The agent that runs your go-to-market.
Operator is the part of LaunchSurface that does the work. It picks who to reach, writes the outreach, sends it from your mailbox, reads what comes back, and schedules its own next run.
What is LaunchSurface Operator?
Operator is the autonomous agent at the centre of LaunchSurface. On a cadence you choose, from every 3 hours to once a day, it runs a full go-to-market cycle for one product. It reads fresh buying signals, picks the prospects who fit your ideal customer, writes a two to five step email sequence for each, sends from your own mailbox, and reads the replies. Each cycle draws on a prepaid wallet and stops when the wallet, or the daily or monthly cap you set, cannot cover the next action. Everything it does is logged with its reason, and anything you have marked for approval waits for you.
The default time between Operator cycles. Slower cadences run every 6, 12 or 24 hours, each with a little jitter.
In Operator's experiments a reply scores 1 and a click 0.25. An open scores nothing. A bounce counts as 2 failures and a spam complaint as 3.
Median cost of one discovery cycle, almost all of it contact enrichment. The measured range is 31¢ to 52¢.
Reads the market
Starts each cycle from the buying signals Signals collected since the last run.
Picks the prospects
Ranks prospects by fit to your ideal customer and by how fresh their signal is.
Writes and sends
Drafts a two to five step sequence per prospect and sends from your connected mailbox.
Learns from replies
Scores every variant on replies, not opens, and shifts toward what works.
A cycle you can watch, step by step.
One cycle, end to end, on its own.
A cycle moves from signals to sends to reading outcomes without a hand-off. It ends when the work is done, when your wallet cannot cover the next action, or when your daily or monthly cap is reached.
Scored on replies, not opens.
Operator runs subject lines, angles and timing as Thompson-sampling experiments. A reply is a success, a click is partial credit, and an open is ignored because it no longer says anything reliable. Bounces and complaints count against the variant that caused them.
Spends its budget on the likeliest prospects.
Prospects are ranked by fit to your ideal customer and by the strength of their signal, so each cycle reaches the people with the best reason to reply first.
Built to run, not to sit in a dashboard.
It schedules itself.
After each cycle Operator enqueues the next one on the cadence you picked. There is no campaign to relaunch.
Every send passes a gate.
Suppression list, address verification, a per-prospect weekly cap and a per-inbox daily cap are checked before anything leaves your mailbox.
It shows its reasoning.
Each cycle records the signals it used, the prospects it chose, what it sent and why, so you can audit any action.
The details, answered.
How autonomous is it, really?
Operator runs whole cycles without waiting on you: it chooses prospects, writes sequences, sends, and reads replies. You decide the boundaries. The cadence, the daily and monthly spend caps, and which kinds of change need your approval are all settings in your operating contract. Replies from large accounts (1,000 or more employees by default) pause automated responses and come to you for review.
Can I approve messages before they go out?
Yes. Drafts can wait in an approval queue before they send, and every message that does go out is logged with the prospect, the signal behind it and the full text.
How does it avoid burning my domain?
Every send goes through a gate. By default it will not email the same person more than twice in 7 days, will not send more than 20 emails from one inbox in 24 hours, requires a verified address for cold email, and follows your warmup ramp. Every 6 hours it also checks each inbox's last 7 days and pauses any inbox whose bounce rate passes 4%, whose complaint rate passes 0.3%, or whose reply rate falls under 2% after 100 sends.
What does it cost to run?
Your plan includes a monthly send allowance and Operator credit. Usage beyond that, such as extra enrichment, is debited from a prepaid wallet at cost plus 60%. A measured discovery cycle costs about 45¢ at the median, and a cycle stops cleanly when the wallet or your cap cannot cover the next action.
Can I steer it?
Yes. You can edit the ideal customer, exclude accounts, change tone, or leave standing instructions in plain language. Operator reads them at the start of every cycle, so direction carries forward instead of resetting.