Enterprises that are building Knowledge Graphs are rapidly getting a grip on unstructured data with current advances in Natural Language Processing (NLP) techniques. But there is still a large mass of unstructured data that is untapped and that is spoken conversations with customers. Speech to text for general purpose conversations (e.g. Google, Alexa, Siri) have proven themselves in the market to be highly accurate. However, speech recognition technology for domain specific industries with lots of product names, industry lingo, and acronyms often creates a challenge for accuracy and usefulness of the content.
In this Webinar we will demonstrate how taxonomy driven speech recognition helps solve these industry specific terminology challenges for real-time voice capture and how this process augments an Enterprise Knowledge Graph for customer insights.
video youtube.com/allegrograph
3. So Why is Speech Technology important
for Knowledge Graphs?
• Most Knowledge Graphs in the enterprise are 360 views on
customers
• To really understand your customer you need to listen to
what they say!
• Add that understanding to your knowledge graph and learn
and become better at understanding your customer
4. What could you do in your enterprise if Speech
Technology was perfect?
• Really hear and understand your customer!
• What do they think about you and your products and your processes
• Automatic Transcription of the interaction with your customer
• Legal compliance,
• Analyze complaints,
• Call and caller classification,
• Find patterns in calls
• Identify callers using biometrics
• And mitigate risks
• Are your agents saying the right things?
• Are they really helping the caller, are they conveying the right information, are they
hitting all the required talking points, do they use the right objection handing, are they
polite, non-confrontational, etc, etc.
5. What could you do in your enterprise if Speech
Technology was perfect?
• Optimize your sales process
• Analyze successful and failing conversations
• Did your sales agent talk about Budget, Authority, Need and Timeline? (BANT)
• Did your agent talk about the right products given the expressed needs of the caller
• When the customer brings up a competitive product, can you instantly help the agent
with the right object handling?
• When you now the industry of the caller, can you provide the agent with use cases for
that industry
• Based on the nature of the call: should you transfer to technical help or technical
consultant or try to close a deal.
• Help the sales agent write the perfect ‘sales notes’
10. The big players (Google, MS, IBM, Apple, and many more) now all
claim that they turned the dream of speech recognition into a reality
11. But be that as it may
In the context of understanding your
customers YOUR enterprise will say
Wow!
95 % is 95 % more
(or infinitely more)
than I can do now
12. But how can you tie NLP and Speech Technology together in one coherent
context so you really can learn, understand and help your customers?
• Our answer: a Semantic Knowledge Graph
13. So what do we do with NLP and Speech Technology
and Knowledge Graphs.
14. We do NLP in various domains
• Do you want to live forever?
• Biography Knowledge Graphs for
interesting (and/or public) figures
• What do my agents talk about? What
conversation style leads to more sales?
• A knowledge graph for
Intelligent Call Centers
16. Challenge: understand conversations to make agents smarter
u What are my customers and agents talking
about (via chat and email and voice conversations)
u Product categories, and sellable products
u Competitive products (and objection handling)
u What are the features and functions my customers care about (and
how do I related that to my main products and the competitors
products)
u What is the momentum for product X
u What is the sentiment as a function of the type of product
u Why is agent X so good at selling product Y
u Does agent ?X steer a conversation in the direction of Y?
u Where is the agent in the sales cycle (did she hit every BANT
category?)
17. Challenge: how do I make my agents smarter
u What recommended products and (industry) use cases should I present to my
agent based on
u Persona
u Industry
u Demand scenario
u Competitor
u Sales Cycle
u How can I help my agent with
u Objection handling
u Understanding the functionality of
products
19. What did we build?
u Industry Knowledge Graph (KG)
u Around companies, subsidiaries, contacts, products and services, industry
and product taxonomies, competing products, current technology stack,
propensity to buy, company news, hiring?, white papers – updated over time
u Taxonomies
u Taxonomy Based Entity Extraction &Sentiment Analysis
u Examples of analytics
u Text classification: find personas, demand scenarios, industry type
u Product Recommendations
u Speech recognition
u Custom services and how to make life for application developers easier
20. Foundational
shared
taxonomy
N3 universal sales-cycle and IT terms & concepts
taxonomy
Campaign Notes & Chats Taxonomies in N3 Knowledge graph
The Taxonomy enables identification of -
• Words
• Terms
• Titles
• Concepts
• Products
• Buyer intent
• BANT (the actual sales qualifying stages
of BANT)
• and Sentiment
in the Chats
21. N3 - Sales Cycle Foundational Shared Taxonomy
Hierarchy
of SALES
concepts
and
terms Synonyms,
Broader,
Narrower,
Related
concepts
and terms
This application is
where the Taxonomies
are created – these are
built to the W3C SKOS
industry standard
23. Taxonomy based entity extraction
u Regular taxonomy editors pretty good at entity extraction when prefLabels and altLabels are regular
words.
u AllegroGraph has entity extractor with specializations for product names.
u Providing altLabels for every product too time consuming, can be automated
u Also needed for post processing when doing speech technology
u Built in from AG 6.6
u When you need place names, people names, organizations, currencies, etc. we use specialized entity
extractors like Cogito or IBM Natural Language Understanding
u Also come with automatic linking to dbpedia, geonames, etc…
u Python Spacy if you need NLP capabilities -> POS, special language models, rules
u All of the above offer some form of Sentiment Analysis
26. 0 5 10 15 20 25 30
Tiona Hill
Christopher Spade
Devin Smith
Sidney Carr
Laura Pugh
Billy Young
JacobHolmes
Yannick Souna
George Hanna
LeahWagner
Mary Rowland-Doud
Robert Edwards
Paul Worley
Brad Mcdougald
Daulton Tyler
AllisonSlocomb
Ashley Etheridge
Elise Nemeth
Meghan Bush
Wesley Sites
Nicole Robinette
Jeff Thompson
Number of Negative, Neutral and Positive Chats
1200 1000 800 600 400 200 0
Ratio of Positive to Negative Chats
Query Results- Chat Sentiment – AI deduced
Some BDR agents are more positive than others
30. Product % Product %
Are BDRs spending their
time on the right
products?
Query Results – for BDRs Laura vs Sidney
% mention of specific SKUs for overall product mentions
32. Product Recommendations
u If a customer talks about X a BDR should bring up Y
u Based on Oddsratios
u Temporal co-occurrence corrected for frequency of each element in pair
u Used in Logistics, Health Care, Chomsky Graph
33. The Green boxes are AI
created likelihoods that these
other Products (connected
blue boxes) will also be
discussed.
34. Taxonomy BuildingEntity Extractor
AGraph Sem-Ingest
Analytic Tools
Real-time
Decision support
Ad Hoc Queries
AG Sentiment
Analyzer
Machine
Learning
Text
Classifier
CCOR Generator:
CoOccurence,
Correlation, Odds
Ratio
REST-Semantic
Search
Taxonomy Tools
AI Labelers
Guided Advisors
Taxonomy
Entities
Sentiments
Classification
Statistical Relations
Labels
Event Knowledge Graph
Text
Text, email, chats Voice to TextDocuments
(PDF, PPT, Word)
AllegroGraph KG
Platform
Knowledge Graph
Platform
35. Taxonomy Tools
AI Labelers
Guided Advisors
Taxonomy
Building
Taxonomy BuildingEntity Extractor
AG Sentiment
Analyzer
Machine
Learning
Text
Classifier
CCOR Generator:
CoOccurence,
Correlation, Odds
Ratio
REST-Semantic
Search
Taxonomy
Entities
Sentiments
Classification
Statistical Relations
Labels
Event Knowledge Graph
Text
AllegroGraph KG
Platform
36. Taxonomy Tools
AI Labelers
Guided Advisors
Entity Discovery
and Extraction
Taxonomy BuildingEntity Extractor
AG Sentiment
Analyzer
Machine
Learning
Text
Classifier
CCOR Generator:
CoOccurence,
Correlation, Odds
Ratio
REST-Semantic
Search
Taxonomy
Entities
Sentiments
Classification
Statistical Relations
Labels
Event Knowledge Graph
Text
AllegroGraph KG
Platform
37. Taxonomy Tools
AI Labelers
Guided Advisors
Sentiment
Analysis
Taxonomy BuildingEntity Extractor
AG Sentiment
Analyzer
Machine
Learning
Text
Classifier
CCOR Generator:
CoOccurence,
Correlation, Odds
Ratio
REST-Semantic
Search
Taxonomy
Entities
Sentiments
Classification
Statistical Relations
Labels
Event Knowledge Graph
Text
AllegroGraph KG
Platform
38. Taxonomy Tools
AI Labelers
Guided Advisors
Classification
Taxonomy BuildingEntity Extractor
AG Sentiment
Analyzer
Machine
Learning
Text
Classifier
CCOR Generator:
CoOccurence,
Correlation, Odds
Ratio
REST-Semantic
Search
Taxonomy
Entities
Sentiments
Classification
Statistical Relations
Labels
Event Knowledge Graph
Text
AllegroGraph KG
Platform
39. Taxonomy Tools
AI Labelers
Guided Advisors
Statistical
Relationships
Taxonomy BuildingEntity Extractor
AG Sentiment
Analyzer
Machine
Learning
Text
Classifier
CCOR Generator:
CoOccurence,
Correlation, Odds
Ratio
REST-Semantic
Search
Taxonomy
Entities
Sentiments
Classification
Statistical Relations
Labels
Event Knowledge Graph
Text
AllegroGraph KG
Platform
40. Taxonomy Tools
AI Labelers
Guided Advisors
Semantic
Exploration
“Search”
Taxonomy BuildingEntity Extractor
AG Sentiment
Analyzer
Machine
Learning
Text
Classifier
CCOR Generator:
CoOccurence,
Correlation, Odds
Ratio
REST-Semantic
Search
Taxonomy
Entities
Sentiments
Classification
Statistical Relations
Labels
Event Knowledge Graph
Text
AllegroGraph KG
Platform
Semantic
Exploration
41. Taxonomy Tools
AI Labelers
Guided Advisors
Applications
Taxonomy BuildingEntity Extractor
AG Sentiment
Analyzer
Machine
Learning
Text
Classifier
CCOR Generator:
CoOccurence,
Correlation, Odds
Ratio
REST-Semantic
Search
Taxonomy
Entities
Sentiments
Classification
Statistical Relations
Labels
Event Knowledge Graph
Text
AllegroGraph KG
Platform
Analytic Tools
Real-time
Decision support
Ad Hoc Queries
45. Conclusion
• If you want to understand your customer better
• Then really really really listen to him/her
• Put your understanding and learning in a knowledge graph
• And we are here to help!