Hey everyone,
We just launched Katalyst for sales teams on Salesforce.
Simple idea: reps don't hate selling, they hate the hours they lose every week feeding the CRM. Logging calls, fixing stages and close dates, writing next steps, reconstructing what happened on a deal from three weeks ago. The CRM they "quietly hate."
Most tools just made data entry slightly less painful. It's still the rep doing the work.
We're a smaller team - with just 8 reps on Salesforce. Is this built for us or are we going to feel like we're paying for enterprise features we don't use?
Katalyst
@peeyush_agarwal Yes, this would work well for your team. And there are different tiers to subscribe to where you get to decide how many total reps will use the platform.
Katalyst
@peeyush_agarwal Adding to yajwin's point, smaller teams often feel the impact faster, there's no ops person to keep Salesforce clean, so the admin falls on the reps directly. Taking that off 8 people who are all selling tends to show up quicker than at a 200-rep org.
Katalyst
@sahil217 The more complex, the better, honestly. We import your full configuration and the agent adapts dynamically. Custom objects, dependent picklists, formula fields, lookup relationships, record type rules, all of it. We've specifically built for messy enterprise Salesforce instances because that's where the real pain is. A vanilla instance is easy. Making it work on a 5-year-old instance with hundreds of custom fields is the challenge we've solved.
Katalyst
@sahil217 Adding to Avneet, the harder half of a complex instance is writing back into it, validation rules and required fields that reject a write. So we treat a write as done only when Salesforce confirms it, and surface the blocker if a rule rejects it, rather than silently failing. Reading the config is table stakes; writing cleanly is the real tes
We build account plans manually before every QBR and it takes our AEs hours. Can Katalyst generate those?
Katalyst
@armaaninder_singh Yes, our AI account plans are built exactly for this. Katalyst pulls in company research plus your own calls, emails, and signals to generate a full account plan automatically, turning hours of AE work into minutes. And because it's built on live data, the plan going into your QBR reflects where the deal actually stands right now, not a snapshot someone stitched together last week.
Katalyst
@armaaninder_singh adding to Avneet, the generated plan is a starting point you shape, not a locked output, so your AEs still bring their judgment, they're just editing something 90% done instead of building from a blank page at 11pm before the QBR. The hours saved are the grunt work, not the thinking.
Genuine question, couldn't I just paste my deal notes into ChatGPT and get basically the same output? What am I getting here that I'm not getting from a $20/month ChatGPT subscription?
Katalyst
@nilesh_phapale ChatGPT would only have context for that specific prompt you provided with the deal information. Whereas, the Katalyst agent works on signals, meeting information, emails, data across your Salesforce and more.
Katalyst
@nilesh_phapale The other half of it: ChatGPT hands you an answer you then re-key into Salesforce yourself. Katalyst does the gathering and the doing, across every deal, not just the one you remembered to paste in. A $20 tool helps you think about one deal; this runs the whole pipeline.
For Katalyst, when you say it works your Salesforce Pipeline, how much autonomy does the agent actually have inside Salesforce? For example, can it update fields, create tasks, and draft next steps on its own, or is it more of a recommendation layer that a rep approves before anything changes? That distinction would help people understand the workflow pretty quickly.
Katalyst
@mia_qiao Great question, and the distinction matters, the answer is it's both, and the rep controls which. Out of the box it runs as a recommendation layer: it drafts the field update, the task, the next step, and shows its reasoning, and the rep approves before anything changes. As it proves itself, you can hand specific fields or actions full autonomy so it writes them on its own, while keeping the higher-stakes stuff approval-only. So it's not "auto vs. suggest" as a fixed setting, it's a dial each rep moves as trust builds.
Katalyst
@mia_qiao One thing to add to Div's answer. The agent doesn't get special access to your Salesforce. It works through each rep's own Salesforce login, so it can only edit what that rep can edit. All your existing Salesforce permissions and rules still apply, and every change shows up under the rep's name, not a bot account. So even at full autonomy, it can never do more than the rep themselves could.
does the meeting recorder handle in-person conversations or just calls through your system, and how much do you lean on Salesforce data versus outside sources like LinkedIn for the deal patterns
Katalyst
@zilan42dm Right now the recorder handles your virtual meetings, Zoom, Meet, Teams, Webex, so calls through those are covered; in-person isn't something it captures today. On data, Salesforce is the backbone since it's your system of record, but we layer outside signals on top, funding, hiring, new contacts, partnerships, so the deal picture isn't limited to what's already been logged.
Katalyst
@zilan42dm To build on Div, keeping Salesforce as the backbone was a deliberate call, it's the one source that reflects how your team actually sells, so the deal patterns stay grounded in reality rather than generic web scraping. The outside signals sharpen that picture instead of replacing it. In-person capture is a real ask we've heard a few times now, so it's genuinely useful to see it come up again here.
This is awesome but I'm an early-stage startup and we're not on Salesforce. Any plans to support HubSpot or go CRM-agnostic?
Katalyst
@binit_agarwala Hey Binit! Hubspot is coming very soon and in the roadmap is our AI-native CRM offering as well.
Katalyst
@binit_agarwala sure binit! stay tuned even more exciting times ahead!!