Private user: email helper

A tactical email operator for teams that cannot afford to leak context, time or data.

The problem

Fifteen operators, consultants and account leads, each running a dozen live client threads a day. Every reply costs the same hidden tax: re-reading the thread to rebuild context, deciding the right tone under pressure, drafting from scratch, and hoping nothing was over-promised in the rush. The slow replies lose deals. The fast ones leak commitments nobody authorised. And every "helpful AI" on the market asks the same price for the privilege: your client correspondence, uploaded to someone else's cloud, trained on, metered, and filtered through alignment rules you did not write.

What we built

NexOps Operator is a browser-side assistant that sits in a side panel next to the inbox the team already uses. One button reads the open email, decides the strongest tactical response on its own, and hands back a finished draft plus a one-line read of what the sender is actually doing and what pressure they are applying. The operator reviews and approves. That is the entire workflow. It is deliberately an assistant, not a cockpit: no menus to learn, no modes to configure, no training afternoon.

Time

The model is conservative on purpose. Baseline: a substantive operational reply takes an operator around six minutes to read, contextualise and compose. With Operator: analyse, review, approve, roughly ninety seconds. Call the saving four minutes per reply, twenty replies a day per person.

Across fifteen users that is about twenty hours recovered every working day, or roughly 4,400 hours a year. At a productive full-time load of 1,800 hours, that is the equivalent of about 2.4 people given back to revenue-generating work. At a loaded cost of £40k per head, on the order of £90k to £100k a year, before counting a single faster reply that closed a deal sooner.

Convenience

The design principle is that the tool should think, not ask. Operator picks the tactic itself based on the email and an optional one-line intent from the user, so nobody chooses from a menu of seven options under time pressure. Alternatives exist, but they stay collapsed until you want them. A background red-team audits every draft before it can be sent and flags the four ways a reply goes wrong: it breaks the chosen tactic, it states something outside the company profile, it commits to something nobody authorised, or it drifts into filler. A commitment ledger quietly records every real promise a draft makes, per contact, so nothing slips through a busy Friday.

Integration

No new mail client, no migration, no IT project. Operator runs as a native side panel inside the browser the team already has, on Chrome, Brave and Edge, directly against Gmail and Outlook 365 on the web. It detects which mailbox it is looking at and adapts. Deployment to fifteen seats is load-and-go, with zero per-user configuration beyond pasting a key once. It meets the team inside their existing workflow rather than asking them to move house.

Security

This is where Operator stops being a convenience and becomes a policy decision. The API key is stored locally on each machine and never touches a central server. There is no NexOps backend aggregating anyone's mail, no telemetry, no usage tracking, no training on client correspondence. The red-team acts as a quality gate so a careless draft cannot quietly promise the wrong thing on company letterhead. And the company profile that grounds every reply is held by the client, not by us.

One honesty note, because our clients expect it: this deployment generates drafts through a commercial model API, so the email content does leave the building for inference. For clients who require zero data egress, the identical operator workflow runs fully on-premise on NexOps Private AI, inside the client's own network, on their own hardware, at a fixed cost regardless of volume. Same assistant, no data leaves the walls.


The numbers, plainly