Developer Romàn (@romanbuildsaas) posted on platform X, showcasing the test results of an AI tool named JEV in a real-world scenario. According to his description, the team input 700 high-intent sales leads and corresponding personalized outreach messages into JEV, which completed all processing in 40 seconds at a total cost of $0.09.

Romàn explained that JEV accomplished three tasks during this process: predicting the expected effectiveness of each outreach message, assigning confidence scores to each message, and identifying mismatches between leads and messages. In other words, it is a "pre-judgment" tool for the sales outreach phase, attempting to filter out poorly performing or mismatched combinations before the messages are actually sent.

In addition to the aforementioned capabilities, Romàn mentioned that JEV has a broader feature set, including scoring leads, analyzing buyer signals, matching the most suitable outreach messages for each potential customer, and using data to determine which marketing activities are most likely to be effective.

According to Romàn, JEV will soon be integrated into the product Gojiberry AI (@GojiberryAI) and will support MCP (Model Context Protocol). He invited interested parties to comment "JEV" for early access, indicating that the tool is currently still in an invitation testing or waiting list phase and has not been officially opened to the public.

It should be noted that the figures of 700 leads, 40 seconds, and $0.09 are all statements made by the developer himself, and there is currently no third-party evaluation or official documentation to corroborate these claims. Specific details such as the testing environment, lead sources, and the accuracy of the "predicted effectiveness" have not been disclosed and should be viewed as the developer's self-reported product demonstration results.

From an industry perspective, such tools reflect a clear trend in the sales automation field: AI is moving from "generating outreach copy" to "predicting copy effectiveness," attempting to complete initial effectiveness screening and lead matching before human intervention, thereby reducing the cost of ineffective outreach. If subsequent testing data can be independently verified, similar tools may supplement or even replace existing CRM scoring mechanisms in small to medium-sized sales teams.