Recall + Attention Case Study

“We were able to save about 9 months of engineering time using the Recall API because we didn't have to build our own meeting bots.”

What is Attention?

Attention is building the next generation of sales intelligence tools to help businesses achieve their full potential. Our fully-integrated AI-powered sales assistant provides real-time coaching, automatically fills out CRMs after sales calls, and drafts intelligent follow-up emails that pull from conversations. Revenue leaders have perfect visibility on their sales activity, providing them with accurate and updated enablement content and reports in minutes instead of days. We’ve helped hundreds of sales teams close deals faster and boost revenue.

Why does Attention need meeting data?

Attention is primarily for sales reps that do demos over platforms like Zoom, Google Meet, or Teams, but also who use dialers like Salesloft or Aircall. The follow up emails, CRM population, and real-time sales call guidance that Attention provides is based off of the source of truth, which is the actual customer conversation that happened. Because of this, we need to work with the audio streams from meetings. These long and unstructured conversations are a gold-mine of information for orgs, not only for the revenue teams, but also for product and marketing teams too.

How does Attention use Recall?

We use Recall to send a meeting bot into the call, so we can access the real-time audio streams from platforms like Zoom, Meet, Teams and more.

What did you do before Recall, and what was the problem?

Before Recall, we built a prototype with the Zoom RTMP streaming endpoint to get the real-time audio streams. However, there were a few problems with this.

  1. With the RTMP integration method, there were a lot of different tricky settings the Zoom admin had to activate. Users needing to go through the entire Zoom approval process was creating a lot of friction.
  2. This integration only worked with Zoom. So we would have to reinvent the wheel for every other video conferencing platform like Meet and Teams.
  3. The Zoom RTMP link would pop up this live streaming notification, which actually scared some of our users.
  4. It made speaker diarization extremely hard. We could not tell one speaker from another that easily. It was not very accurate.

So we definitely we had to move away from the live streaming side of things.

What led you to work with Recall?

When we evaluated you guys, it was just that perfect moment where we wanted a more seamless integration method getting the real time transcripts from Zoom and the other platforms. And because you guys provided that out of the box meeting bot API, we didn't need to build a meeting bot from scratch for each different platform.

Can you share some quantifiable metrics that stand out from your Recall integration so far?

I would say we were probably able to save about nine months of engineering time using the Recall API because we didn't have to build our own meeting bots. And actually even more than that, right? Because at the end of the day working with you guys also allows us to make sure that everything is stable. Allowing us to outsource all of the bot maintenance to you guys allows us to focus on all our efforts on the other stuff and we're able to build all these other features faster today.

Last thoughts?

I feel like you guys are part of our team today. You're very responsive. The product is stable and setting Recall up was extremely fast and seamless.