Cognitive System Value Model
The Cognitive System Value Model

The Cognitive System Value Model

The race to develop and deploy cognitive systems is on. Immature and rapidly changing, the current cognitive systems landscape is made up of a small but steadily growing group of cognitive tools, APIs, platforms, and application vendors. The cognitive “holy grail” they seek: high-impact, rapidly-delivered cognitive systems.

In the Cognitive Systems market, five deployment types have emerged:

1. Cognitive Tools & APIs
2. Embedded Cognitive/AI Functionality
3. Native Cognitive Applications
4. Cognitive System Deployments
5. Autonomous Cognitive Systems

Current Cognitive System deployments have a variety of approaches, ambitions, and complexities. Early developers and funders of Cognitive Systems place a premium on how long it takes to “go live” and recognize material cognitive benefit.

For each Cognitive System deployment type, the following Cognitive Value Model provides context, illustrates the time-to-value versus the relative benefit realized, and identifies sample technologies and implementations.

1. Cognitive Tools & APIs

Developer / user tooling, plug-ins and callable APIs
Bound and/or semi-structured, finite corpus (if any)

Examples:
- IBM Watson Analytics
- IBM Watson Conversation API
- IBM Watson Chatbot
- IBM Watson Visual Recognition API
- Microsoft Cognitive Services

Additional context:

Cognitive Tools and APIs have low-level elements that developers use to turn programmatic applications into cognitive deployments. They have a relatively fast time-to-value set against incremental benefit.

2. Embedded Cognitive/AI Functionality

Adaptive, contextual, application-native cognitive features
Controlled, pre-defined corpus and data

Examples:
- Adobe Sensei
- Google Waze
- Google Search
- H&R Block
- Salesforce.com Einstein
- IBM Watson Security
- IOT Applications
- Weather Applications

Additional context:

Embedded Cognitive/AI Functionality is application-specific, and performs pre-defined and function-specific tasks. This Cognitive Deployment Type has a shorter time-to-value as these capabilities are typically built on a cognitive platform or application. The resulting functionality is rarely seen as overtly cognitive by the application end-user because it is augmentative to the functionality provided in core programmatic applications.

3. Native Cognitive Applications

Purpose-built, contextual, cognitive platform-based applications
Multi-level, hybrid corpus; standardized data

Example:
- NorthPage Digital Marketing Intelligence

Additional context:

Native Cognitive Applications are question-based systems that intelligently identify associations, make recommendations, and process volumes of data. Single-domain, Native Cognitive System implementations achieve high levels of cognitive value in minimal time. Additionally, Native Cognitive Applications are designed and built around a cognitive core – as opposed to adding cognitive capabilities to a programmatic system.

4. Cognitive System Deployments

Corpus formulation; Data sourcing & normalization
Cognitive system training, testing, refinement, deployment
Ongoing management / maintenance
ERP-like in scope & resource requirements

Example:
- IBM Watson ground-up applications

Additional context:

Cognitive System Deployments are large-scale, custom enterprise applications that are cognitive-based or cognitive-augmented. They typically require resource-intensive, multi-year efforts before the system goes live. The levels of investment and project complexity represent the greatest risks to recognizing desired benefits from Cognitive System Deployments.

5. Autonomous Cognitive Systems

Multi-application, multi-dimensional cognitive systems
Real-time, multi-corpus development and factoring
Multi-mode and multidimensional data

Example:
- Self-piloting vehicles

Additional context:

Autonomous Cognitive Systems are the current - and temporary - pinnacle of cognitive systems. Wrapped in complexity, they aggressively advance the state of cognitive systems.

Take the Guesswork Out of Optimizing Complex Digital Programs

Master your digital priorities, performance and competition through frictionless, actionable intelligence.