Replies and what the numbers mean

Outreach is only half the job. What happens after someone replies is where the revenue actually is.

The Inbox

Replies to your outreach come into Inbox classified and threaded, so you are not reading a flat pile of messages trying to work out which are worth your time. Each reply is sorted as interested, neutral question, not now, referral, unsubscribe, wrong person or negative, and you can change the classification if it is wrong. Filter with All, Unread and Interested, or search by sender, subject or message text.

A reply stops the rest of that person's campaign emails, so nobody gets a follow-up after they have answered.

To answer, open the reply and either write your own and click Save draft, or click Draft with AI for a suggestion grounded in the conversation and your locked voice. Without a locked voice, you get a starter template instead. Either way, you review and edit it.

A draft sits there until you click Approve and send. That sends one email through your connected mailbox, and it counts toward that mailbox's daily cap. Nothing goes back out on its own. The same principle as the rest of the product: it prepares, you decide.

While sending is switched off for your workspace, Approve and send runs as a dry run and sends nothing. The draft is marked approved. Once sending is on, click Approve and send again to send it.

Analytics, and why it looks conservative

Analytics deliberately keeps different things apart, and it is worth understanding why, because most outreach tools blur them into one flattering number.

• Provider accepted. The email provider accepted the message for delivery. This is not proof it arrived.
• Confirmed delivered. Delivery was actually confirmed.
• Human replies and Positive or referral. A person actually replied, and how many of those were interested or a referral.

Bounces, complaints and replies not yet classified are counted separately under Delivery attention.

Collapsing these into a single "delivered" or "success" figure is how outreach dashboards end up flattering and useless. Keeping them separate means the number you look at is the number you can act on.

You will not find open or click rates here. Noli does not report them without authoritative event data, and that is a deliberate choice rather than a missing feature.

Outcome counts stay at zero until durable events actually exist. An empty report is an honest report, not a broken one. Campaign outcomes only appear once a play has become an approved campaign, and the counts you see are for each campaign's current approved version.

Usage and tokens

Settings, then Usage, lists each research run with Estimated tokens, what it was quoted, and Reconciled tokens, what it actually settled to. A run that has not settled yet says Not settled yet. It also shows the projected send tokens for each campaign draft.

Outside data is the main cost, and the Strategist's own AI turns are metered too. It all comes out of the same monthly token pool as the rest of Noli, so there is no separate balance to top up. If no provider operations have been recorded yet, it says so rather than showing a fabricated zero-state. Pricing a run reserves nothing and costs nothing, and every actual run shows its quote before it can reserve tokens.

Daily lead finding

Daily lead finding keeps a campaign topped up without you starting each run. On a campaign's Settings step you set a daily target of qualified people, a raw candidate ceiling, and a daily token cap of up to 2,500,000 tokens a day. It runs one cycle per campaign each weekday and only queues people for your review. It never sends anything.

This is released in stages. If the panel says it is not active in your workspace yet, it has not been switched on. See Keep the queue full with daily lead finding.

Related: Review and approve a campaign and Create plays and run research.