Sorun Cevaplayalım

İşinizle ilgili öngörüler edinin, gerçek zamanlı bilgilere göre karar alın.

SAP Eğitim ve Sertifika Dönemleri

Uzmanlığınızı ve deneyiminizi SAP çözümleri kullanarak tasdik edin.

Çözüm Ekibi Başvurusu

Kullanıcılara hızlı ve pratik çözümler üreterek görev almak isteyenler.

SAP BLOG SAP Cash Application – How Accounts Receivable Benefits from Machine Learning #ASUG Webcast Recap

SAP Blog

Kayıtlı Üye
Katılım
22 Ara 2017
Mesajlar
990
Tepki puanı
6
This was an ASUG webcast from last week, from the Finance community. Below are my notes.

1cash.png

Figure 1 Source: SAP
Proactive marketing, more digital services
Services to apps, smart phones
New tax regulations, GDPR

2cash.png


Figure 2: Source: SAP
Finance transformation, machine learning, make suggestions, clear line items so Finance has more time such as Analytics

3cash.png

Figure 3: Source: SAP
“digital transformation”
Finance is part of the digital core, with pillars, more ideas
FP&A is EPM (BPC, planning)
Accounting & Financial Close – ERP, GRC, EPM – not just legal reporting, includes part of management reporting (profit center)
Finance Operations – accounts payable, receivable, expense reporting, real estate
Treasurey management – bank communications, statements, cash management, working capital, hedging, investments
Enterprise Risk & Compliance – GRC, access management, segregation of duties
Cybersecurity & data protection – new topics

4cash.png


Figure 4: Source: SAP
Leonardo – cloud deployments, user experiences, robotics machine learning, blockchain, new technology – tools for finance to do their job more effectively

Allow finance to automate processes so spend more time doing analytics

Make more strategic decisions

5cash.png


Figure 5: Source: SAP
Financial close – do it end to end, take individual processes, former batch job, GR/IR, cash application – batch jobs – automate them, more accessible without having to wait overnight

Fraud prevention, detect duplicate invoices

Context sensitive help – co-pilot “siri for business” – ask business for context type situations

Predict future values – look at trends, model what if scenarios, profitability aspect – predict profitability by customer, product line

6cash.png


Figure 6: Source: SAP

Machine learning, AI – rule engines, what we have in the past, configure rules, batch processes look at rules, determine reconciliation

Issue with the rules – rare for company to go back and reconfigure them

Robotics process automation – running a macro “over and over” – not revisited often if still valid

Middle of Figure 6, machine learning looks at the patterns, looks at the exceptions, and process of exceptions

Machine learning learns from actions of finance team; not need to reconfigure the rules

7cash.png

Figure 7: Source: SAP
Real time information from SAP HANA; not wait for BW
Finance can review analytics in real time

8cash.png


Figure 8: Source: SAP
CFO portfolio – more machine learning aspects

GR/IR coming out with machine learning

9cash.png

Figure 9: Source: SAP
KPI – improve days sales outstanding, able to clear items more quickly

Instead of going through exceptions manually, automate machine learning

Less manual time with machine learning

Central Finance is an implementation of S/4HANA finance, allows consolidation of processes

10cash.png

Figure 10: Source: SAP

Why are rules less effective over time? Often configuration rules are not reviewed

Exceptions – credit blocked sales orders, see how professional handles, and machine learning will learn from those decisions

11cash.png

Figure 11: Source: SAP

Customer pays for invoices, could be for 1 invoice or multiple invoices with one transaction; if multiple, how match? May only have a 30% match rate, everything else manual (missing information)

Could have exchange rate differences

Customer could have called, manual information document, system may not understand (unstructured information)

Finance needs to deal with this on an exception basis, takes time

12cash.png

Figure 12: Source: SAP
Cash application looks at your history; needs at least 5K records; sees how Finance team member executed on the exceptions, looks at the matching criteria and learn from it

It looks at matching proposals

Cash application – you can decide if you want it to automatically clear or give you a proposal; configure confidence level

Benefits to finance – less time for finance, better DSO KPI; do not need to go back to review rules

Based on SAP Cloud Platform – runs on cloud or on-premise

13cash.png


Figure 13: Source: SAP
Customer sends document, take that the information to process cash application
“intelligent matching”

Look at the documents, and involved in clearing processes

14cash.png


Figure 14: Source: SAP

Runs in cloud or on premise, same application

Hybrid architecture

15cash.png


Figure 15: Source: SAP

Customers can logon, see their payments, see them online, integrate with credit history

16cash.png


Figure 16: Source: SAP

Now have lockbox; available since 1805 S/4HANA

17cash.png


Figure 17: Source: SAP
Releases supported

18cash.png


Figure 18: Source: SAP
5K of the documents above to train machine learning
Take historical documents, run through machine learning engine
50 matching criteria as of now (customer number, PO number, actual value)
Configure tolerance, confidence

19cash.png

Figure 19: Source: SAP
Finance clears manually, creates model, recommends clearing, not clear, output will have probability % of accurate match

20cash.png

Figure 20: Source: SAP
Probability of matching; can review reports

21cash.png


Figure 21: Source: SAP

Steps to cash application

Step 3 is to run cash application job, it will return proposals – clear automatically, and if not in tolerance, not clear

Happens in real time, based on HANA

Order of steps to happen

22cash.png


Figure 22: Source: SAP
Cash comes in, match with payment advice, clear based on tolerance level

23cash.png


Figure 23: Source: SAP
Requires S/4HANA; looking at connectors to backport
Automation at scale
Coming out with more apps to leverage machine learning

Question & Answer


Q:Backend has to be S/4HANA?
A: Yes, looking at backporting to ECC, no schedule for that
Q: Does cash app cloud service look at our system?
A: Provides recommendations, part of configuration to auto clear
75% confidence level is typical
Q: Do we have a risk from losing data?
A: Still look at what is cleared and why; not losing data
Q: Is this solution used by consumer good companies?
A: Working with co-innovation customers, it is cross industry now
Q: How long does it take for machine learning to learn the model?
A: Learning is fast; a few days



I would be interested in this application if it also applied to SAP FI-CA. What do you think?

Okumaya devam et...
 
Üst