AIN System Development:
A Customer-Centered Approach
James T. Smith
Briefing to
GTE Laboratories

ABSTRACT
An Architecture Which Places the Customer as the Focus
Of All Service Design, Development, Deployment, & Management
Several Technologies from the Field of Distributed Artificial Intelligence
Are Integrated & Leveraged with Far-Reaching Consequences

The One-Size-Fits-One Product
In the old days--about three or four years ago--we consumers asked the companies we did business with for higher quality and greater responsiveness.  They did not let us down.
Today, however, responsiveness and quality no longer guarantee that consumers will be loyal to those we do business with.  Aware that companies will give us what we want, we’re asking for more, and that “more” is flexibility.
We want them to give us what we want, not what they want to give us.
Flexibility means we’re looking for companies to treat us as individuals and not like members of the herd.
Gary Heil & Tom Parker, “One Size No Longer Fits All,” Information Week, 27 Feb 95, p. 112

The Customer Revolution
As they did with responsiveness and quality, businesses of all types and sizes have heeded the call and are giving customers the opportunity to have it their way.
At one time we were satisfied with a one-size-fits-all product or service; now we want businesses to bend to our will.
On a rational level, this means we’re more insistent that a product meet our special needs.
Information and new technology are at the heart of the move toward customizing service.
Gary Heil, Tom Parker, & Rick Tate, Leadership And The Customer Revolution, Van Nostrand Reinhold, 1995

Listen to Your Customers
Customers today expect greater variety and customization, more sophisticated product features, and better service.
In response, savvy marketers are shifting from a traditional “make and sell” approach to a knowledge-intensive “sense-and-service” strategy.
Raw data is transformed into information and, through the use of sophisticated tools and educated users, converted into useful knowledge.
Information systems can impose greater structure on a customer-service process, which offers the potential for greater management control and impressive efficiencies.
But too much structure can result in rigid operating procedures that run counter to customer expectations of reasonable flexibility.
James Cash, “Listen to Your Customers,” Information Week, 27-Feb-95, p. 108

Gaining Customer Loyalty
The Ritz-Carlton Hotel Co. exemplifies the value of orchestrating processes, systems, structures, information technology, and people.  All work in concert to support a strategy of providing customized, personal, high-level services to guests of the company’s hotels.
The Ritz vision for IT, explained in a recent Harvard Business School case study, “The Ritz-Carlton: Using Information Systems to Better Serve The Customer,” focuses on turning guest preference data into actionable information.  Such information is available to each employee, who can use it to provide a more personal level of service.
The result:  In July 1994, Consumer Reports rated the Ritz-Carlton first among luxury hotels.
James Cash, “Gaining Customer Loyalty,” Information Week, 10-Apr-95, p. 108

Customer-Centered = Problem-Centered
Network’s Solution-Space
AIN Call Models Evolve & Service Implementations Are Modified
Customer’s Problem-Space
Semantics of Service Context and Behavior for the Customer’s Management Must Remain Consistent

Problems with Current Approaches
Cumbersome Service Management Interfaces
Designer-Centered  v. Customer-Centered Organization
Duplication  v. Synergism of Service Information

Problems with Current Approaches
Cumbersome Service Management Interfaces
Many Services Available Today Are Not Used by the Customer
Not Because the Services Are Not Useful
But Because They Are Too Difficult or Awkward for the Customer to Manage or Operate

Problems with Current Approaches
Designer-Centered  v. Customer-Centered Organization
Services Designed from the Call-Model Up  v. from the Customer View Down
Service Management Is a ‘Window’ into the Service Implementation
Emphasis on Procedural Implementation v. Declarative Functionality of the Service

Designer-Centered v. Customer-Centered
A Simple Implementation Example
Currently Manage Lists of Numbers
Maps Naturally to the Service Designer’s Solution
Should Manage lists of Parties, i.e., Names
To Whom Telephone Numbers, Email ID’s, etc. May Be Associated
‘Phonebook’ of ALL Customer Contacts
Customer & Problem Centered Implementation Approach

Problems with Current Approaches
Duplication v. Synergism of Service Information
Each Service Designed Independently of All Others
Need of the Same or Complementary Information by Multiple Services
Example — a Number Added to One Service List Must Be Removed from Another

Synergism of Service Information
Customer Information Should Be Entered & Managed
In Customer’s Problem-Centered Terms
One Logical Time [One Place] Only
In a Consistent Manner [Whatever the UI]
Independent of:
Which, How Many, or What Types of Services Are Subscribed
Which Do or Could Depend Upon that Information

Top-Level Service Management & Delivery
Information & Processing
Customer-Specific
Service Subscription
Customer Context
Service Context
AIN System-Specific
Service Description
Service Management
Service Control Loop

Top-Level Service Management & Delivery
Administration Manager [AM]
Controls Customer’s Modification & Management of Information
To Maximize Benefit of Positive Feature Interactions,
To Minimize Negative Interactions
Feature Manager [FM]
Controls & Manages Real-Time Execution of the Customer’s Services
As Configured by the Customer’s AM

Customer-Centered Profiles

Customer Service Profile: Spreadsheet Form

Far-Reaching Consequences of this Design
Customer-Centered Service Management
User Interface Synergism
User Information Synergism
Normalized Information & Processing
Symbolic Declarative Information
Flexible Intelligent Processing Strategy
Information Abstraction & Late-Binding
Feature Interaction & Negotiation Management

Customer-Centered Service Management
The Customer Uses Terms that Are Meaningful to Him
To Interact with Service Management
To Organize and Manage Information
At a Declarative Problem-Statement Level
Independent of the Procedural Implementations of those Services

User Interface Synergism
Consistent ‘Look-n-Feel’ Across All User Interfaces & Paradigms
Symbolic Label — Multiple Interface Representations
e.g., Textual ‘Family,’  Verbal ‘Family,’ &  Icon ‘Family’
Traditional DTMF, PC, GUI Interfaces
Multi-Lingual Natural Language Interfaces

User Information Synergism
All Customer-Providable Information Should Be Entered & Managed
In Customer-Centered Terms
In a Consistent Manner Independent of UI
Once — per Item Update
Independent of
Which, How Many, or What Types of Services Are Subscribed
Which Do or Could Depend Upon that Information

Normalized Information & Processing
Vertically Normalize the Management of Each Synergized Information Type
Encapsulate the Representation & Processing of Each Information Type
Plug-n-Play Model of Information Management and Usage
The Management & Use of Each Type of Plug-n-Play Information Is a Distinct Subscribable Capability

Symbolic Declarative Information
Symbolic Entities & Statements
Independent of Their Underlying Representations and Implementations
Express What Service Behavior or Activity Is Desired by the Customer
Problem-Centered Rather than Solution-Centered

Flexible Intelligent Processing Strategy
Symbolic Declarative Representation
Neutral to Specific Service Features
Provides Basis for Unlimited Flexibility
Applied to Many Different Applications
Utilized in Different Ways [Purposes] under Different Circumstances
Only the Customer’s ‘What’ Is Specified
‘When’ & ‘How’ Resolution Are Determined by the Network

Information Abstraction & Late Binding
Applicable to ALL Network Elements
Horizontal Layers of Abstraction
Symbolic Information at Each Layer
Declarative Abstraction [Model] of Processing [Procedures] of Lower Layers
Different [Real-Time] Circumstances
Applicable to Different Requirements
Mapped to Different Procedural Realizations
Information Abstraction = Service Capability Reuse

Feature Interaction & Negotiation
General Service Execution
Customer Contexts = Preferences
FM Heuristically Selects Service Context
Better Matches Customer Preferences
Is Compatible with System States and Capabilities
FM Utilizes Customer & System Preferences
Negotiate Behavior to Be Delivered
Provide Acceptable Behavior to Involved Parties
Meaningful Degradation of Services

Far-Reaching Consequences of this Design
Customer-Centered Service Management
User Interface Synergism
User Information Synergism
Normalized Information & Processing
Symbolic Declarative Information
Flexible Intelligent Processing Strategy
Information Abstraction & Late-Binding
Feature Interaction & Negotiation Management

Customer-Centered Service Management
The Customer Uses Terms that Are Meaningful to Him
To Interact with Service Management
To Organize and Manage Information
At a Declarative Problem-Statement Level
Independent of the Procedural Implementations of those Services

Customer-Centered Service Management
Customer-Meaningful Terminology — Information Abstraction
Customer-Defined Terms
Examples — ‘OfficeHours,’  ‘MyOffice,’  ‘Family,’  ‘DoNotDisturb,’  ‘InMeeting’
System-Defined Terms
Examples — ‘TeleMail,’  ‘MailBox,’  ‘Alert,’  ‘Busy,’  ‘ForwardCall’

Customer-Meaningful Terminology
Only Declarative Statements of What the Customer’s Problem Is
No Procedural Statements of How to Achieve a Solution

Customer-Centered Label System
Facilitate Customer-Centered Management of Services
Assignment of Logical Names for Easy Reference & Recall
Abstract User Interface Information from Its Underlying Representations
Symbolic Rather than Direct Manipulation of Information

Customer-Centered = Problem-Centered
Customer’s Service Problems
Specification of Service Context and Behavior Should Remain Consistent
Customer’s Problem Statement Should Be Invariant to Service Designer’s Solutions
Service Designer’s Solutions
AIN Call-Models. etc. Will Evolve, and Implementations Will Be Modified
Mappings of Problems to Solutions
Re-Interpreted, Re-Evaluated, Updated

Customer Service Profile: Spreadsheet Form

Far-Reaching Consequences of this Design
Customer-Centered Service Management
User Interface Synergism
User Information Synergism
Normalized Information & Processing
Symbolic Declarative Information
Flexible Intelligent Processing Strategy
Information Abstraction & Late-Binding
Feature Interaction & Negotiation Management

User Interface Synergism
Consistent ‘Look-n-Feel’ Across All User Interfaces & Paradigms
Symbolic Label — Multiple Interface Representations
e.g., Textual ‘Family,’  Verbal ‘Family,’ &  Icon ‘Family’
Traditional DTMF, PC, GUI Interfaces
Multi-Lingual Natural Language Interfaces

Consistency Across All User Interfaces
Support Multiple User Interfaces
Multiple Access Methods Supported
Examples — DTMF, Voice-Activation, ADSI Terminal, Windows PC, Macintosh
Extendible as New UI’s Are Developed
Example — Icons for GUI’s for PC’s, & PIC’s for Video Phones PC’s
All Interfaces Are Created Equal
Example — Configurable F5 Key on a PC Matched with Configurable #5 of DTMF Phone

Consistency Across All Interfaces
Pseudo-Natural Language
Interface-Neutral Interaction Model
Provides Framework for Customer Expression
User-Specific Grammar & Vocabulary
Subscribed Service Capabilities Define Domain of Permissible Statement Types
Mappable to Each User Interface
Consistent Navigation Rules Across DTMF, IVR’s, PC Menu’s, etc.
Encapsulatable in other Applications

Example Usage of the Label System
English-like Natural-Language UI
Add Johnny to Friends
New Individual MyFax 214-718-6398
If Work And RcvFax, Then Fwd2 MyFax And Page MyPager
New Service MailPage Is VoiceMail And Page
If Boss And TimeOut, Then MailPage MyVMail And MyPager

Example Usage of the Label System
Spanish-like Natural-Language UI
Anadir Juan A Amigos
Persona Nueva MiFax 214-718-6398
Si Trabajo Y RcvFax Entonces Transferir MyFax Y Page MiBeeper
Servicio Nuevo CorreoBeeper Es Correo Audio Y Beeper
Si Jefe Y TempoExpirar Entonces CorreoBeeper MiCorreoAudio Y MiBeeper
by Ricardo A. Negrette

Label-Centered = Customer-Centered
Focal Point of Customer Interaction’s with & Management of Services

Far-Reaching Consequences of this Design
Customer-Centered Service Management
User Interface Synergism
User Information Synergism
Normalized Information & Processing
Symbolic Declarative Information
Flexible Intelligent Processing Strategy
Information Abstraction & Late-Binding
Feature Interaction & Negotiation Management

User Information Synergism
All Customer-Providable Information Should Be Entered & Managed
In Customer-Centered Terms
In a Consistent Manner Independent of UI
Once [per Item Update]
Independent of
Which, How Many, or What Types of Services Are Subscribed
Which Do or Could Depend Upon that Information

Customer-Centered Information Synergism
Customer-Centered Organization of Information
Much Customer-Entered Information May Be Applicable to Many Services
Express Such Information in a Service-Neutral Manner
Apply this Information Consistently to Provision All the Customer’s Services

Customer-Centered Information Synergism
Customer PIM -- Personal Information Manager
Customer Context Profile — Contains Customer-Defined Information
Identifies -- Who, When, Where, Why
Customer Content Profile — [Futuristic]
Content Attributes -- FAX, Movie, E-Mail
Service Context Profile — Associates this Information with Subscribed Services
What to Do — Whenever, Whoever, Whatever

Customer-Centered Profiles

Far-Reaching Consequences of this Design
Customer-Centered Service Management
User Interface Synergism
User Information Synergism
Normalized Information & Processing
Symbolic Declarative Information
Flexible Intelligent Processing Strategy
Information Abstraction & Late-Binding
Feature Interaction & Negotiation Management

Normalized Information & Processing
Vertically Normalize the Management of Each Synergized Information Type
Encapsulate the Representation & Processing of Each Information Type
Plug-n-Play Model of Information Management and Usage
The Management & Use of Each Type of Plug-n-Play Information Is a Distinct Subscribable Capability

Independent Information Capabilities
Examples of Independent Information Types
Temporal — Daytimer
Personal — Phonebook
Customer State — ‘Involved,’ ‘Leisure’
Content — Voicemail & Email Messages, Multi-Media Data Streams

Service-Driven v. Capability-Driven
Traditional Service-Driven View
Tight [Hardcoded] Integration of Service Interdependences by Service Designer
Blurring of Independent Capabilities within the Application Service
Object-Model Is one of Inheritance
White-Box Interdependences
NOT Plug-n-Play

Service-Driven v. Capability-Driven
New Capabilities-Driven View
Independent Building-Blocks from Which
Customer May Customize Desired Service Behavior
Object-Model Is One of Delegation
Black-Box Independences
Use Plug-n-Play Capabilities

SIB’s Service Independent Building Blocks
SIC’s -- Service Independent Capabilities
Example -- the DayTimer Is Completely Independent of the PhoneBook
Plug-n-Play Capabilities
One DayTimer Interchangeable for Another
Independent of All other Customer Chosen Capabilities and Services
Brand-Name v. Generic Implementations of Capabilities

Customer Service Profile: Spreadsheet Form

Far-Reaching Consequences of this Design
Customer-Centered Service Management
User Interface Synergism
User Information Synergism
Normalized Information & Processing
Symbolic Declarative Information
Flexible Intelligent Processing Strategy
Information Abstraction & Late-Binding
Feature Interaction & Negotiation Management

Symbolic Declarative Information
Symbolic Entities & Statements
Independent of Their Underlying Representations and Implementations
Express What Service Behavior or Activity Is Desired by the Customer
Problem-Centered Rather than Solution-Centered

Phone Number = Telephony Symbol
Phone Number Is a Universal Symbol
Has Multiple Implementations
Rotary Dialer, Touch-Tone Keypad
Pulse-Dialed, DTMF, etc.
BCD-encoded, ASCII-string, Binary, etc.
Overloaded for Multiple Meanings
Caller-ID -- Who Is Calling, NOT What Device
800, 500-like Services, ‘500-CALL-GTE’
Black-Box Processing by the Network
Has Universal Meaning to the User

Extending the Set of Telephony Symbols
One Number Corresponds to One Symbolic Statement
No Longer Rich Enough User Interface
Extension of Telephony Symbolics
IF Nancy CALLS,  DURING WorkHours, SEND2 Secretary
DURING WorkHours,  IF Nancy CALLS, SEND2 Secretary
SEND2 Secretary,  DURING WorkHours,      IF Nancy CALLS

Declarative v. Procedural Knowledge
Declarative Knowledge
Encoded in Passive Data Structures
Distinct from the Procedures that Implement that Knowledge
Procedural Knowledge
Encapsulated, or Compiled, into the Procedures that Implement It
Ultimately — a Trade-off
What to Encode & What to Compile ?

A Better Knowledge Paradigm
Bi-Model Knowledge
Incremental Compilation of Knowledge
Duality — Compiled & Interpreted Forms
One-to-Many — One Declaration May Have Many Procedural Implementations
Analogies of this Paradigm
Floating-Point in Hardware & Software
Object-Code & Source Code Libraries
Plug-n-Play Components (H/W & S/W)

Declarative Customer Service Context
Symbolic Interpretation of Meaning
Independent of the Implementations of these Symbolic Entities
‘DURING WorkHours’ Is Independent of the Implementation of the Customer’s Daytimer
Customer’s Desires Are Preserved
BUT No Longer Hardcoded [Bound] to a Particular Implementation or Technology
The Network Intelligently Understands
& Processes the Customer’s Desires

Customer Service Profile: Spreadsheet Form

Symbolic Abstract Reasoning
Symbolic Analysis of the Spreadsheet Form of the Service Context Profile
Independent of their Implementations
Five Service Contexts Share the Common Time Period: ‘Office’
Two Involve the Customer State: ‘Lunch’
Three Involve the System State: ‘Busy’
One Each Are Particular to the ‘Boss’ & the ‘Secretary’
The Other Defaults to ‘ANY’ [*] Customer State & Calling Party

Far-Reaching Consequences of this Design
Customer-Centered Service Management
User Interface Synergism
User Information Synergism
Normalized Information & Processing
Symbolic Declarative Information
Flexible Intelligent Processing Strategy
Information Abstraction & Late-Binding
Feature Interaction & Negotiation Management

Flexible Intelligent Processing Strategy
Symbolic Declarative Representation
Neutral to Specific Service Features
Provides Basis for Unlimited Flexibility
Applied to Many Different Applications
Utilized in Different Ways [Purposes] under Different Circumstances
Only the Customer’s ‘What’ Is Specified
‘When’ & ‘How’ Resolution Are Performed by the Network

Domains of Intelligent Flexibility
Parallel & Distributed Processing
Facilitated by the Normalization of Information Types
Static Analysis & Feedback
Select Specific Processing Strategies
Directed Real-Time Processing
to Yield a Better Performance
Fault-Tolerant Behavior
Real-Time Re-Evaluation & Adaptation

Support of Parallel Distributed Processing
Facilitated by the Normalization of Information Types
Daytimer, Phonebook, Customer States, System States, etc.
Are Independently Managed
Service Context Profile Maintains Their Factored Nature
They May Be Processed in Parallel
The Results from One Area Can Focus Processing of the Others

Static Analysis & Feedback Adaptation
Analysis & Feedback Source & Usage
Feedback from Billing & Usage Systems for Prediction of Future Expected Use
Adapt to [Learn] the Customer’s Preferences
Capture Implicit Customer Characteristics
[This Is Applicable to Marketing ! ]
Organize Profile Information -- Improve Real-Time Performance & Customer Satisfaction
Consider Most Likely [to Occur]
Consider [Customer] Preferred Cases

Directed Real-Time Processing
Many Possible Real-Time Strategies
Compile Integrated Customer-Specific Decision-Graphs
from Static Analysis & Declarative Customer Service Management
Compile Each Capability Independently
Daytimer Distinct from Phonebook, etc.
General Rule-Engine Compile
Compile Symbolic Links
Real-Time Variations of RETE, TREAT, etc.

Real-Time Compilation Strategies
Procedural Compilation
Compiles the Information Along with Its Representation
Suppresses Original Symbolic Knowledge
Artificially Reorganizes & Restructures
Declarative Compilation
Compiles the Representation But NOT the Information
Preserves Original Symbolic Knowledge
Natural Factorization -- an Example

INC-2 Service Specification

INC-2 Service Specification

INC-2 Service Specification

INC-2 Service Specification

Far-Reaching Consequences of this Design
Customer-Centered Service Management
User Interface Synergism
User Information Synergism
Normalized Information & Processing
Symbolic Declarative Information
Flexible Intelligent Processing Strategy
Information Abstraction & Late-Binding
Feature Interaction & Negotiation Management

Information Abstraction & Late Binding
Applicable to ALL Network Elements
Horizontal Layers of Abstraction
Symbolic Information at Each Layer
Declarative Abstraction [Model] of Processing [Procedures] of Lower Layers
Different [Real-Time] Circumstances
Applicable to Different Requirements
Mapped to Different Procedural Realizations
Information Abstraction = Service Capability Reuse
Run-Time Rather than Compiled Reuse

Bi-Model Layers of Abstraction
Information Duality of Each Layer of Abstraction -- Recursive Downward
Encoded Declarative Self-Knowledge
Compiled Procedural Functionality
Bi-Model Functionality
Incremental Compilation of Knowledge
Natural Factorization
One-to-Many
One Declarative Description May Map to
Many Possible Procedural Implementations

Information Abstraction & Late Binding
Realization of Symbolic Information
Subject to Real-Time Circumstances
Evolves in Response to Changes by Customer or Technology
Example:  ‘Alert Loud’ & ‘Alert Quiet’
Dependent upon Available Resources
Buzzing Beeper  v.  Blinking Screen
Dependent on Customer Circumstances
Hearing Impaired,  Personal Preferences

Service Description Profile
Comprehensive Declarative View
AIN System Features & Capabilities
Their Applicability [Heuristics] to Various Categories of Customer Problems
Example:
Active Alerting -- Providable by Distinctive Ringing, Paging, Character Generator [TV]
Passive Alerting -- Providable by Email, Voicemail, Distinctive Dial Tone
Caller Identification -- Providable by ANI, PIN, VoicePrint, Retina Scan

Service Management Profile
Declarative to Procedural Mapping
Associates Declaratively Described Features & Capabilities
Their Procedural Realizations in the AIN
Specific Platforms,  Specific Resources,  Specific Conditions,  etc.
Example:  Access to [Support for]
Distinctive Ringing?  Paging ?  Email ?  ANI ? Voicemail ?  Distinctive Dial Tone ?  PIN ?

The Various Context Profiles
Information Consists of Symbolic Patterns
Meanings May Be Directly Determined
e.g., Resolution of ‘WorkHours,’ ‘MyFamily’
Meanings May Be Indirectly Inferred
e.g., Inference of ‘Alert Loud,’ ‘Alert Quietly,’ ‘ID-Caller’
Heuristics May Be Applied to Deduce Secondary Relationships
e.g., Preference to More Specific Context

Intelligent Entities — Self-Awareness
An Intelligent Entity Understands Its Functions
Can Evaluate Alternate Approaches
Can Explain Itself — Intentions & Actions
Compiled Automatic [Routine] Responses
Can Innovate — Adapt & Apply Non-Routine Approaches
Guided By Heuristics — ‘Rules of Thumb’ about General Approaches, etc.

Normal v. Abnormal Circumstances
General Processing Heuristics
Under Normal Circumstances
Behavior Delivered as Requested
Compiled Procedural Approaches
Self-Knowledge May Not Be Needed
Under Abnormal Circumstances
Behavior NOT Delivered as Requested
Feature Interaction & Negotiation Mgt — Intelligently Provide Acceptable Behavior

Far-Reaching Consequences of this Design
Customer-Centered Service Management
User Interface Synergism
User Information Synergism
Normalized Information & Processing
Symbolic Declarative Information
Flexible Intelligent Processing Strategy
Information Abstraction & Late-Binding
Feature Interaction & Negotiation Management

Feature Interaction & Negotiation
General Service Execution
Customer Contexts = Preferences
FM Heuristically Selects Service Context
Better Matches Customer Preferences
Is Compatible with System States and Capabilities
FM’s Utilize Customers’ & Systems’ Preferences
Negotiate Behavior to Be Delivered
Provide Acceptable Behavior to Involved Parties

Undesirable Feature Interactions
Single-Customer Interactions
Example:  Call-Forwarding  v. Call-Waiting
Solution:  Provide Capability to Identify Preferences for Different Contexts
Result:  Interaction  =>  Coordination
Multi-Customer Interactions
Example:  Caller-ID  v. Call-Blocking
Solution: Provide Capability to Negotiate Acceptable Compromises
Result:  Impasse  =>  Negotiation

Negotiation — Defined
Negotiation — Iterative Process
Announcement — Pose Need to Other Entities
Bid — Receive Offers from Responders
Award — Accept Best Offer of Assistance
Variations in the Process
Counter-Offers
Teaming & Coordination
Subcontracting — Secondary Negotiations

Negotiation Characteristics
Fundamental Components
Exchange of Information
Evaluation of Exchanged Information
Final Agreement
Preparation for Negotiation
Absolute Bounds
Compromise Fall-Back Positions
Evaluation Criteria

Examples of Negotiation in the Network
Types of Simple Negotiation
Credit-Card Calling — Trust Me !
Collect Calling — Will You Accept ?
Person-to-Person — No One But You !
Cellular Roaming Plans
Parties Involved in Negotiation
Between Calling Parties
Between Calling Party & the Network
Between Networks

Multi-Customer Feature Negotiation
Goal of this Process
Secure Acceptable Service Behavior
As Satisfactory As Possible
Avoid Behaviors Unsatisfactory in Context
Criteria:  Behavior, Quality, Cost, etc.
Leverage Concepts Previously Discussed
Symbolic Declarative Constructs of Customer’s Service Administration
Self-Knowledge Incorporated within AIN Features & Capabilities
Feature Manager  =>  Negotiation Agent

Negotiation Infrastructure
Symbolic Declarative Taxonomies
Customer Requirements
Privacy,  Alerting,  Security
AIN Feature & Capabilities
Call-Blocking,  Caller-ID,  PIN-Screening
Associate Requirements & Capabilities
Privacy  => Call-Blocking
Alerting  => Caller-ID
Security => PIN-Screening

Negotiation Infrastructure
Taxonomies & Cross-Reference Requirements
Should Reflect Logical Patterns
From Abstract Generic to Concete Specific
Alert,  Alert Quietly,  Alert Quietly w/ Pager LCD, ...
Acquaintances, Friends,  Close Friends,  Johnny
Directly v. Analogously [Heuristically] Related
Messaging [Email] Is a Form of Passive Alerting

Feature Negotiation Heuristics
Satisfy the Request
At a Level of Specificity >= the Request
At a Level of Specificity < the Request
Based on a Subset of the Request
Based on Economical [Cost, Resource, etc.] Generalization of Solution
Based on Any Generalization of Solution
Based on Any Generalization of Request
...

Feature Negotiation => Mediated Access
Inter-SCP Feature Negotiation Between the Customer’s Resident SCP & that at the POP

Current Development of this AIN System
Have Developed Prototypes & Demos
Customer-Specific Components
Realty World
Have Developed Requirements for the Commercial Versions
Incrementally Incorporate These Concepts
Commercialization of Realty World
Currently Implementing Systems
GTE Developed; R&D w/ ISCP & AI-Net
Telecom’95 in Geneva, Switzerland

Future Goals & Developments
Ubiquitous Application of these Customer-Centered Design Concepts
Re-Engineering of the Total Service Delivery Process
Service Marketing
Service Management
Service Creation
Service Execution
Support Functions