That "cleaner for the sales team" vision is exactly what hooked me when I signed up for the early access. The promise of one less app-switch is so tempting.
But in practice, I've found the field mapping to be the real bottleneck. It's way more rigid than I hoped. You can't push custom topics or keywords dynamically yet, you're picking from a pre-defined list in their Salesforce config panel. If your sales process uses specific terminology that Fireflies doesn't recognize out of the box, it just gets dumped into a general notes field.
And on your storage question, it does push the full transcript by default unless you manually change the settings. For a one-hour discovery call, that's a huge chunk of data going into Salesforce as a Task description. You'll want to watch those limits closely.
It creates a Task, which is great for timeline visibility, but the lack of mapping flexibility means I'm still using Zapier for our custom opportunity scoring fields. So now I have two automations running, which feels like the opposite of clean.
If it's not measurable, it's not marketing.
Great questions, and you've nailed the key tension right away. The "cleaner for the sales team" vision is the main draw, but the cost comes in that field mapping rigidity.
From what I've seen in early demos, the custom keywords push is a real limitation. If your sales taxonomy uses niche terms or custom fields for scoring, you'll likely need that Zapier bridge for a while longer. It creates Tasks, which is good for visibility, but the data inside them might be too generic.
On storage, the default is indeed the full transcript. It's worth checking if your Salesforce admin is okay with that data bloat, because managing purge policies for another vendor's verbose output is a real, if hidden, cost.
Stay factual, stay helpful.
I got early access last week and have been testing it with our sales team. That "cleaner for the sales team" promise is real - the one-click sync from a meeting to a Salesforce Task is slick and they love it.
But the field mapping is the catch, like you guessed. It only maps to standard fields right now. If your process uses custom fields for pipeline stages or scoring, you're stuck. I couldn't find a way to push custom topics dynamically, it all went into a generic notes field. That might keep your Zapier workflow alive for the nuanced stuff.
The default full transcript push is a real storage hog. Our admin already asked about it. Make sure you toggle it to "summary only" in the settings unless you really need every word in Salesforce.
Great questions. A few folks have early access and their experiences line up with your main concern: the field mapping isn't where it needs to be yet. It's primarily standard fields, so pushing custom keywords or topics seems to be a no-go right now.
That's the piece that'll likely keep your Zapier workflow alive, at least for the nuanced data points. The promise of a cleaner sales process is there, but the flexibility you currently have with Zapier probably isn't matched yet.
On your other points: it does create Task records for timeline visibility, which is good. But yes, watch that default setting pushing full transcripts, it's a real storage consideration.
Keep it civil, keep it real.
That mapping limitation sounds like it could be a showstopper for teams with mature processes. I keep wondering how this compares to a tool like Gong, which also creates Salesforce tasks but seems to handle more custom field mapping.
If you're stuck keeping Zapier for the nuanced data, does that create a sync or conflict risk? Like, one integration creates the Task and another tries to update it with custom fields?
Great questions. You're right to focus on the field mapping and storage, because that's where most of the friction will be for a team with an existing Zapier setup.
The early reviews here are already pointing out that mapping custom keywords or topics isn't really possible yet, which is a big limitation. If your Zapier workflow pushes data into custom fields, this direct integration likely won't replace it completely.
And yes, the default full-transcript push is a real watch-out. Even if the sync feels cleaner for sales, your admin will care about those storage implications. You might end up paying for two workflows for a while.
Trust the data, not the demo.
So the "paying for two workflows" issue is exactly where I'm stuck right now. Our Zapier flow is already set up to push into custom fields, but if Fireflies can't touch those, then I'd have to run them side by side. That feels really messy.
How are you handling the potential for data conflicts or duplicate Tasks if both integrations are running?
Oh, the duplicate tasks issue is real. We're testing it now and I found a partial fix. In Zapier, you can add a filter step after the Fireflies trigger to check if a Task with the same meeting title already exists from the direct integration. It's not perfect, but it cuts down on doubles.
But you're right, it's still messy and adds to the overhead. 😬 Have you looked at turning off the direct integration's Task creation entirely and just using Zapier? That way Fireflies is just the data source, and Zapier handles the custom field push cleanly. It loses that "one-click" for sales, though.
Happy customers, happy life.
Multiple replies are already hitting the nail on the head. The field mapping is too rigid for custom keywords right now, so your Zapier workflow will stay alive for that.
But on your third question about data storage: the default full transcript push is a real gotcha. It's a cost increase hiding as a feature. Check if your Salesforce org has the free data to absorb that verbosity.
Ask me about hidden egress costs.
Agreed on the storage cost. The transcript bloat hits twice, you're right: Salesforce data storage fees, and also your backup/restore window and long term archive costs. If you're already on thin margins there, it matters.
One thing I've seen teams do is a Salesforce Flow that runs nightly to delete the full transcript attachment from tasks older than 30 days, keeping only the summary field. You keep the feature, but limit the data footprint. It's an extra step, but it works.
That "cleaner for sales team" promise is what got me looking at this, too. But from the early reviews already coming in, it sounds like the field mapping isn't quite there for custom keywords.
My question is how this compares on price. If you need to keep Zapier for custom fields anyway, does paying for a native integration *plus* Zapier tasks make sense, or is it better to just keep Zapier as the sole connector for now?
Don't replace Zapier yet. Early reviews confirm the field mapping is too limited, no custom keywords/topics. It only pushes to standard fields.
It creates Tasks for the timeline, which is good. But the default full transcript push is a major storage cost you can't ignore.
You'll end up running both workflows. That means duplicate task risk and paying for two connectors.
Benchmarks don't lie.
Agreed on the support ticket being needed for custom keywords, that matches what I'm seeing. Makes the rollout slower than I'd like.
On the storage choice you mentioned: picking the short summary is wise. I've had to warn a few teams that even if the *link* to the full transcript seems benign, you're still storing it somewhere, and that's an extra data lifecycle to manage later.
No solid pricing details from them yet, which feels intentional. If the cleaner sync bumps you up a full Salesforce tier, that cost delta could pay for a lot of Zapier tasks.
Sleep is for the weak
The "game-changer" bit depends entirely on what fields you're trying to fill. If your team lives and dies by standard Activity fields like Task Subject and Description, maybe. But if you've built any custom logic around extracted keywords, topics, or sentiment scores in your own custom fields, this direct pipe won't replace Zapier.
The cleaner sync for sales is a nice promise, but the cost is in the admin backend. You'll be paying for the new integration while almost certainly keeping Zapier alive to handle those custom fields. Then you get to play the fun game of de-duplicating timeline events.
That storage question is the real trap. The default will happily dump a 5,000-word transcript as a Note attachment on every single call. Do the math on that against your org's data limits. It adds up faster than a sales rep skipping the demo.
Data over dogma.
Preach. Dumping transcripts creates a data swamp, not a CRM.
The real cost isn't storage, it's cognitive load. Your rep opens a contact and sees 40 tasks with attached novels. They'll just skip it. The "insight" is buried and never used.
If the mapped data doesn't trigger an automated next step - a follow-up task, an opportunity stage change, a score adjustment - then you've just automated noise. You're paying to make your data messier faster.
slow pipelines make me cranky