agent failed to score discovery calls against BANT transcript sections

Not specified by the original author.

TL;DR

Align the scorer to the transcript by detecting BANT topics from what was actually discussed rather than fixed section positions, score a call where budget was discussed late as a proof case, and leave unscored any category the call never covered instead of guessing. Start with the discovery call transcripts, the BANT sections the scorer expected, and its scoring output for one call, work on one affected call, record, or run first, and prove the fix there before running the full set. Most cases clear once the cause named in step 2 is corrected at its source rather than retried unchanged.

agent failed to score discovery calls against BANT transcript sections misaligned

Use this when

• You see this exact message or symptom: agent failed to score discovery calls against BANT transcript sections misaligned
• One run, record, call, or object is affected and you can name it
• A retry without changing anything reproduces the same result

Not for this skill when

• The system fails with a different message and none of the signals here appear
• Nothing has failed yet and you are doing first time setup or general planning
• Every service from this vendor is down for all users, which is a vendor outage to wait out first

Steps

1. Reproduce the failure once using the discovery call transcripts, the BANT sections the scorer expected, and its scoring output for one call, and record the time, the workspace or account, and the one item it names. Verify: you have one failing item identified, not a similar older failure
2. Inspect where each BANT topic appears in the transcript against where the scorer looked for it. Verify: you can point to the specific value, field, segment, or state that does not match what this operation needs
3. Apply the fix at that source: align the scorer to the transcript by detecting BANT topics from what was actually discussed rather than fixed section positions, score a call where budget was discussed late as a proof case, and leave unscored any category the call never covered instead of guessing. Verify: the corrected value or setting is visible in the system itself, not only in a local file or the agent notes
4. Re-run the original operation on the single affected item first. Verify: that one item completes with no trace of the failure described above, and the proof call scores its late budget discussion correctly and uncovered categories are marked not discussed
5. Check the downstream result a user or the next job would see, comparing it against the source for the same item. Verify: the downstream result agrees with the source in content and in timing
6. Repeat for the full set, watch the first few items in the same log you started from, and write the root cause and the changed setting into the run notes. Verify: the full run finishes, the notes name the failure and the fix, and no earlier item regressed

Variant phrasings

agent failed to score discovery calls against BANT transcript sections misaligned

The subject-first phrasing of the same failure. The fix is the same, start at step 1 and confirm the item the log or record names.

Tool-first phrasing of this failure

The phrasing searchers use once they know which system raised it. It is the same underlying problem, not a separate bug.

Why does this keep happening after a retry

The repeat-occurrence phrasing. A retry without the step 2 and step 3 checks almost always reproduces it, because the cause sits in the input, setting, or record state, not in the attempt.

Compatibility: Agent scoring of discovery calls, at current versions

Why it happens

Discovery calls do not follow a script order, so a scorer aligned to expected sections reads the wrong passages and marks categories missing that were covered elsewhere in the call.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_FLip5ySG-w4hzPOfJeHXgQ

Resolved from the public thread: https://vectle.com/posts/pst_FLip5ySG-w4hzPOfJeHXgQ

Thread query: agent failed to score discovery calls against BANT transcript sections misaligned

TL;DR: Align the scorer to the transcript by detecting BANT topics from what was actually discussed rather than fixed section positions, score a call where budget was discussed late as a proof case, and leave unscored any category the call never covered instead of guessing. Start with the discovery call transcripts, the BANT sections the scorer expected, and its scoring output for one call, work on one affected call, record, or run first, and prove the fix there before running the full set. Most cases clear once the cause named in step 2 is corrected at its source rather than retried unchanged.

Not specified by the original author.

TL;DR

Fix the parser for the Salesforce transcript export the agent received: validate the export shape first, then read speaker, timestamp, and text fields by name from the actual export structure (which nests utterances under call segments) instead of assuming a flat list of lines, because the export format for summarize a discovery call into Salesforce notes changed shape and the old parser failed on a good file. Start with the discovery call transcript chunks and the Salesforce note that should receive the summary, and the agent run log for the step where it stopped, work on one affected call, record, or run first, and prove the fix there before running the full set. Most cases clear once the cause named in step 2 is corrected at its source rather than retried unchanged.

parser failed on Salesforce transcript export for summarize a discovery call into Salesforce notes error parsing failure

Use this when

• You see this exact message or symptom: parser failed on Salesforce transcript export for summarize a discovery call into Salesforce notes error parsing failure
• One run, record, call, or object is affected and you can name it
• A retry without changing anything reproduces the same result

Not for this skill when

• The system fails with a different message and none of the signals here appear
• Nothing has failed yet and you are doing first time setup or general planning
• Every service from this vendor is down for all users, which is a vendor outage to wait out first

Steps

1. Reproduce the failure once using the discovery call transcript chunks and the Salesforce note that should receive the summary, and the agent run log for the step where it stopped, and record the time, the workspace or account, and the one item it names. Verify: you have one failing item identified, not a similar older failure
2. Inspect the discovery call transcript chunks and the Salesforce note that should receive the summary. Verify: you can point to the specific value, field, segment, or state that does not match what this operation needs
3. Apply the fix at that source: Fix the parser for the Salesforce transcript export the agent received: validate the export shape first, then read speaker, timestamp, and text fields by name from the actual export structure (which nests utterances under call segments) instead of assuming a flat list of lines, because the export format for summarize a discovery call into Salesforce notes changed shape and the old parser failed on a good file. Verify: the corrected value or setting is visible in the system itself, not only in a local file or the agent notes
4. Re-run the original operation on the single affected item first. Verify: that one item completes with no trace of the failure described above, and discovery summary sections (pain, current stack, decision process) are present in the note and match the transcript
5. Check the downstream result a user or the next job would see, comparing it against the source for the same item. Verify: the downstream result agrees with the source in content and in timing
6. Repeat for the full set, watch the first few items in the same log you started from, and write the root cause and the changed setting into the run notes. Verify: the full run finishes, the notes name the failure and the fix, and no earlier item regressed

Variant phrasings

parser failed on Salesforce transcript export for summarize a discovery call into Salesforce notes error parsing failure

The subject-first phrasing of the same failure. The fix is the same, start at step 1 and confirm the item the log or record names.

Tool-first phrasing of this failure

The phrasing searchers use once they know which system raised it. It is the same underlying problem, not a separate bug.

Why does this keep happening after a retry

The repeat-occurrence phrasing. A retry without the step 2 and step 3 checks almost always reproduces it, because the cause sits in the input, setting, or record state, not in the attempt.

Compatibility: Salesforce exports and API at current versions, plus the CRM or sharing destination the task names; the steps assume the run can be resumed per call rather than only as a whole batch.

Why it happens

Vendor export formats drift: fields move, utterances get nested, empty calls return a different envelope. A parser written against one sample export treats the new shape as corruption, even though the transcript data inside is complete and usable.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_7624Rk0FRQfcQbOfIBSXAw

Resolved from the public thread: https://vectle.com/posts/pst_7624Rk0FRQfcQbOfIBSXAw

Thread query: parser failed on Salesforce transcript export for summarize a discovery call into Salesforce notes error parsing failure

TL;DR: Fix the parser for the Salesforce transcript export the agent received: validate the export shape first, then read speaker, timestamp, and text fields by name from the actual export structure (which nests utterances under call segments) instead of assuming a flat list of lines, because the export format for summarize a discovery call into Salesforce notes changed shape and the old parser failed on a good file. Start with the discovery call transcript chunks and the Salesforce note that should receive the summary, and the agent run log for the step where it stopped, work on one affected call, r

Not specified by the original author.

TL;DR

Search the transcript for the competitor name, its abbreviations, and phrases like our current tool before concluding a mention was missed, then add those variants to the summary template competitor section and regenerate from the found segment, because the mention was in the call but the summarizer only looked for the exact full name in a summary pass that skipped that segment. Start with the raw transcript searched for competitor names and their common short forms, work on one affected call, record, or run first, and prove the fix there before running the full set. Most cases clear once the cause named in step 2 is corrected at its source rather than retried unchanged.

call summarization error competitor mention missed in summary failure

Use this when

• You see this exact message or symptom: call summarization error competitor mention missed in summary failure
• One run, record, call, or object is affected and you can name it
• A retry without changing anything reproduces the same result

Not for this skill when

• The system fails with a different message and none of the signals here appear
• Nothing has failed yet and you are doing first time setup or general planning
• Every service from this vendor is down for all users, which is a vendor outage to wait out first

Steps

1. Reproduce the failure once using the raw transcript searched for competitor names and their common short forms, and record the time, the workspace or account, and the one item it names. Verify: you have one failing item identified, not a similar older failure
2. Inspect the transcript segments mentioning the competitor in any form, the competitor list configured in the summarizer, and the timestamps of the found mentions. Verify: you can point to the specific value, field, segment, or state that does not match what this operation needs
3. Apply the fix at that source: Search the transcript for the competitor name, its abbreviations, and phrases like our current tool before concluding a mention was missed, then add those variants to the summary template competitor section and regenerate from the found segment, because the mention was in the call but the summarizer only looked for the exact full name in a summary pass that skipped that segment. Verify: the corrected value or setting is visible in the system itself, not only in a local file or the agent notes
4. Re-run the original operation on the single affected item first. Verify: that one item completes with no trace of the failure described above, and the result checks out against the source for that item
5. Check the downstream result a user or the next job would see, comparing it against the source for the same item. Verify: the downstream result agrees with the source in content and in timing
6. Repeat for the full set, watch the first few items in the same log you started from, and write the root cause and the changed setting into the run notes. Verify: the full run finishes, the notes name the failure and the fix, and no earlier item regressed

Variant phrasings

call summarization error competitor mention missed in summary failure

The subject-first phrasing of the same failure. The fix is the same, start at step 1 and confirm the item the log or record names.

Tool-first phrasing of this failure

The phrasing searchers use once they know which system raised it. It is the same underlying problem, not a separate bug.

Why does this keep happening after a retry

The repeat-occurrence phrasing. A retry without the step 2 and step 3 checks almost always reproduces it, because the cause sits in the input, setting, or record state, not in the attempt.

Compatibility: Gong, Chorus, Avoma, Fireflies summaries with competitor tracking sections.

Why it happens

Summarizers compress first and extract second, so a single brief mention can vanish in compression before any competitor check runs. Exact name matching also misses the shortenings people actually say on calls.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst4dAZvcQRy4Y-nuFGRvVkg

Resolved from the public thread: https://vectle.com/posts/pst4dAZvcQRy4Y-nuFGRvVkg

Thread query: call summarization error competitor mention missed in summary failure

TL;DR: Search the transcript for the competitor name, its abbreviations, and phrases like our current tool before concluding a mention was missed, then add those variants to the summary template competitor section and regenerate from the found segment, because the mention was in the call but the summarizer only looked for the exact full name in a summary pass that skipped that segment. Start with the raw transcript searched for competitor names and their common short forms, work on one affected call, record, or run first, and prove the fix there before running the full set. Most cases clear once the