmed meeting flows into the CRM with the originating campaign attached. Google Meet and Teams links are generated with the invite, and a prep brief assembled from the account's full history lands before the call.

The dialer connects calls, meetings, and email threads to a single timeline per lead, so reps stop stitching context together across tools. Numbers are provisioned per workspace, call outcomes are logged automatically, and the queue panel lets you work a list of warm leads end to end.

The MCP server exposes our core workflows as 27 typed tools — lead discovery, list management, campaign control, sending, analytics, and meeting scheduling. Keys are scoped per organization so an assistant only ever touches the data it should. This was a deliberate platform bet: outreach is increasingly orchestrated by agents, and we wanted Gripwise to be the system of record those agents act through.

A classification layer reads every inbound message and assigns intent labels, so positive replies never get buried under auto-responders. Hot threads bubble up first, and the labels feed downstream automations — booking a meeting on interest, pausing a sequence on an objection, or honoring an opt-out instantly.

The visual builder lets you mix channels in a single sequence, branch on opens, replies, and label changes, and set per-step delays that respect each lead's timezone and working hours. Sequences pause themselves the moment a lead replies, so no one gets a follow-up after they've already said yes.

We migrated the product to an organization-first model. Plans, inboxes, leads, and campaigns are scoped to a workspace, row-level security enforces isolation at the database, and seats can be added or removed without touching billing for the rest of the team. This was the foundation that unlocked agencies running many clients side by side.

The new warmup engine paces volume per inbox, gradually building sender reputation while monitoring bounce and spam signals. Alongside it, the deliverability checker validates SPF, DKIM, and DMARC and flags configuration issues before a campaign ever sends. Daily sending limits were added per inbox to keep accounts healthy as volume scales.

By integrating LinkedIn directly, a sequence can touch the same person on two channels from one schedule instead of two racing campaigns. Connection requests, profile visits, and message sync run within safe daily limits, and every interaction is mapped back to the lead so the timeline reflects the full multi-channel conversation.

We connected a large contact database to a natural-language search layer, so you can find leads by intent signals like recent funding, active job posts, or new product launches instead of fiddling with rigid filters. Results enrich automatically and drop straight into a list ready for outreach.

The generator pulls signals from a lead's company and role, then writes a first line and angle tailored to them — not a mail-merge token. This was an early product direction we committed to: personalization at scale only works when the research is real, so we built the research step into the core of every send.

We shipped the data model the rest of the product is built on — leads, lists, companies, and a usage-based wallet that meters discovery and enrichment fairly. Getting this layer right early is why later features like sequences, the dialer, and MCP could move quickly.