How Artificial Intelligence is Transforming the Financial EcosystemMFM Localism EventMarch 2019
Copyright © 2019 Deloitte Development LLC. All rights reserved. 1
Baris [email protected]
Tech Strategy
Introductions
Dennis [email protected]
Strategy and Analytics
Copyright © 2019 Deloitte Development LLC. All rights reserved. 2
1900s
Tabulating System EraMechanical devices helped us with tasks such as organizing data and making calculations for tasks such as evaluating the sales performances of a company’s employees (Hollerith Tabulating Machine).
1950s
Programmable Systems EraProgrammable systems enabled major accomplishments such as space exploration and the development of the Internet. The programmable systems era will continue indefinitely, and will underpin the era of computing that we are in now (Windows Desktops/ERP systems).
2000s
CognitiveEraCognitive capabilities augment and amplify human intelligence. It is critical to recognize and know when and how to automate and augment to achieve the cognitive advantage.
2020s
Amplified Intelligence EraIn the future, systems will be able to augment human intelligence and replicate human interactions.
Welcome to the Cognitive Era
Copyright © 2019 Deloitte Development LLC. All rights reserved. 3
Exponential Growth of DataIncreasing access to multi-source, multi-structured data is creating all new capabilities, inspiring innovation, and generating completely new business models
Smarter AlgorithmsAdvanced algorithm development and machine learning techniques are enhancing the way automated reasoning occurs
Reduced Computing CostsThe cost of computing has been slashed thanks to technologies such as cloud computing and virtualization that lower infrastructure costs
Faster Processing SpeedsImproved tools and frameworks for handling and managing larger datasets accelerate capabilities and reduce time-to-data
Advances in key disciplines led to the prominence of Artificial Intelligence (AI)
Copyright © 2019 Deloitte Development LLC. All rights reserved. 4
20151970 2000 2020Data storage to review prior performance Data analytics AI
Ad
ded
Val
ue
Hindsight
Historical reporting on key performance
indicators
DataCapture
DesktopcomputingMainframes
Manual
ERP
Foresight
New technologies (Cloud, Robotics, AI, ML,
NLP, Blockchain)
Cloud
In-memory Computing
Visualization Process Robotics
AdvancedAnalytics
CognitiveComputing
Blockchain
Digital
Insight
Reporting and analytic software
Profitabilityanalysis
Mobiledevices
EPM
Performancetracking
New technologies have driven the development of Finance for decades
Copyright © 2019 Deloitte Development LLC. All rights reserved. 5
Finance transformations present the opportunity to make smarter decisions, at a faster pace
The future disruptors within finance enhance the ability of finance organizations to process transactions, scale, and provide strategic decision support. The key question for finance organizations is how best to deploy the extra capacity created by the these disruptors.
HOW HUMANS IN FINANCE NEED TO RESPOND
Talent Model
• Transactional processes that rely on large, disconnected teams today will be automated.
• Finance teams will require a high degree of business acumen and ease with analytic models.
Organization Structure
• Transition from large shared services to smaller, flexible tiger teams.
• Structure as a more matrixed organization with constant ties to the business.
Service Delivery Model
• Adapt finance service delivery to be more nimble and lead from the front
• Leverage cognitive technologies to quickly quantify dollar value and deliver financial impact for business decisions
Copyright © 2019 Deloitte Development LLC. All rights reserved. 6
Structured DataDeterministic Outcomes
Unstructured DataProbabilistic Outcomes
Robotic Process Automation Cognitive Automation
“Mimics Human Actions”
Potential RPA Applications:• Back Office – “Swivel
chair” clerical tasks• Supply Chain –
Cross-system verification and coordination
• Customer Service –Mass customization of automated CRM
RPA Technologies:• Robotic Process
Automation• Rules engines• Event stream / complex
event processing• Human-in-the-loop
process automation
RPA Realm:• Rules-based tasks• Operational processes
“Augments Human Intelligence”
Potential Cognitive Applications:• Supply Chain – Predictive
analytics in inventory management
• Customer Support –Algorithms to optimize supply logistics networks
• Manufacturing –Computer vision systems for inspection and quality assurance
Cognitive Technologies:• Optical Character
Recog• Natural language
Processing• Machine learning• Integrated cognitive
computing platforms
Cognitive Realm:• Cognitive analytics• Decision making
Through AI, companies can overcome the labor and scaling challenges arising from complex and ever changing technology, compliance and process landscape
Copyright © 2019 Deloitte Development LLC. All rights reserved. 7
An automated robot begins populating the net revenue report and output to its analytics engine
5:00 AM
The data from the analytics engine is staged on a tableau or BI server and dashboards are refreshed
6:00 AM
The CFO has access to the updated reports and narratives as soon as they arrive in the office
The CFO further drills down into cost variance analyses for products/services with dropping margins
The natural language generator adds in narratives to the dashboard
7:00am
8:00am
The CFO interacts with the chatbot in natural language to drill down into splits by channel/product
8:10am
8:15am
In less than 5 years, leading media Finance organizations could look very different
All of the technologies required to realize this vision of the future exist today
Copyright © 2019 Deloitte Development LLC. All rights reserved. 8
Robotic Process Automation Video
Copyright © 2019 Deloitte Development LLC. All rights reserved. 9
Cognitive Automation Video
Copyright © 2019 Deloitte Development LLC. All rights reserved. 10
Speed Increase
Cost Reduction
Quality
Talent
Scalability24/7 Operations
Decision Making
Internal Control
AI Benefits
Reduce turn-around time
Reduce cost through automating activities and requiring fewer FTEs
Increase quality by avoiding human error
Boost employee engagement by shifting agents to more interesting tasks
Increased capacity without long build-up phase
Non-stop performance
Improve executive decision making
Reduce potential for human fraud
Reduces time to action based on accuracy and confidence in data available
Ensuring consistency / accuracy of data in reporting by eliminating manual errors by 80% to 99%
Eliminates manual intervention and reduces employees needed to execute task by 20% to 60%
Reduces average time to execute transactional processes by 60% to 80%
Decreases processing time by up to 300%with overnight / weekend processing
Build automated system interfaces without investment in IT architecture by 20% to 50%
Shift FTE focus from report generation to analysis by 30% to 60%
Reduce time for vendor fraud detection from days to minutes
AI can quickly provide financial and non-financial benefits that impact the most common organizational performance measures
Copyright © 2019 Deloitte Development LLC. All rights reserved. 11
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
of respondents report positive ROI from AI efforts
Median ROI from AI efforts17%
AI investment and ROI: Relative landscape of industries
Source: State of AI in the Enterprise, 2nd Edition, Deloitte
AI investments pay off—some more than othersIt’s raining returns
Copyright © 2019 Deloitte Development LLC. All rights reserved. 12
Achieving financial benefit
85%see a positive ROI from AI investment
Strategically important
76%say AI technologies are very or critically important to business success today
Facing challenges
42%cite implementation as a top three challenge; data and integrationwere also highly ranked
Source: Deloitte State of AI in the Enterprise, 2nd Edition, 2018
High level of enthusiasm for AI in TMT, but success can depend on execution
Copyright © 2019 Deloitte Development LLC. All rights reserved. 13
Deep learning
54%+19 points from 2017,
the largest increase
Machine learning
61%+4 points from 2017
Natural language processing
66%+10 points from 2017
Source: Deloitte State of AI in the Enterprise, 2nd Edition, 2018
Early adopters are creating strong AI foundations
Copyright © 2019 Deloitte Development LLC. All rights reserved. 14
ChallengeHow to achieve “best-in-class” in Finance in 3 years?
• Finance factory• Business Insights and Service• Real-time • Self-service• Talent
Putting it all together
SolutionA Digital Finance program with 4 key pillars:
• RPA/Cognitive Automation• Blockchain• Predictive analytics and visualization• Analytics CoE
Impact
Analytics COE established300K hours in Bill to Cash process automation70K delivered in the 1st year
52K hours savings, 100% NBA Compliance, AR improvement thru
automated emails and call scripts for collections
One-click, automated executive reports with
human sounding commentary
Other: Promotion effectiveness, Blockchain for settlements, CS chatbots, NLP for contracts