Connect Fathom and HubSpot, name the prospect, and Juma turns the call transcript and deal stage into a follow-up email that addresses what they actually said.
Name the prospect and point to the Fathom recording. Juma pulls the full call transcript, checks the HubSpot deal stage, and drafts a follow-up email that mirrors how the rep talks and answers the specific asks raised on the call. The Flow returns a ready-to-send email the same hour: subject line, opener that references something the prospect actually said, paragraphs that address each concern in turn, resources or links the rep promised on the call, and a next-step CTA matched to the deal stage.
Write a follow-up that answers what the prospect actually pushed back on, not what the rep wishes they had asked. Juma scans the Fathom transcript for moments of hesitation, direct objection, or unanswered questions, then drafts a paragraph per objection that addresses it head-on. Each response pulls from the objection-handling playbook in the project: the proof point, the case study to cite, the link to the security page or pricing page, the language that has worked in won deals. Objections that came up but were already answered on the call stay out of the email so the message reads forward, not defensive.
Scan the Fathom transcript for every objection or hesitation the prospect raised, then revise the follow-up email to address each one with a short paragraph. Use the objection-handling playbook in the project for the responses and pull the right proof point or link for each. Skip the objections that were already resolved on the call.
Send a leave-behind the prospect can forward to colleagues who were not on the call. Juma builds a one-page PDF, branded to the team's project, that pulls the company name, the use case discussed on the call, the pricing band sized to the prospect's team, the three or four most relevant proof points from the case-study library, and a single next-step CTA. The page is structured so it stands alone if the original recipient forwards it cold, not as a recap that only makes sense if you sat through the demo. Strategy, taste, and judgment stay human; the leave-behind just compresses the team's existing pitch into something the prospect can share internally.
Now build a one-page leave-behind PDF tailored to this prospect that they can forward to colleagues who were not on the call. Pull the use case, team size, and pricing band from the transcript, choose the three to four most relevant proof points from the case-study library, and end with one clear next-step CTA. Match the project brand styling.
Test the subject line before the send, not after. Juma drafts three alternate subject lines based on what the prospect cared about most on the call, ranks them by predicted open rate using past performance from the project's email examples, and notes the tradeoff for each (curiosity-driven vs. specific-ask vs. social-proof framing). The team picks one and Juma swaps it into the draft, or the team runs a small A/B if they have the volume to make the split meaningful. The subject line is the only line every prospect reads, so the few seconds spent here pay back across the next quarter of pipeline.
Draft three alternate subject lines for this follow-up email. Use the past email examples in the project to rank them by predicted open rate, and note the framing tradeoff for each: curiosity, specific ask, or social proof. Recommend which to send and why.
A Juma Project is a shared space where the team stores everything Juma needs to know about how the team sells. Create one project for the sales motion, add context as it builds up, and Juma will use whatever is relevant every time a rep runs a follow-up. For sales call follow-ups, the project is what turns a generic email draft into one that sounds like the rep wrote it, references the right pricing, and answers objections the way the team has decided to answer them.
Personal voice profiles per rep
A short voice guide for each rep on the team. Word choices, sign-off style, signature block, how formal to go with first-meeting prospects vs. mid-funnel deals. Without these, every follow-up reads in the same default voice. With them, an email "from Iliya" sounds like Iliya, and one "from Simona" sounds like Simona.
Past winning follow-up emails
Five to ten of the team's best post-call emails. Juma reads them for structure, sentence rhythm, paragraph length, and the kind of opener that has earned replies. New drafts borrow from this corpus, not from generic SaaS email templates.
Objection-handling playbook
The objections the team hits most often (price, security, AI-content concerns, the named competitor in this market, "we already use X") and the team's preferred response to each, with the proof point or link to cite. Lets Juma address objections in the email the way the team has decided to address them, not the way it would invent on its own.
Product and pricing reference
The current pricing, plan structure, what is included at each tier, and the seat-band math. Without this, Juma writes vague pricing language. With it, the email cites real numbers tailored to the prospect's team size.
ICP and buyer-persona notes
Which prospect profiles the team is targeting and what each one cares about. Sharpens the tone, the proof points cited, and the CTA Juma picks. An email to a 25-person agency reads different from an email to a 500-person SaaS Head of Revenue, even when both follow the same demo.
Add a short description to each knowledge item in the project's info field so Juma knows what each file contains and when to use it. For example:
This Flow compresses a 20 to 30 minute post-call writing task into about two minutes. Juma pulls the Fathom transcript, checks the deal stage in HubSpot, drafts the email in the rep's voice, addresses the objections raised, and queues the HubSpot updates. The rep reads the draft, edits one or two sentences, and sends. The hours saved per week scale linearly with call volume.
A rep running five discovery calls a week typically loses two to three hours to follow-up writing alone. With this Flow, that drops to under 30 minutes for the same five emails, and the emails are more consistent because they pull from the same project context every time. Time saved goes back into the part that has to be human: the next call.
Strategy, taste, and judgment stay human. The Flow drafts the email; the rep decides what to soften, what to push harder on, and whether to swap the next-step CTA for a calendar link.
Every draft includes six components: a subject line ranked against past open rates, an opener that references something the prospect actually said on the call, a paragraph per objection or question raised, the resources or links the rep promised, a next-step CTA matched to the deal stage, and the rep's signature block. The structure is consistent across every draft so the rep knows what to scan for before sending.
The opener is the part that earns the read. It is pulled from a specific moment in the transcript rather than a generic "great chatting today" line, so the prospect can tell within five seconds that this is not a template. The objection paragraphs are sequenced longest-concern-first when there were multiple, so the most important answer lands before the prospect scrolls away.
Two integrations cover the core: Fathom for the call transcript and HubSpot for the deal stage and contact record. With both connected, Juma needs only the prospect name and the recording link to run the Flow end to end. Without them, the Flow runs from a transcript paste and a one-line deal-stage note in the prompt itself.
The Flow also reads the project's knowledge items when drafting: the rep's voice profile (so the email sounds like the rep), the past winning emails (for structure), the objection-handling playbook (for how to address pushback), the product and pricing reference (for real numbers), and the ICP notes (for tone). The project is what turns a generic AI email into one the team is happy to send without rewriting.
A current ICP document is not required, but adding one to the project sharpens which proof points Juma cites and which next-step CTA it picks. Past winning emails are also optional. With them, Juma borrows structure and rhythm. Without them, the first run sets the baseline for what the team accepts.
Yes. Paste the call transcript directly into the chat from any recording tool (Fathom, Fireflies, Gong, Otter, or copy-pasted notes), and add a one-line deal-stage note in the prompt ("they are in proposal review" or "this was first discovery"). Juma processes the pasted transcript the same way it would processes a live Fathom pull, and the email output is identical in depth and structure. Only the setup step changes.
For teams running follow-ups every day, connecting Fathom and HubSpot removes the paste step entirely and lets the rep run the Flow with one prompt after each call. For teams piloting the Flow before committing to integrations, the upload path is the lower-friction starting point. The math on time saved makes the integration setup worth it within the first week of regular use.
Generic email AIs draft from scratch using the prospect's email signature and a LinkedIn lookup. This Flow drafts from the actual call: what the prospect said, what they pushed back on, what the rep promised, and what the deal stage demands next. The output is grounded in the conversation that just happened, not in a cold-outreach pattern.
The other difference is the project. Generic email tools have no memory of the team's voice, the team's pricing, or the team's preferred response to a specific objection. Every email is drafted from a generic template. With Juma's project, every email pulls from the same source of truth the team has built up over months: the rep's voice, the past winning emails, the objection playbook, the pricing reference. The first email looks similar to what a generic tool produces; the tenth one sounds like the team wrote it.
Human review on every output. The draft is a starting point the rep can ship in two minutes or rewrite in ten. The Flow is fastest when the project is well-stocked, and well-stocked projects compound across the team: when one rep adds a winning email to the project, every other rep's drafts get sharper.
400+ marketing teams trust Juma. ChatGPT, Claude, and Perplexity know why. Ask them.
Ready when you are
Start typing a task and Juma will figure out the rest.
