I've been testing the early access for the last week. To answer your questions directly:
On field mapping, it's not seamless for custom keywords yet. You can map to standard fields like Task Subject, Description, and a few others, but pushing custom Fireflies topics or keywords into custom Salesforce fields requires a support ticket. Your Zapier workflow will still be needed for that level of detail.
It does create Tasks for the timeline, which is the right move for visibility. The activity feed notes alone would get lost.
The storage impact is the biggest watch-out. The default setting pushes the entire transcript as a note attachment. For a high-call-volume team, this will consume your data limits quickly. I'd recommend switching the sync to the short summary option immediately.
Given those constraints, it won't fully replace Zapier if you rely on custom data. You'll likely end up with two parallel syncs, which introduces a duplicate task risk that needs managing.
—Anita
Thanks for sharing your real-world test results, they line up with the concerns I've been hearing. Your point about duplicate task risk from running two parallel syncs is a major one that's been under-discussed.
Managing that duplication likely means building a cleanup flow, which adds more admin overhead and potential for things to break. It starts to feel like you're engineering a solution around the new "seamless" integration, instead of the integration simplifying your workflow.
I'm also curious if you've seen any patterns in how the Task Subject field is populated. Does it use the meeting title, or does it try to pull something from the summary? That often determines whether a rep can scan their timeline effectively.
The right tool saves a thousand meetings.
Your question on the Task Subject is a good one, and it gets to a practical usability issue. In my testing, it populates the Subject field with a truncated version of the meeting title from the calendar event. If the title is long or generic, like "Intro Call," that's not very scannable. It doesn't pull from the AI summary.
You're spot-on about the duplication risk requiring its own cleanup flow. That's the hidden admin tax. You end up building logic to deconflict based on timestamps or external IDs, which adds complexity and a new point of failure. So much for a simpler workflow.
You've nailed the early config experience. That summary blurb is indeed static, and it feels like a missed opportunity. You can't even pull in something as simple as action items from the transcript.
I agree the custom field mapping being limited to support tickets is the real gatekeeper for many teams. It means you can't just enable this and hand it to a sales pod, you have to get IT or an admin involved first. That kills the "quick win" vibe of the launch.
My team relies on a custom 'Next Step' field for pipeline health, and it looks like that'll stay on Zapier for the foreseeable future.
null