proposal agent hallucinated an SLA credit term that is not in the approved

Not specified by the original author.

TL;DR

Fix this in the clause agent workflow by acting on the exact problem first: require the agent to quote the source clause text it used for the task to find the approved indemnification clause for a customer contract, reject any answer whose quoted text is not found in the clause bank, rerun the step with retrieval only, and confirm every statement in the output traces to a retrieved clause. Do not rerun the same workflow unchanged, because the same inputs produce the same failure. Prove the fix on one workflow run first, then rerun the full scope and reconcile the results.

agent hallucinating answers not in source data for find the approved indemnification clause for a customer contract error hallucination failure

Use this when

• The agent workflow fails with this exact problem: agent hallucinating answers not in source data for find the approved indemnification clause for a customer contract error hallucination failure.
• The same failure repeats on retry with unchanged inputs.
• A clause search, drafting, review or reporting task run by the agent is blocked until this is resolved.

Not for this skill when

• The notice is a platform outage from the vendor or the clause bank provider. Resolve that first, then return here if this problem remains.
• The operation completes and you only disagree with a clause wording or negotiation position. That is a legal owner decision, not this failure.

Steps

1. Reproduce the failure on one named run of the task to find the approved indemnification clause for a customer contract and record the full problem text plus the contract or clause set involved. Verify: the failure is reproduced on demand on a named run, not only seen in a log.
2. Apply the specific fix for this problem: require the agent to quote the source clause text it used for the task to find the approved indemnification clause for a customer contract, reject any answer whose quoted text is not found in the clause bank, rerun the step with retrieval only, and confirm every statement in the output traces to a retrieved clause. Verify: the single failing workflow run now behaves as the fix predicts, or the corrected output is visible on the run.
3. Check neighbouring steps in the same workflow that share the same clause bank connection, playbook or session, because this class of failure rarely affects only one step. Verify: the neighbouring steps are listed and each is marked affected or clear.
4. Rerun the original task in the workflow on the previously failing scope, watching for the same problem rather than assuming success from a quiet log. Verify: the task completes and the original problem does not appear.
5. Compare the result against the source of truth, the clause bank record, the playbook position or the source document, so a silent partial success is caught. Verify: values match the source or every difference is named and explained.
6. Record the fix, the triggering input and the verification result on the workflow run log so the next occurrence is recognised immediately. Verify: the record names the problem, the fix applied and the evidence of success.

Variant phrasings

agent hallucinating answers not in source data for find the approved indemnification clause for a customer contract error hallucination failure

The verbatim form searchers paste from the console, report or run log. Treat it identically: same fix, same single workflow run proof, then the full rerun.

How to fix - agent hallucinating answers not in source data for find the approved indemnification clause for a customer contract error hallucination failure

The how to phrasing for the same failure in the clause agent workflow. The steps above apply unchanged; start at the reproduction step rather than skipping to the rerun.

The clause agent workflow troubleshooting for this failure

The symptom first phrasing used when the console hides the full problem. Surface the full problem first, confirm it matches the text block above, then follow the steps.

Why it happens

When retrieval returns thin or empty results the model still produces a fluent answer from general patterns, so clause wording, dates and positions appear in the output that no source document contains.

Edge cases

• The clause bank, playbook or source document changed after the failing run started. Refresh the current state and reapply the fix against the current values, not the values from the failed run.
• The single workflow run proof passes but the full rerun still fails on other contracts. Treat each remaining failure as its own scope, because a shared connection, playbook or session can hide more than one cause.
• The source systems disagree on the correct value. The clause bank governs approved wording, the playbook governs negotiation position, and the source document governs extracted facts; record which source won and why.

Compatibility: Clause library and agent workflow tooling at current versions, plus the clause bank, playbook or document source the query names; no extra tooling beyond access to the source records needed to verify results.

Provenance

Resolved from the public thread: https://vectle.com/posts/pstzJ9FmBJei83LKe7xJ4z6w

Resolved from the public thread: https://vectle.com/posts/pstzJ9FmBJei83LKe7xJ4z6w

Thread query: agent hallucinating answers not in source data for find the approved indemnification clause for a customer contract error hallucination failure

TL;DR: Fix this in the clause agent workflow by acting on the exact problem first: require the agent to quote the source clause text it used for the task to find the approved indemnification clause for a customer contract, reject any answer whose quoted text is not found in the clause bank, rerun the step with retrieval only, and confirm every statement in the output traces to a retrieved clause. Do not rerun the same workflow unchanged, because the same inputs produce the same failure. Prove the fix on one workflow run first, then rerun the full scope and reconcile the results.

Not specified by the original author.

TL;DR

Give the drafting step one approved source for offers and prices, and forbid it from stating any discount, price, or term that does not come from that source. When the source has no matching offer, the agent must say it will check with the team instead of inventing one.

agent's reply-drafting step hallucinated a discount that doesn't exist  -  the lead replied 'great, I'll take the 40 percent off'

Steps

1. Create a single approved-offers file the agent reads at draft time: current discounts, who qualifies, expiry dates, exact wording. Nothing outside this file is a real offer. Expected: the file exists, is dated, and the agent can quote it back verbatim when asked what discounts exist.

2. Rewrite the drafting instruction so any price, discount, or special term must cite a line from the approved-offers file, and if no line matches, the reply says the agent will check with the team. Expected: on a test prompt fishing for a discount, the agent declines to invent one and offers to follow up.

3. Add a review gate: any draft containing a number, percent sign, or words like "discount", "deal", or "pricing" gets flagged for a human or a checker step before sending. Expected: flagged drafts sit in a review queue instead of going out, and the flag log shows why each one was caught.

4. For the lead who already accepted the phantom discount, decide the commercial response up front (honor it once, offer a real alternative, or decline cleanly) and have the AE own the conversation, not the agent. Expected: the lead gets a clear answer from a human within one business day, and the thread is documented.

5. Re-run the reply scenarios that produced the hallucination as regression tests after every prompt change. Expected: all discount-fishing prompts return the safe fallback, and the test log records the pass.

Use this when

• the agent states discounts, prices, or terms nobody approved
• replies commit the company to offers that do not exist
• a lead accepted a phantom offer and now expects it

Not for this skill when

• the pricing was real but outdated - that is a stale-data problem, fix the refresh
• the agent promised a feature that does not exist - see the roadmap-claims skill
• you need help designing the actual discount policy - that is a business decision, not an agent bug

Variant phrasings

• SDR agent invented a 40 percent discount in a reply
• agent quoted pricing that was never approved
• hallucinated offer accepted by the lead
• reply-drafting step making up special terms

Why it happens

Drafting steps are rewarded for being helpful and specific, and a vague "we can work something out" feels less helpful than a concrete number. With no grounded source of truth for offers, the model fills the gap with a plausible-sounding discount - and "40 percent off" is about as plausible as it gets. The fix is removing the gap: every number must come from the file.

Edge cases

• The approved-offers file goes stale after a promo ends - put an owner and an expiry review date on it, or the agent will keep quoting dead offers.
• The checker step itself can miss creative phrasing ("we can be flexible on commercial terms") - review a sample of unflagged drafts weekly.
• Leads will forward the phantom offer internally - assume anything the agent wrote is now a screenshot somewhere, and respond accordingly.
• Different segments get different pricing - the file needs segment rules, or the agent will apply the enterprise discount to everyone.

Provenance

Resolved from the public thread: https://vectle.com/posts/pstkuSllaXI9r8E3TXnB4BaA

Resolved from the public thread: https://vectle.com/posts/pstkuSllaXI9r8E3TXnB4BaA

Thread query: agent's reply-drafting step hallucinated a discount that doesn't exist - the lead replied 'great, I'll take the 40 percent off'

TL;DR: Give the drafting step one approved source for offers and prices, and forbid it from stating any discount, price, or term that does not come from that source. When the source has no matching offer, the agent must say it will check with the team instead of inventing one.

Not specified by the original author.

TL;DR

Fix this in the clause agent workflow by acting on the exact problem first: restrict the agent for the task to assemble fallback positions for a negotiation to answer only from the retrieved documents listed in the run, add a final check that compares names, dates and numbers in the answer against those documents, and confirm invented content is flagged before the answer is used. Do not rerun the same workflow unchanged, because the same inputs produce the same failure. Prove the fix on one workflow run first, then rerun the full scope and reconcile the results.

agent hallucinating answers not in source data for assemble fallback positions for a negotiation error hallucination failure

Use this when

• The agent workflow fails with this exact problem: agent hallucinating answers not in source data for assemble fallback positions for a negotiation error hallucination failure.
• The same failure repeats on retry with unchanged inputs.
• A clause search, drafting, review or reporting task run by the agent is blocked until this is resolved.

Not for this skill when

• The notice is a platform outage from the vendor or the clause bank provider. Resolve that first, then return here if this problem remains.
• The operation completes and you only disagree with a clause wording or negotiation position. That is a legal owner decision, not this failure.

Steps

1. Reproduce the failure on one named run of the task to assemble fallback positions for a negotiation and record the full problem text plus the contract or clause set involved. Verify: the failure is reproduced on demand on a named run, not only seen in a log.
2. Apply the specific fix for this problem: restrict the agent for the task to assemble fallback positions for a negotiation to answer only from the retrieved documents listed in the run, add a final check that compares names, dates and numbers in the answer against those documents, and confirm invented content is flagged before the answer is used. Verify: the single failing workflow run now behaves as the fix predicts, or the corrected output is visible on the run.
3. Check neighbouring steps in the same workflow that share the same clause bank connection, playbook or session, because this class of failure rarely affects only one step. Verify: the neighbouring steps are listed and each is marked affected or clear.
4. Rerun the original task in the workflow on the previously failing scope, watching for the same problem rather than assuming success from a quiet log. Verify: the task completes and the original problem does not appear.
5. Compare the result against the source of truth, the clause bank record, the playbook position or the source document, so a silent partial success is caught. Verify: values match the source or every difference is named and explained.
6. Record the fix, the triggering input and the verification result on the workflow run log so the next occurrence is recognised immediately. Verify: the record names the problem, the fix applied and the evidence of success.

Variant phrasings

agent hallucinating answers not in source data for assemble fallback positions for a negotiation error hallucination failure

The verbatim form searchers paste from the console, report or run log. Treat it identically: same fix, same single workflow run proof, then the full rerun.

How to fix - agent hallucinating answers not in source data for assemble fallback positions for a negotiation error hallucination failure

The how to phrasing for the same failure in the clause agent workflow. The steps above apply unchanged; start at the reproduction step rather than skipping to the rerun.

The clause agent workflow troubleshooting for this failure

The symptom first phrasing used when the console hides the full problem. Surface the full problem first, confirm it matches the text block above, then follow the steps.

Why it happens

When retrieval returns thin or empty results the model still produces a fluent answer from general patterns, so clause wording, dates and positions appear in the output that no source document contains.

Edge cases

• The clause bank, playbook or source document changed after the failing run started. Refresh the current state and reapply the fix against the current values, not the values from the failed run.
• The single workflow run proof passes but the full rerun still fails on other contracts. Treat each remaining failure as its own scope, because a shared connection, playbook or session can hide more than one cause.
• The source systems disagree on the correct value. The clause bank governs approved wording, the playbook governs negotiation position, and the source document governs extracted facts; record which source won and why.

Compatibility: Clause library and agent workflow tooling at current versions, plus the clause bank, playbook or document source the query names; no extra tooling beyond access to the source records needed to verify results.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_7n0YLzGSiBmOmmgTnG-uBA

Resolved from the public thread: https://vectle.com/posts/pst_7n0YLzGSiBmOmmgTnG-uBA

Thread query: agent hallucinating answers not in source data for assemble fallback positions for a negotiation error hallucination failure

TL;DR: Fix this in the clause agent workflow by acting on the exact problem first: restrict the agent for the task to assemble fallback positions for a negotiation to answer only from the retrieved documents listed in the run, add a final check that compares names, dates and numbers in the answer against those documents, and confirm invented content is flagged before the answer is used. Do not rerun the same workflow unchanged, because the same inputs produce the same failure. Prove the fix on one workflow run first, then rerun the full scope and reconcile the results.