What Does a SkillStream Performance Engine Do?
A SkillStream Performance Engine (SPE) applies a known method to perform a task effectively in a given context. It answers the question: "Given this situation, how should this known method be applied well?" and then applies it.
An SPE allows novices to reliably produce consistent, high-quality results on demand while progressively building the ability to perform the work independently. It also allows experts to work faster while maintaining consistency, surfacing errors, and keeping their skills sharp and up to date.
An SPE:
- Encodes a coherent, evidence-based model of excellent practice
- Constrains execution to prevent common failure modes
- Translates weak or vague intent into strong action
- Surfaces the right judgment points at the right time
- Explains why certain choices matter
- Tracks decisions, assumptions, and reasoning
- Builds user capability through repeated use
Crucially, an SPE does not just help people perform better. It helps them become better performers over time. SPEs make apprenticeship at scale possible.
What Is an SPE?
An SPE is an advanced expert system that enables someone to perform a task proficiently in a given context and teaches them how to perform it independently. It is a virtual expert that allows apprenticeship at scale.
How Does an SPE Work?
An SPE provides explanations, examples, templates, and other information relevant to performing and learning a task. It guides performance by decomposing the task into specific steps that the person follows with support from a specialized AI expert.
Each step is designed to make the task easier for the person and the AI expert to understand the step and contextualize the method being applied to the specific situation. In addition, each step is designed to teach the person, through repetition, the method and necessary skill(s) to perform the task independently.
An SPE is not only a performance tool, but also a learning modality that can be used in lieu of or as reinforcement to traditional training.
How Is an SPE Different Than a Traditional Expert System?
Traditional expert systems combine a knowledge base and an inference engine to help a person perform a task. They generally work well in narrow, well-defined fields where a human can define objective rules. They don't work well when the rules become complex, the situation is ambiguous, or applying a method requires judgment. They are rigid and not well-suited to ambiguous, novel knowledge work situations.
An SPE's specialized AI expert can contextualize the method for performing the task to the relevant situation, even when it is novel or ambiguous. It acts like a mentor by suggesting how to perform a step, giving feedback, or discussing factors that may influence how it is performed.
How Can an SPE Contextualize a Task Method to a Given Situation?
SkillStream Performance Engines approximate tacit skill application by statistically modeling how experts condition codified steps on situational variables and apply judgment within a known method, and reproduce that conditional behavior in new but similar contexts. This contextualization wasn't usually possible before generative AI.
Key capabilities that make it possible include:
- Large-scale pattern induction from expert behavior
- Representation of situational features and constraints
- Conditional generation based on multiple interacting variables
- Abstraction across cases, not memorization of rules
- Instruction-following tuned to intent, not syntax
- Ability to surface rationales and tradeoffs
- Tolerance for ambiguity and underspecification
- Long-context processing
What Skills Do You Need to Build SPEs?
Some new skills are needed to build SkillStream Performance Engines, but they overlap significantly with instructional design and development program deployment. In addition, SPEs can be built to assist in building SPEs.