Aisera Prompts

Act as an IT director at [COMPANY] ([EMPLOYEE COUNT] employees, [TICKETING SYSTEM] today). Deploy Aisera's AI Service Desk: connect the 400+ integration to [TICKETING SYSTEM] and our knowledge base ([KB URL]); enable Ticket AI auto-classification and routing for the top [NUMBER] ticket types ([PASSWORD RESET], [VPN ISSUE], [SOFTWARE REQUEST]); set the auto-resolution target ([PERCENT] percent in [DAYS] days) for the safe categories; keep [CATEGORY] human-handled. Include the Teams/Slack listening mode setup (which channels, the trigger phrases), the pilot group ([DEPARTMENT], [NUMBER] users, [WEEKS] weeks), and the change-management message ('the service desk now answers in [TIME], here is what changed').

Act as an HR operations lead at [COMPANY] hiring [NUMBER] people/quarter. Build the Aisera onboarding automation: when HR marks a hire in [HRIS], the AI agent provisions [ACCOUNT 1], [ACCOUNT 2], and [HARDWARE/SOFTWARE]; sends the day-one checklist to the new hire via [CHANNEL]; answers the inevitable questions ('how do I [COMMON QUESTION]?') from the HR knowledge base; and nudges the manager at day [N] for the [CHECKPOINT] check-in. Include the [NUMBER] questions new hires ask most ([Q1], [Q2], [Q3]) with their grounded answers, the exception path (contractors get [DIFFERENT FLOW]), and the metric (time-to-productive from [DAYS] to [TARGET DAYS]).

Act as a service owner at [COMPANY] who does not code. I need an AI agent that handles [PROCESS] (currently [NUMBER] tickets/month, owned by [TEAM]). Use Aisera's Agent Composer in natural language: describe the agent's job ('when an employee asks about [TOPIC], check [SYSTEM], then [ACTION]'); define its knowledge sources ([DOC 1], [DOC 2]); set its boundaries (never [FORBIDDEN ACTION], escalate when [CONDITION]); and test it with [NUMBER] real historical tickets before going live. Include the plain-language agent description to paste into Composer, the test ticket set, and the sign-off checklist (accuracy above [PERCENT] percent on the test set).

Act as a contact center manager at [COMPANY] with [NUMBER] agents handling [VOLUME] tickets/month. Deploy Aisera Agent Assist: real-time triage that classifies each incoming ticket ([CATEGORY], urgency, sentiment); auto-summarization so agents see the [NUMBER]-sentence gist instead of reading [NUMBER] messages; answer generation grounded in the KB; and next-best-action recommendations ([ACTION 1], [ACTION 2]). Include the agent training (one [HOUR]-hour session: how to accept, edit, or reject suggestions), the quality gate (suggestions below [CONFIDENCE] are hidden), and the metric (average handle time from [MINUTES] to [TARGET]).

Act as an SRE lead at [COMPANY] running [INFRA, e.g. Kubernetes on AWS]. Configure Aisera AIOps: connect observability ([DATADOG/PROMETHEUS/OTHER]); enable proactive incident detection on [SERVICE 1] and [SERVICE 2] with noise suppression tuned to our alert baseline ([ALERTS/DAY] today); set up automated root-cause analysis for [INCIDENT TYPE]; and define the auto-remediation runbooks the AI may execute unsupervised ([RUNBOOK 1]) versus those needing approval ([RUNBOOK 2]). Include the on-call integration (page [TEAM] via [PAGERDUTY/OPSGENIE] with the AI's diagnosis attached) and the post-incident review template the AI pre-fills.

Act as an IT service manager at [COMPANY]. Too many tickets are 'my [THING] expired/broke and nobody told me.' Deploy Aisera Ticket Concierge for proactive resolution: identify the [NUMBER] recurring issues ([CERT EXPIRY], [LICENSE RENEWAL], [PASSWORD EXPIRY]); have the AI detect them early from [SYSTEM] signals and message the affected user on [CHANNEL] with the fix ('your [THING] expires in [DAYS] days; click here to renew'); only open a ticket if the user does not act in [HOURS] hours. Include the message templates, the detection rules, and the metric (reactive tickets for these issues down [PERCENT] percent in [MONTHS] months).

Act as an operations analyst at [COMPANY]. The [PROCESS] workflow ([NUMBER] steps across [SYSTEM 1] and [SYSTEM 2]) eats [HOURS] hours/week. Rebuild it in Aisera's Prompt Studio with low-code: map each step to a prompt block with its inputs ([INPUT 1], [INPUT 2]) and outputs; add the approval gate at step [N] (approver: [ROLE], SLA [HOURS] hours); handle the [NUMBER] failure modes ([FAILURE 1]: retry; [FAILURE 2]: escalate to [TEAM]). Include the natural-language description of the workflow to start from, the test run with [NUMBER] sample inputs, and the handover doc so [TEAMMATE] can maintain it.

Act as a global IT leader at [COMPANY] ([COUNTRIES] countries, [EMPLOYEE COUNT] employees). Aisera supports 100+ languages with built-in detection. Plan the rollout: phase 1 ([LANGUAGE 1], [LANGUAGE 2], [PERCENT] percent of users); validate intent matching on [DOMAIN TERMS] per language with local IT reviewing [NUMBER] test conversations; configure language detection fallback (default to [LANGUAGE] when unsure); and keep the knowledge base canonical in [LANGUAGE] with AI handling the translation layer. Include the phase 2 language list ([LANGUAGES]), the per-language auto-resolution target (within [POINTS] points of English), and the governance (who approves a new language going live).

Act as an AI platform owner at [COMPANY]. Generic LLMs misunderstand our internal jargon ([TERM 1], [TERM 2]) and our [TICKETING SYSTEM] categories. Use Aisera LLM Studio: select the tuning dataset ([NUMBER] resolved tickets from [DATE RANGE], filtered to quality resolutions); define the evaluation set ([NUMBER] held-out tickets); tune for [TASK: classification/summarization/answer generation]; and compare tuned versus base model on accuracy for [CATEGORY]. Include the data hygiene rules (strip [PII FIELDS] before tuning), the retraining cadence ([FREQUENCY]), and the rollback criterion (tuned model underperforms base on [METRIC] by [POINTS] points).

Act as a CISO evaluating Aisera for [COMPANY] in [REGULATED INDUSTRY]. Build the security review package around the TRAPS framework (trusted, responsible, auditable, private, secure): document data flows (ticket text goes to [LOCATION], embeddings stored in [LOCATION], retention [DAYS] days); confirm the compliance mappings ([SOC 2], [GDPR/HIPAA], [OTHER]); verify the on-premises LLM option for [DATA CLASSIFICATION] data; and define the audit cadence (quarterly review of [NUMBER] agent decisions for bias and correctness). Include the [NUMBER] questions for the vendor ([Q1: data use in training], [Q2: sub-processors], [Q3: breach notification SLA]) and the red lines that would block the deal ([RED LINE 1]).