Intelligent Automation and AI: Deliver What Matters To Your Procurement Function!
Intelligent Automation and AI: Deliver What Matters To Your Procurement Function!
Intelligent Automation and AI: Deliver What Matters To Your Procurement Function!
and AI
Deliver what matters to your
procurement function!
Srivatsan Abhishek
B Tammina
Solution Consultant, Solution Consultant,
Nia AssistEdge
Internal
Setting the context…
Shifting gears in
procurement…
ROBOTIC ARTIFICIAL
PROCESS INTELLIGENCE
AUTOMATION
How far along are you with AI/Automation plans for your procurement function?
• I want to get started immediately
• Trailing and Initial stages(PoC)
• Production
• Scaling
• Not yet
• Start immediately
• Near term
•
Digitization
especially renewals
Future proof
Precise table technology stack
Contract • Contract review and risk analysis extraction of
Risk & automation any kind
Clause • Integration and support for User Friendly
Variance contracting playbooks and Easy to Use
Location attribute
HEADER
• What Clauses/Terms exist in my existing legacy contracts • How do we identify clause variances in 3rd party contracts
• How can we create a Master Clause/Term Repository, with • How can we search for specific Clauses/Terms in specific
Fallbacks contracts
• How do we assess Risk in contracts • How do we identify what types of contracts to migrate ?
Motivation Solution
One of the largest multi-national conglomerate Infosys Nia Legal solution ingested vendor contracts
company headquartered in Japan deals with spanning various topics and legal entities to create a
Benefits tons of vendor contracts (buyer suppliers) and neural network based knowledge model which in
spend considerable effort and time in reviewing turn enabled identifying and comparing the
these contract documents. Their value chain variations in wording for these topics, across
Reduced legal costs and manual efforts integration division spends months of time and vendors (buyers, suppliers) for all the associated
effort dealing with these contracts throughout contracts. Also based on the entities extracted a risk
the year right from drafting stage to identification solution is developed and deployed. A
negotiations to approval to reposition. They end user interface summarizing the entire contract
Improved Staff Productivity
need to summarize these contracts by extracting as well as providing the capability to compare 2 or
business critical information like obligations, more contracts was provided saving the team
warranty, compliance, timeframes and so on as significant effort and money.
Early Risk identification and mitigation well as identify risks in each of the contract
Contract
Risk
Benefits &
Compliance
Policies
Procurement
Enterprise
Legal
Evolution of AI based contract analysis in the enterprise – Use cases
Large US Financial Services Provider
Nia CA ingested tens of thousands of Procurement contracts to analyze clauses and terms for a
variety of risks including Legal, Business, Financial, Compliance etc. Nia also provided analytics on
risky contracts, clauses and terms. The organization could then focus on contracts with highest
risks. Nia CA saved enormous amount of money and time to provide accurate risk analysis
Contract
Risk
Benefits &
Compliance
Policies
Procurement
Enterprise
Legal
Evolution of AI based contract analysis in the enterprise – Use cases
Large US Financial Services Provider
Nia CA ingested tens of thousands of Procurement contracts to analyze clauses and terms for a
variety of risks including Legal, Business, Financial, Compliance etc. Nia also provided analytics on
risky contracts, clauses and terms. The organization could then focus on contracts with highest
risks. Nia CA saved enormous amount of money and time to provide accurate risk analysis
Contract
Risk Large US Railroad operator
Nia CA ingested all labor contract agreements to create a
knowledge model that captured rules, exceptions and
Benefits & resolutions. The model responded to incoming queries and
Compliance
Policies provided resolution to exceptions based on skills, availability,
compliance rules etc. captured from the labor contracts. A
conversational interface provided responses to natural
language queries
Procurement
Enterprise
Legal
Evolution of AI based contract analysis in the enterprise – Use cases
Large US Financial Services Provider
Nia CA ingested tens of thousands of Procurement contracts to analyze clauses and terms for a
variety of risks including Legal, Business, Financial, Compliance etc. Nia also provided analytics on
risky contracts, clauses and terms. The organization could then focus on contracts with highest
risks. Nia CA saved enormous amount of money and time to provide accurate risk analysis
Contract
Risk Large US Railroad operator
Nia CA ingested all labor contract agreements to create a
knowledge model that captured rules, exceptions and
Benefits & resolutions. The model responded to incoming queries and
Compliance
Policies provided resolution to exceptions based on skills, availability,
compliance rules etc. captured from the labor contracts. A
conversational interface provided responses to natural
language queries
Procurement
Enterprise
Large US Telecom provider
Nia CA ingested contracts that had complex billing terms for
a variety of services, and provided configuration rules for
Real estate Revenues Billing system. Nia CA used a 100+ layer Deep Learning
network to analyze document structure, extract tables and
associate billing terms. This resulted in faster onboarding of
clients and more accurate billing
Legal
Evolution of AI based contract analysis in the enterprise – Use cases
Large US Financial Services Provider
Nia CA ingested tens of thousands of Procurement contracts to analyze clauses and terms for a
variety of risks including Legal, Business, Financial, Compliance etc. Nia also provided analytics on
risky contracts, clauses and terms. The organization could then focus on contracts with highest
risks. Nia CA saved enormous amount of money and time to provide accurate risk analysis
Contract
Risk Large US Railroad operator
Nia CA ingested all labor contract agreements to create a
knowledge model that captured rules, exceptions and
Benefits & resolutions. The model responded to incoming queries and
Compliance
Policies provided resolution to exceptions based on skills, availability,
compliance rules etc. captured from the labor contracts. A
conversational interface provided responses to natural
language queries
Procurement
Enterprise
Large US Telecom provider
Nia CA ingested contracts that had complex billing terms for
a variety of services, and provided configuration rules for
Real estate Revenues Billing system. Nia CA used a 100+ layer Deep Learning
network to analyze document structure, extract tables and
associate billing terms. This resulted in faster onboarding of
clients and more accurate billing
Legal
Contract
Risk Large US Railroad operator
Nia CA ingested all labor contract agreements to create a
knowledge model that captured rules, exceptions and
Benefits & resolutions. The model responded to incoming queries and
Compliance
Policies provided resolution to exceptions based on skills, availability,
compliance rules etc. captured from the labor contracts. A
conversational interface provided responses to natural
language queries
Procurement
Enterprise
Large US Telecom provider
Nia CA ingested contracts that had complex billing terms for
Large US Telecom provider
a variety of services, and provided configuration rules for
Nia CA ingested tens of thousands of real estate contracts
connected to assets in the Wireless business for compliance
Real estate Revenues Billing system. Nia CA used a 100+ layer Deep Learning
network to analyze document structure, extract tables and
and minimize revenue leakage
associate billing terms. This resulted in faster onboarding of
clients and more accurate billing
Legal
Contract
Risk Large US Railroad operator
Large US Health Insurer Nia CA ingested all labor contract agreements to create a
Nia CA ingested all Health Plan documents and provided a knowledge model that captured rules, exceptions and
natural language benefits search capability so clients can get Benefits & resolutions. The model responded to incoming queries and
Compliance
consistent and accurate responses to their benefits queries Policies provided resolution to exceptions based on skills, availability,
regarding In/out of network benefits, Co-payment and co- compliance rules etc. captured from the labor contracts. A
insurance and other similar benefits conversational interface provided responses to natural
language queries
Procurement
Enterprise
Large US Telecom provider
Nia CA ingested contracts that had complex billing terms for
Large US Telecom provider
a variety of services, and provided configuration rules for
Nia CA ingested tens of thousands of real estate contracts
connected to assets in the Wireless business for compliance
Real estate Revenues Billing system. Nia CA used a 100+ layer Deep Learning
network to analyze document structure, extract tables and
and minimize revenue leakage
associate billing terms. This resulted in faster onboarding of
clients and more accurate billing
Legal
ROBOTIC ARTIFICIAL
PROCESS INTELLIGENCE
AUTOMATION
Automation
Knowledge Maturity
Rule Based Data Driven
driven
Process Decision
maintenance
Automation Making
& self-learn
Sophistication
Document Approval
Handling
digitization workflows
Unstructured
and
Data
extraction
Reporting,
ERP
Control and
integrations
Audit
Goods
Order placed delivered
1 3
2 4
Purchase Delivery
Retail Business with order issued Ordering goods from Receipt Goods are supplied to
turnover of >$6B 2000 vendors Received over 60 warehouses
5
Vendors submit 3 documents
Purchase order + Delivery Receipt + Invoice
Tax Invoice, BoS, Receipt Delivery Receipt, PO, Challan, Damage Notes, Bills
Challenges:
• Most digital copies had multiple documents included in one copy
• Scanned documents making data extraction a challenge
• Handwritten documents which affects automation %
Human-Intervention Challenges
Challenges
• Very minimal success % in end to end invoice processing
• Lot of manual interaction required leading to resource
delays and human errors
Recommendation
Threshold % of 100%
Perform Process
confidence
RPA, Apply Automation
rules Learn from
decisions made
by experts Conditional
Decision Making
SME or Domain
expert involvement
Our Solution
Data Extraction
from Complex docs
Using AssistEdge
Computer Vision &
Advanced OCR Process related
efficiencies
100% established &
Process savings
Conditional Automation delivered
Decision Making
Using AssistEdge
Decision Workbench
Our client was able to automate this process end to end using
AssistEdge, including data extraction and conditional decision making
Data Extraction
from Complex docs
Using AssistEdge
Computer Vision &
Advanced OCR Process related
efficiencies
100% established &
Process savings
Conditional Automation delivered
Decision Making
Using AssistEdge
Decision Workbench
Our client was able to automate this process end to end using
AssistEdge, including data extraction and conditional decision making
AssistEdge
AssistEdge
Decision Workbench
learns from
Remaining were routed to
decision
domain experts with
recommendations from
AssistEdge
With AssistEdge Decision Model,
With existing model only 18%
70% invoices were processed
invoices were processed
without human intervention
Key KPIs improved
3 days
Invoice processing time
15 days
$1
Invoice processing cost
$5
Internal
Appendix
Contextual
Enables
To identify process Insights for Leverages cognitive
Cognitive
redesign opportunities Albie services on tap
Services
Workbench
Top CXOs on what will have greatest impact on their business in next 12 months
Customers feel RPA has Customers actually see an Customers feel with RPA they
provided agility in improvement in CX are well placed to scale
responding to customers their business
In a recent study, Enterprises who have begun their RPA journey have seen
improvements in form of Better CX and Agility in responding to Business
Increase Automation GDPR ready Albie, the cognitive engine 18+ Patent Assets
Footprint & counting
Automate applications AssistEdge RPA is now Albie observes, learns and Around Bot orchestration,
in Citrix, VDI GDPR ready helps solve business problems performance prediction & AI
Automate in your language Integrated with Smart Resource Innovation around Exception
of choice CyberArk Management Management
Multi-lingual support for e.g. AssistEdge is integrated with Predictive alerts around bot Intelligence to learn from
Japanese Spanish, French CyberArk secure Credential performance & SLA failures exceptions & automate them
Vault
Dynamically scale Bots as per OpenID and Albie Decision Pre and Post processing
business needs SAML compliant Workbench Cognitive services
Scale no. of bots dynamically AssistEdge enables secure Seamlessly unifies Human- Working with cognitive services
for increased bot reliability authentication with SAML, OIDC Digital worker to improve their outcome