Connect HubSpot and Fathom, name a time period, and Juma turns closed-deal data and lost-call transcripts into a win/loss analysis with a sales objection playbook attached.
Name the client, the time period, and the deals to analyze. Juma pulls closed-won and closed-lost deals from HubSpot, reads call transcripts from Fathom for every lost deal, and researches each named competitor's positioning on the web. The Flow returns two artifacts the same day: a win/loss analysis report with patterns, objection themes, and competitive gaps, and a sales objection-handling playbook with the top objections the team keeps losing on plus how to handle each.
Read every closed-lost call as if the team sat in on it. Juma pulls transcripts from Fathom for every deal that closed lost in the selected time period, reads them in full, and clusters what the prospect actually said into a small set of recurring objection themes. Each theme is named in plain language ("we already use a free alternative," "the migration cost was too high," "the AE could not answer the security question"), counted across deals, and tied back to the specific deal IDs so the team can replay the calls. Themes that only appeared once stay separate from themes that recurred, so the team knows what to fix first.
Pull the Fathom transcripts for every deal that closed lost in the last six months and group what prospects said into recurring objection themes. For each theme, count how often it appeared, name it in plain language, and list the deal IDs where it showed up. Separate one-off objections from patterns the team is hitting repeatedly.
Turn the analysis into something the sales team uses on Monday. For each of the top three to five objection themes, the playbook drafts a one-page entry covering the objection in the prospect's own words, the underlying concern behind it, the response that has worked in won deals (pulled from the won-deal transcripts for cross-reference), the proof point or case study to cite, and the worst-case fallback when the objection holds. The output is a PDF the team can drop into Notion or hand to AEs in onboarding. Strategy, taste, and judgment stay human; the playbook just compresses what the team already knows into a usable format.
Now turn the top three to five objection themes into a sales objection-handling playbook. For each one, cover the objection in the prospect's words, the underlying concern, a response that has worked in won deals, the proof point or case study to cite, and the fallback when the objection holds. Output as a one-page-per-objection PDF the team can hand to AEs.
Find out which named competitor is actually winning the pipeline, not which one the team thinks is. Juma scans HubSpot closed-lost reasons and Fathom transcripts for every named competitor, counts mentions per deal, and ranks competitors by how often they appeared in losses. For the top two or three, the Flow then researches each one's current positioning on the web (homepage, pricing page, latest comparison pages they have published) and produces a side-by-side competitive gap table showing where they are winning, where they are at parity, and where the client has an unaddressed advantage the sales team is not leading with.
Scan HubSpot closed-lost reasons and Fathom transcripts for every competitor mention, then rank competitors by how often they appeared in lost deals. For the top two or three, research each one's current positioning on the web and build a competitive gap table covering where they are winning, where we are at parity, and where we have an advantage the team is not leading with.
Sort the win/loss picture by segment, not just by deal. Juma takes the ICP segments the team is targeting (company size bands, industry tiers, sales motion, regions) and calculates win rate, average deal size, and average sales cycle for each. The output is a segment scorecard that shows which segments the team is winning in, which segments are eating cycles without closing, and which segments the playbook still has not been pressure-tested against. The scorecard pairs with the Define Your ICP Flow, so what the win/loss data exposes can flow straight back into the next ICP refresh.
Compare win rates by ICP segment for the last six months. For each segment we target (size band, industry, sales motion, region), calculate win rate, average deal size, and average sales cycle. Build a segment scorecard that flags where we are winning, where we are losing time without closing, and which segments have too few deals to judge yet.
A Juma Project is a shared space where the team stores everything Juma needs to know about a client. Create one project per client, add context as it builds up, and Juma will use what is relevant every time the team runs a flow. For win/loss, the project pulls together the documents that turn raw CRM data into a sharp analysis: the most recent ICP, the sales positioning the team is being trained on, and any past win/loss reports to compare against.
Current ICP doc
The ICP segments the team is targeting. The win/loss analysis cross-references closed deals against each segment to flag which segments the team is winning in and which are eating cycles. Without this, the segment scorecard is built from inferred segments rather than the team's actual targeting.
Sales positioning and objection guidance
The pitch the team is being trained to give, the value props the AEs are leading with, and any objection-handling guidance the team already has. Lets Juma compare what the team said versus what the lost deals heard, and skip rewriting playbook entries that are already strong.
Competitor briefs
Anything the team has on the competitors they actively lose to: pricing, positioning, known strengths, the comparison pages those competitors have published about the client. Sharpens the competitive gap table beyond what web research alone can pull.
Past win/loss reports
The most recent win/loss analysis. When refreshing, Juma compares the new patterns against the previous report and calls out what shifted: new objections emerging, old objections that have been resolved, and competitive moves that altered the loss landscape.
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 two-week win/loss project into a single working session. Juma pulls deal data from HubSpot, reads every lost-deal transcript from Fathom, and synthesizes patterns in one pass. The team reviews the draft analysis the same day rather than waiting on a research sprint and manual transcript reading.
A manual win/loss project usually starts with pulling a deal export from the CRM, reading or skimming dozens of call recordings, taking notes by hand on objection themes, cross-checking competitor mentions, and then formatting all of it into a deck or doc for leadership. Most agencies and in-house teams run win/loss once a quarter at best because the cost of doing it well is so high. Many skip it entirely and run on instinct instead.
This Flow compresses the reading, clustering, and formatting work into one chat session, and pairs the analysis with a sales objection-handling playbook in the same run. The hours saved get redirected to the part that actually has to be human: deciding which patterns to act on, and how to change the pitch.
The full report covers six components: a deal-level breakdown of every win and loss, recurring objection themes with frequency counts, a named-competitor ranking based on how often each appeared in lost deals, a competitive gap table for the top two or three competitors, a win/loss scorecard by ICP segment, and a list of pattern shifts versus the previous quarter.
Each component is fully built out, not summarized. Objection themes are named in the prospect's own words and tied back to specific deal IDs so the team can replay the calls. The competitive gap table cites the source page on each competitor's website. The segment scorecard flags segments with too few deals to judge yet, so the team is not acting on noise.
Two integrations cover most of the analysis: HubSpot for closed-won and closed-lost deal data with reasons and custom fields, and Fathom for the call transcripts the analysis reads to cluster objection themes. With both connected, Juma needs only the time period and the named competitors to run the Flow end to end.
The Flow runs without integrations too. The team can upload a CSV export of closed deals from any CRM and a folder of call transcripts (Fathom export, Fireflies export, Gong export, or copy-pasted notes), and Juma will treat the export the same way it would treat live data. The output is identical in depth and structure. Integrations only change the setup time, not the substance of the analysis.
A current ICP document is not required, but adding one to the project sharpens the segment scorecard and links the win/loss output back into the next ICP refresh. Past win/loss reports are also optional. With them, Juma flags what shifted; without them, the first run establishes the baseline.
Some closed-lost deals end with no clear reason and no call transcript. Juma surfaces these as a separate bucket called "unattributed losses," counts them, and flags them as a process gap rather than guessing at the cause. A high unattributed-loss percentage is the signal, not noise to filter out.
When unattributed losses sit above a threshold (typically 30 percent of closed-lost deals), the analysis treats the gap itself as the first finding. Before the team can fix objection themes or competitive positioning, the deal-loss capture process needs to surface what is actually happening. The Flow's recommendations include practical fixes: a closed-lost reason field that is required at deal close, a 10-minute loss debrief logged in HubSpot, or a quarterly call with the AE to back-fill recent losses while they are still fresh.
Strategy, taste, and judgment stay human. The Flow flags the gap and proposes the fix; the team owns the call on what process change actually gets shipped.
Yes. Upload a CSV of closed deals from any CRM (Salesforce, Pipedrive, Attio, or a spreadsheet) and a folder of call transcripts from any source. The Flow will process the upload the same way it processes the live integrations. The deliverables, structure, and depth do not change. Only the setup step does.
For teams running win/loss as a one-off project for a specific quarter, the upload path is often faster than connecting integrations the team would not otherwise need. For teams running win/loss every quarter, connecting HubSpot and Fathom removes the export step entirely and lets the Flow rerun with one prompt each cycle.
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.
