Converting inbound leads into customers at OpenAI
2:32

Converting inbound leads into customers at OpenAI

OpenAI 29.09.2025 43 053 просмотров 763 лайков

Machine-readable: Markdown · JSON API · Site index

Поделиться Telegram VK Бот
Транскрипт Скачать .md
Анализ с AI
Описание видео
Turning inbound leads into customers takes quick, personal follow-up. Tailor Assist gives every prospect a thoughtful response at scale, using AI to craft answers that feel authentic and helpful.

Оглавление (1 сегментов)

Segment 1 (00:00 - 02:00)

It's one of our highest converting channels by an order of like magnitude. We have like 13,000 uh leads that come inbound every single month. And of those, we are really only like able to have a hyperpersonalized touch with about a thousand of them. A lot of people this is a very big purchase for them. and we want to make it as comfortable as possible. And by just sending them an automated email, we weren't actually accomplishing that. Taylor's able to take all of your internal knowledge, uh, your policies, your writings, all of your docs, um, your public knowledge, your case studies, your website, and turn that into like knowledge that an agent can read and then act on. When a lead submits a form, Taylor's basically taking all that information that they've given us and able to figure out what's the best product for that customer and write an email and offer to assist them in any of their product support or like purchasing journey. When we were deciding how to make tailor, we were talking to uh mostly the sales team um and trying to understand here's like the ideal customer journey. We found that the best way to improve our agents behavior and systemically improve his performance was to take a small subset of examples and with a subject matter expert write very detailed criteria on what makes a good email or interaction. Basically tell the model if this was like a 10 out of 10 email like how would you get there? Then you're able to basically like keep running it, keep fixing it. you're able to see like where your gaps are and consistently perform. So, we're able to build like a very complex um eval system like just me and one other engineer um to do multi-turn evals. We're measuring like you know how often are customers responding, how often are they purchasing? And once you start doing those repeatedly, you're able to ultimately increase your performance. If we just give someone like a pretty personalized experience and answer some of their important questions, they're willing to buy really quickly. Since we've launched this a couple months ago, it's unlocked multi-millions of dollars in incremental revenue. Now, we're able to basically interact very personally with the whole world. Nice.

Другие видео автора — OpenAI

Ctrl+V

Экстракт Знаний в Telegram

Экстракты и дистилляты из лучших YouTube-каналов — сразу после публикации.

Подписаться

Дайджест Экстрактов

Лучшие методички за неделю — каждый понедельник