Phoenix - Your AI AssistantPrompt Engineering for Customer AI — Brand-Safe Outputs
Delivery in
5 days
- Views 20
Amount of days required to complete work for this Offer as set by the freelancer.
Rating of the Offer as calculated from other buyers' reviews.
Average time for the freelancer to first reply on the workstream after purchase or contact on this Offer.
What you get with this Offer
I will engineer prompts for your customer-facing AI application — combining persona definition matching your brand voice, constraint engineering preventing off-brand or inappropriate responses, topic boundary enforcement keeping the AI within its intended scope, and escalation behaviour design for queries the AI should not handle. Customer-facing AI prompts have safety requirements that internal AI tools don't — a customer-facing AI that uses your brand's competitor names, makes unverifiable claims, discusses sensitive topics, or produces legally inadvisable statements creates reputational and legal risk that prompt engineering must prevent structurally rather than relying on the model's default behaviour.
The engineering covers brand persona definition, positive behaviour instructions, explicit negative constraints for off-brand responses, topic boundary enforcement, safe escalation language for out-of-scope queries, adversarial input testing, and a brand-safety test suite.
The engineering covers brand persona definition, positive behaviour instructions, explicit negative constraints for off-brand responses, topic boundary enforcement, safe escalation language for out-of-scope queries, adversarial input testing, and a brand-safety test suite.
What the Freelancer needs to start the work
Please share your brand voice guidelines, your AI application's intended scope, topics that must never be discussed, your escalation process for out-of-scope queries, and your LLM provider.