Artificial Intelligence -- Potential and Implications for Finance

Timo Elliott, Innovation Evangelist

Artificial Intelligence: The Potential and Implications for Finance Leaders

3

Artificial intelligence

AI is a “sociotechnical construct” indicating machine capabilities which solve complex tasks that were

recently only possible by humans (equally well or better)

Technical disciplines that

solve business problems

through the extraction of

knowledge from data.

Deep learning

Machine learning

Data Science

Big Data

Advanced

Analytics

Predictive

Analytics

Data

mining

4

Why now?

Big DataAlgorithms Hardware

5

Algorithms are taking over

Voice transcription Lip reading Image Descriptions

6

What’s different with machine learning?

Classical software must

be programmed

Humans learn from

experience

Machine learning learns

from data

7

Data Training Inference

Apply

model

Services(such as invoice processing,

profile matching)

…and more

Applications (such as cash application)

Text

Image

Video

Speech

… and more

Train

model

Prepare

data

Capture

feedback

How does machine learning work?From data to insights

8

In the future, all processes will be “self-improving”

Do your financial processes

improve automatically over time?

9

40% 70% 94%

10

Machine learning everywhere

Source; McKinsey & Company

Machine learning has

broad potential across

industries and use cases

http://www.mckinsey.com/global-themes/digital-disruption/whats-now-and-next-in-analytics-ai-and-automation

11

Automating end-to-end processesIncrease efficiency and reduce costs

Detect and prevent Detect and rank information out of Big Data

PredictDerive knowledge from historical information to

increase the accuracy of predictive scenarios

Proactive context-sensitive supportDigital assistants boost productivity of financial experts

Machine learning for financeChanging the world of finance by adding intelligence to applications

Automate

PredictDetect

Assist

12

Big opportunities

Source: http://gartner.com/SmarterWithGartner

“In 2018, half a billion users will save two hours a day

thanks to AI-powered tools.” Gartner

13

Finance of the futureInverting the effort pyramid through automation

Strategy

and growth

Keeping the

lights on

Compliance, efficiency,

and business performance

Strategy and

growth

Keeping the

lights on

Today 2020

Automation

“Up to 70% of finance tasks are potentially automatable

with next-generation technologies”(McKinsey 2016)

14

Example: accounts receivable teams

Joni Chen

AR accountant

Lump sum

Missing infoDiscounts

Exchange

Customer

call

???

Payments

15

Intelligent Invoice Matching

History

Payments

Invoices

Matching proposals

Intelligently learn matching criteria from your history and

automatically clear payments.

Machine Learning

Improves days sales

outstanding

Allows shared services to

scale as the business

grows

Empowers finance to

focus on strategic tasks

and service quality

16

Automatic extraction of information from invoices

17

Companies with 10,000 employees

and more process 300,000+ invoices

every year

A subset of these needs to be corrected,

because of formal or content errors

Historical data is used to identify

patterns for invoices that need

corrections

These patterns are used to warn the

approver proactively

Intelligent invoice correctionProactively improve invoice accuracy

18

Automatic extraction of information from employee expenses

19

Record

to Report

Procure

to Pay

Order

to Cash

Eliminate manual FI/CO reconciliation efforts and make business self-service access

to data easier and more intuitive

Improve forecast accuracy, accelerate and automate the financial close along with lower

compliance and auditing costs

Significantly simplify and automate the interaction between buyers and suppliers through

digitizing the exchange of information and achieve high invoice automation rates

Automate core Accounts Receivable processes like Credit Management, Dispute

Management and Cash Application

Plan to

Forecast

Using AI to help automate End-to-End processesIncrease efficiency and reduce costs

20







Automate

Predict

Assist

Detect

21

Fraud investigators can detect unknown fraud patterns and reduce false

positives leveraging their company’s investigative history – without expert

knowledge in data science and algorithm tuning.

Detection of new fraud patterns

Reduction of false positives

Predictive detection methods allow your business analysts and fraud

investigators to

Automatically detect and rank attributes within classified data that positively

correlate with fraudulent cases;

Incorporate them with existing detection methods into new fraud management

strategies.

Business integrity screeningDetect and rank information that positively correlates with fraud

Detect Fraud

Reduce false positives

Save money

Fraud Management Team

22

Assist

Detect







Automate

Predict

23

Augmenting financial analysis with artificial intelligence Bringing together actuals, forecast and simulation to uncover market trends before they happen

Cleanse and

match data from

different sources

Spot outliers, perform

forecasts, and

determine causality

Share and

operationalize data

more intelligently

Run what-if analysis

and zero-in on key

influencers for any

business challenge

24

Predictive accountingAccelerate accounting processes by machine learning powered forecasting

Create a common view

of all financial &

operational data

Predicts key

accounting KPIs & how

they impact key metrics

Provides easily

consumable reporting

Anticipates the impact

of currency fluctuations

for the end of a period

Optimize processes to close the financial books at the end of every fiscal period, and

reduce costs, cycle times, and error rates.

Data of

former

period end

positions

ML used to forecast

costs, revenues &

recurring costsMachine Learning

25

Constantly-updated predictive key performance indicators

Group

Prediction

26

Predictive accounting vision

27

DetectPredict







Automate

Assist

28

Digital enterprise assistants and self-service finance

Business context awarenessUnderstanding the business context, and pro-actively

suggesting solutions using predictive functionality

Conversational user interfaceConversational interface that uses natural language processing

functionality to create a human-like experience

Cross applicationsAllows seamless transition across platforms; start a task on a

mobile device and continue later, on a desktop or vice versa

Self learning Using machine learning functionality to gain knowledge based

on historic data, experience, and take action in response to new

or unforeseen events“By 2020, the average person

will have more conversations

with bots than with their spouse”

Gartner

29

Your data is your most important asset

Investments in systems to ensure high-quality, consistent data will be even more amply rewarded

Invest in gathering leading “signal” data (e.g. social sentiment data, or foot-traffic statistics) to

augmented “lagging” finance data

It’s about people

Automation is a huge opportunity, but it’s also about “augmented intelligence” – displacing work, not

replacing workers

Watch out for side-effects

You’re delegating more decision-making to machines: make sure there’s still oversight and auditing

New ways of working

Implementing AI in new ways requires new skills – your strategic partners can help!

Implementing AI: requirements for success

Thank you.

Contact information:

Timo Elliott

VP, Innovation Evangelist

SAP

timo.elliott@sap

@timoelliott