3. Intuit Proprietary & Confidential
The Semantic Technology Vision
3
"The Semantic Web is an extension of the
current web in which information is given well-
defined meaning, better enabling computers
and people to work in cooperation.“
--Tim Berners-Lee, James Hendler, Ora Lassila
(2001)
12. Intuit Proprietary & Confidential
Benefits and Next Steps
Immediate Benefits
– Cross promotion of articles with shared tags
– Provides better help to call-center agents
Next Steps
– Experiment with more advanced tagging: User generated
triples
– Use Common Tag to link concepts and deal with synonyms
Outcome: Leverage user contributed data to
– Improve knowledge base taxonomy: Better semantic graphs
that complements the directory tree
– Quicker and more relevant information for decision making
12
15. Intuit Proprietary & Confidential15
Problem #1: Payee Identification
Our Favorites:
DIOCESE OF SJ DES:REG SALARY
ID:RGD000000970668
INDN:RUKONIC,ALEKSANDRA CO
ID:1942734503 PPD
REF:000144317946121108250007029016260316
07013122
PAYPAL DES:INST XFER ID:5XD222SDHQ582
INDN:MARKO RUKONIC CO ID:PAYPALSI66
WEB
REF:000144317946121108250007031001321376
07013127
16. Intuit Proprietary & Confidential
SAFEWAY STORE 05-03 CUSTOMER
1114 PURCHASE #934054 FREMONTCA
FRESH CHOICE #4 04/30 CARD #6312
PURCHASE
#24210737120286180800552 SAN
JOSE, CA
PARAMOUNT'S GR AMERICA 04/28
CARD #1114 PURCHASE
#24610437119004024018539 SANTA
CLARA, CA
DIOCESE OF SJ DES:REG SALARY
ID:RGD000000970668
INDN:RUKONIC,ALEKSANDRA CO
ID:1942734503 PPD
REF:0001443179461211082500070290
1626031607013122
PAYPAL DES:INST XFER
ID:5XD222SDHQ582 INDN:MARKO
RUKONIC CO ID:PAYPALSI66 WEB
REF:0001443179461211082500070310
0132137607013127
16
From: To
• Safeway
• Fresh Choice
• Great America
• MY PAYCHECK
• Paypal Income
Quicken User Contributed Payee Identification
17. Intuit Proprietary & Confidential17
Problem #2: Categorization
04/20/08 CHEF CHU'S $23.67
04/19/08 CHEVY'S $11.34
03/18/08 STARBUCKS $3.78
02/01/08 FRESH CHOICE $17.22
How much did I spend this year on lunch? Coffee?
The chore of repetitious work:
Average customer has a checking and a credit card account
They have 50 transactions per month
Most of them are recurring – keep shopping at favorite stores
It takes a lot of time to categorize all those
Imagine 1000s of customers doing this work all the time, all
over again
- LUNCH
- LUNCH
- COFFEE
- DINING
18. Intuit Proprietary & Confidential18
Quicken ACE: Automatic Categorization Engine
• Quicken ACE collects information from all its users
and builds up its knowledge database.
• Some databases:
– Payee recognition
– Transaction categorization
– Bill and check recognition
– Personalized category management
– Popular categories for a payee
– Payee name completion, Quicken style
Every time a user categorizes a transaction or
renames a payee, we store that.
Lots of data – our valuable asset
19. Intuit Proprietary & Confidential19
1. Seeded with large static merchant list
2. Personalized category and payee renaming (with
new category, new merchant)
3. User contributed data is used to improve/enrich the
seed knowledge base
4. Community voting as the guardian
5. Statistics of different semantic attributes is
reflected in GUI
Quicken ACE: Evolve Semantics with User
Contributed Data
20. Intuit Proprietary & Confidential20
Tax Time!
• ACE can map every
expense to a tax line item
– Itemized deductions are not a
chore any more
– Where there is a link between
category and tax line item
– Small businesses and rental
property owners win!
– ACE supports any number of
category sets
21. Intuit Proprietary & Confidential
Lessons Learned
• User want to contribute when
– it is easy to make such contribution
– there is pain/value
• Tagging is a powerful way
–How complicated can we push it?
–Common Tags
21
22. Intuit Proprietary & Confidential
View Points
• Knowledge is a living breathing thing
–Ivory tower approach for building semantic webs won’t
work
• User contribution will stay
–As an iconic feature of Web 2.0, user contribution will
stay and will pave the way to Web 3.0
• Call for user centric tool design
–Need tools to enable, improve, and fuse user
contributed data with expert maintained
ontology/semantics.
22
Intuit.Semantic Technology in a software company?Intuit as a content management organization.
The semantic web is envisioned to be a universal medium for data, information, and knowledge exchange.The semantic web comprises a set of design principles, collaborative working groups, and a variety of enabling technologies. A variety of data exchange formats provide a formal description of concepts, terms, and relationships within a given knowledge domain.Applications: Search, data reconciliation
Intuit’s roles in content management: Publisher, consumer, aggregatorIdeal: Intuitive connection among all documents of our entire financial life.How?The Semantic Technology vision.Implementation dilemma: Are we dead in the water?
Simple example.Introduction of QuickBooks.Overview of the support content model.
Review of QuickBooks support site.Knowledge management ecosystem for a growing business.
The irony: Standardization and usability.Lacking: Labor intensive cataloging effort; timeliness; flexibility.
Our implementation of the user tagging system.Folksonomy and Web 2.0.How it works.
Early adoption.Funny tags.Two takeaways: Unmet user need; Semantic Technology opportunity with user contribution.Revisit the question: Are we dead in the water?Proposal: user generated semantic grid.Zoom in on some early forms of user triple statements.
Examples for analysis:http://support.quickbooks.intuit.com/support/pages/knowledgebasearticle/1002282http://support.quickbooks.intuit.com/support/pages/knowledgebasearticle/1011293http://support.quickbooks.intuit.com/support/pages/knowledgebasearticle/1000335
Message: Social as an easy way (maybe the only way) that the Semantic Web can happen.Technology gap between Web 2.0 and Web 3.0.Efforts underway: CommonTag initiative; OpenMind initiative.Call for joining force: Immediate steps.Asking for help: Adoption of Semantic Technology in our partnering institutions.Closing remarks: A living and breathing knowledge management eco-system.
Yet another example on leveraging crowd in semantic knowledge.Background about Quicken: personal finance management -- aggregate all your spendings/incomes in one place with an intuitive interface -- Automated (download info from Financial Institutions) -- classify your spending into It looks neat and easy and nice, right? Not so actually if you have peek at the internal.
Two problem:Payee identifications ( each CC has its own way to identify merchants) Categorization problem
What ACE does: Realize that it will take forever for all the FI to adopt any standardized semantic description of expendituresRealize that Quicken is a great platform for creating semantic rich features because of its large user base Leverage user contribution to solve both the payee identification problem and categorization problem Easy user participation Talk about Quicken ACE/iPhone?Talk about the airplane/twitter thingy?
80 categories of spending in Quicken, based on user contribution