Revive deals that went cold

Closed-lost is where deals go to be forgotten. The reason for losing gets typed into a dropdown, the thread stops, and eighteen months later nobody remembers that the only blocker was a budget cycle that has since turned over twice, or a feature that shipped last spring. Re-reading them is a job that would take a rep a full day and produce maybe four emails worth sending, so it never gets to the top of anyone's list.

Everything needed to revive a closed-lost deal is already written down, in four different places. The CRM holds the loss reason, the public record holds what has happened at the company since, the profiles hold who is still there, and the issue tracker holds what you have since built. The expensive part is reading every deal history against four systems to find the few worth reopening, and that reading is what an AI agent does here once a month.

Which closed-lost deals are in scope

The run lists the deals in Attio marked lost more than 90 days ago. Ninety days is the floor: inside it the reason for losing is almost always still true, and the email reads as not having listened. Twenty-four months is the ceiling, since past that the people involved have moved on and the message is a cold email wearing a follow-up's clothes. Three kinds of deal are excluded outright: those lost to a competitor on a multi-year contract that has not yet expired, anyone who asked not to be contacted again, and accounts with an open opportunity right now.

Why the closed-lost reason field is not enough

The closed-lost reason field is treated as a hint. It is a dropdown chosen at the end of a bad week, and almost everything lands on "price". The notes, logged calls and email thread underneath tell a different story often enough to change the email entirely, since "price" often turns out to mean no budget this year, or a security review that never got through. "I know price was an issue" is a template. "I know the security review stalled in October" proves somebody read the history.

Each deal is classified into one of four reasons, with the evidence recorded: timing or budget, a missing capability named precisely, lost to a competitor, or no decision. The precision is what the later steps rely on. "Lost on features" cannot be matched against anything. "Lost because there was no audit trail" can be searched against every issue closed since, and when that search comes back with a release date the win-back has a subject.

What changed since the deal was lost

Three things are checked for each account, every item with a date.

• The company. Apify looks for hiring, especially in the function you sell into, along with funding, an acquisition, a new office or market, and leadership changes in or above the buying role. The source URL is kept for anything the email will reference, because a win-back that cites something the reader can check is a different message from one that gestures at momentum.
• The people. Each contact is checked against their LinkedIn profile. Someone still there gets the email. Someone gone, with a replacement findable in the same role, is replaced as the recipient, and the opener never references a conversation the new person was not part of. Someone who moved to a company already in your CRM is flagged separately as a warm introduction there, which is frequently worth more than the win-back itself.
• The product. For every deal lost over a missing capability, Linear is searched for what has been released since the deal closed, matched on substance rather than wording. A buyer who wanted "a way to see who changed what" and a shipped issue titled "audit log" are the same thing.

The release date matters as much as the match. A capability that shipped last week is news. One that shipped fourteen months ago is an admission that nobody told them at the time, and the email is better for saying so plainly than for presenting a year-old release as though it just happened.

Picking the deals worth reopening

With the loss reason, the fresh signals, the contact situation and anything shipped side by side, a deal is worth reopening when one of them answers the loss reason:

• lost on a missing capability, and that capability now exists
• lost on timing or budget, and there is a funding round, a new fiscal year or a new person in the buying role
• lost to a competitor, and their contract is near its renewal window
• went quiet, and the person who went quiet has been replaced

Where nothing answers the loss reason, the deal is dropped and counted. The instinct is to write to everyone, because the list is right there and the drafting is free. These are accounts that already said no once, though, and a second contact with no new information is how a domain's reputation gets spent. Counting the drops rather than hiding them keeps that visible.

Each deal that stays gets one draft in Gmail: what happened, roughly when, what changed, and one question. If a capability shipped, the email says what it does. Nothing is sent. The batch goes to Slack with how many deals were in scope, how many were dropped and why, and the drafts grouped by loss reason, so each account owner can scan their own.