$ loading_
Prepare reviewable retention and service-recovery options for churn-related customer cases.
Copy the install command and let the AI configure it · recommended for beginners
Please install the "retention-actions" skill from askskill: 1. Download https://raw.githubusercontent.com/microsoft/aibast-agents-library/main/solutions/customer-sentiment-churn/manual/skills/aibast_retention-actions_03/SKILL.md 2. Save it as ~/.claude/skills/aibast_retention-actions_03/SKILL.md 3. Reload skills and tell me it's ready
Prepare reviewable retention and service-recovery options for Marcus Johnson. Use only the provided synthetic operating snapshot, clearly separate observed evidence, inferred results, and proposed next steps, and state that no customer contact or fee change has occurred.
A reviewable draft with evidence, analysis, proposed next steps, and required disclaimers.
Generate retention options for the specified fictional customer record only; do not substitute another record. Return source-backed evidence for the request and identify what authorized human review is required before action.
A recommendation list tied to the exact record, with source-backed evidence and required human review.
Prepare reviewable pre-outreach options without contacting the customer or making any offer or account change. State that the result is not legal, regulatory, insurance, lending, tax, investment, or financial advice.
An internal review draft of pre-outreach options with clear limitation statements.
Marketing, relationship, or product teams can prepare reviewable retention and service-recovery options for an at-risk customer case. It is meant for pre-review preparation and does not contact customers or execute changes.
When answering retention-preparation questions, the skill can extract source-backed evidence from a synthetic operating snapshot. The output separates observed information, calculated or heuristic results, and proposed next steps.
Before any outreach, fee adjustment, or external action, teams can generate a draft for authorized human review. The result states that no contact, approval, transaction, or account change has occurred.
The README describes a retention-option preparation skill for a synthetic customer sentiment and churn prediction pilot. It prepares reviewable service-recovery and outreach options before any customer contact or offer. The procedure requires using the exact fictional record scope, grounding results in a synthetic operating snapshot, separating observed evidence from inferred output and next steps, adding explicit non-advice and no-action disclaimers, and naming the authorized human review required before any action.
Prepares reviewable service-recovery and outreach options without contacting customers or making offers.
Persona: Relationship Manager
Prompt: Prepare options for Marcus that I can review before anyone contacts him or changes a fee.
Expected synthetic evidence: Marcus Johnson, No customer was contacted.
It is used in the synthetic pilot for the Customer Sentiment and Churn Prediction Agent to prepare reviewable retention options. The docs say it generates service-recovery and outreach options without contacting customers or making offers.
The documentation requires stating that the result is not legal, regulatory, insurance, lending, tax, investment, or financial advice. It also must state that no approval, communication, filing, account change, payment, order, transaction, or other external action occurred.
According to the docs, you need to identify the exact fictional record or report scope and use the synthetic operating snapshot to provide source-backed evidence. The provided materials do not specify installation, keys, or runtime requirements; see the source repository.
Rank a fictional contract portfolio to prioritize legal review attention.
Explain synthetic cross-sell affinity rules and benchmark assumptions from fixed evidence.
Compare contract clauses against internal policy requirements and identify gaps.
Draft prioritized amendment positions and legal escalation points for contract renegotiations.
Scan synthetic customer records to find evidence-backed cross-sell product gaps.
Draft prioritized cross-sell recommendations and a reviewable engagement plan from synthetic data.
Prioritize churn reviews with transparent evidence and suggested next steps.
Aggregate customer feedback into theme insights and weekly action priorities.
Identify top fixable customer issues and draft response templates from support data.
Prepare customer record fields for cleaner reporting and standardized analysis.
Review pipeline health, surface risks, and plan weekly deal priorities.
Draft common legal responses from templates with built-in escalation checks.