The AI SDR that reads every reply and plans the answer.

A sales AI agent that does the reply side of a sales development rep's work: it reads what a lead wrote back, works out what they meant and writes the next message toward your goal. In GetSales every workflow hands its replies to it by default once its Brief is written and it is live, and it works in Unibox, the unified inbox.

Mostly yes, under another job title: both mean an AI agent that works outbound conversations. In GetSales that agent is the AI SDR, and it takes the replies.

No. It takes the reply side of the work, and a person stays in charge of it. You write its Brief, choose how far it goes on its own and take every thread it hands you.

In GetSales, AI Search finds the people from a sentence, Workflows send the first touches on LinkedIn and email from your own sender profiles, and the AI SDR takes the replies. To start it, answer the five questions of its Brief: who it answers, what you sell, where a good conversation ends, what it never does and how it writes. The app can draft most of them from your website, and you keep what fits. Then press Go live.

Start with where it stops. Ask who sees a draft before it goes, which replies it leaves to a person, what it answers from and what a thread can cost. The AI SDR in GetSales answers LinkedIn and email replies in their thread, as that thread's sender profile, from your Brief, your Knowledge and the replies your team sent. Each scenario sets how far it goes on its own. The replies it leaves to a person come to you as a task, and every thread has a credits cap.

It reads the reply in its thread and labels it, Interested or Info request for example, with how sure it is and the reason in plain words. The label picks the scenario, and the scenario's plan runs step by step: react to the message, move the pipeline stage, and answer through the Playbook. The plan shows in the thread, and the scenario's level sets whether it runs on its own or waits there for your OK. The 15 labels and what each one does are on the reply classification page.

Through your link: it holds no calendar of yours. You set the goal in its Playbook, a call, a sign-up or a webinar spot, with its link, and it sends that link once they say yes, never before. If they name a time instead, it does not confirm it: you get a task. A third request to move the same meeting comes to you too.

Where a scenario lets it. Each scenario sets its level: Autopilot sends on its own, within the caps you set. Copilot drafts and waits for your OK. Conditional reads your conditions first. By default it sends at most 4 messages on its own in a thread, then it hands the thread to you. Whatever the level, it hands you a thread at its credits cap and any thread it should not answer alone, such as a call, a price question, a legal or security question, or a reply it cannot read: a task opens for you, the thread waits in Needs you and, by default, an email goes to whoever its sender profile is assigned to. Write in a thread yourself and it stops sending there and plans around your message.

From replies your team has sent. Its Golden set fills from your sender profiles' past threads, you mark the best ones as golden, and its voice forms from the set, the golden ones first. Your edits to its drafts join the set after your OK. Its Stop words are checked on every draft before it is sent or shown to you.

It spends credits from your workspace's balance, the same ones the pricing page sells as data credits. Its work on a reply spends credits, and sending spends none. Each thread has a cap, 40 credits by default, and at the cap the thread comes to you with a task. Its Settings show what a thread cost at the median over the last 30 days. Sign up free: Playtables stay free with your own API keys, and LinkedIn and email come with a 7-day trial.