Learn more about KPMG Artificial Intelligence

Intelligent Automation   |  Acceleration   |   Enhanced insights 

Unlocking the value of AI to accelerate strategies for Intelligent Automation and cost management, growth and customer engagement, and risk and regulatory policy



Business implications of AI

Cost Efficiency

Estimates suggest that a software robot is approximately 1/3 of the cost of an offshore FTE. When RPA is further enabled using AI programming approaches the benefits continue at an exponential rate

 

Quality/Reliability

Expect reductions in mistakes, accidents, regulatory violations and fraud. When trained properly, AI enabled processes will reduce risk and improve the quality of results

 

Productivity/ Performance

Artificial intelligence is “on” 24/7, 365 days a year; does not take vacations; and can perform tasks at digital speeds


 

Productivity/ Performance

Artificial intelligence is “on” 24/7, 365 days a year; does not take vacations; and can perform tasks at digital speeds



Consistency/Predictability

By design, machines are programmed to perform consistently, they do not bring personal bias into processing.  There is no ambiguity in “what created this insight”, as you can trace results in the method used to design the model 

 

Scalability

AI enabled solutions are designed to handle big data and evolving problems. There is also no overtime, or hiring challenges when your machine needs to handle more data or processing.

 






 

Where do I begin?

 

Grow your AI talent

Grow your AI talent, from technical level up to senior leadership. Simultaneously, nurture current process experts—those humans in the loop —to start training AI systems, and devote resources to increasing the AI literacy of your entire workforce. Organizations should have training programs to increase the skills of existing resources, as well as, teach leadership how to consume insights driven by complex algorithms they may not be able to see or fully understand. This may also include a Center of Excellence in the AI function dedicated to driving solutions, empowering teams to grow into a data driven organization.

Promote and enable a culture of innovation

Promote and enable a culture of innovation in which employees are expected to participate in AI transformation, and prioritize use cases for accelerated adoption. Get comfortable managing the portfolio in an environment with “known unknowns” and “unknown unknowns.” Create the portfolio so each project team learns from every experiment. Having more than one project under way will help you learn quickly, leverage learnings across projects, adapt and keep moving forward.


Test and learn

Test and learn. Know your high-value early use cases before you purchase technology. Consider “renting” capabilities – both technology and resources – before committing to a large expenditure. Start with proof of value rather than proof of technology. Resist the temptation to answer the question “can we do it” first; instead start with the question “should we do it”.  Expect a journey and that courage and leadership will be a critical element of progress and success

Allocate meaningful funding

Allocate meaningful funding and treat it as a long-term commitment measured in solutions successfully rolled out to production, not in weeks or months -  but longer term with ongoing enhancement and scaling. Partner with organizations that have a deep understanding of AI, not so that they can “do” AI for you, but so that they can accelerate your efforts. To succeed on this journey, enterprises should build a support network of firms that can help them get their journey under way, assist in the transformation, stay up to date on the latest innovations and work toward a goal of becoming self-reliant.






 

Key lessons learned


Don’t underestimate the power of good data

Sufficient volumes of quality data must exist to train models properly. Ensuring accessibility and availability of data can help scientists to build accurate solutions, or equally inhibit their ability to build trustworthy models.


Produce more with the same number of people

Use AI and automation to reduce employee administrative task load, freeing them to perform higher value tactical and strategic work. Equally important, use AI to drive insights and detect data  issues and opportunities too large for traditional approaches.


Artificial intelligence solutions are not plug and play

While many APIs and pre-built platforms are great accelerators, most solutions also require custom programming and training to attain target accuracy and results. Long-term, efficient models need to be well trained and improved over time.




Carefully select opportunities for deploying AI

Make sure the cost to implement is being balanced with expected ROI from day one. Prioritize back office computer-to-computer interactions use cases, particularly in IT, finance, and accounting as good places to start. These applications are currently driving out high value ad do not put your customers at risk while you are developing your AI acumen and early solutions.






Brad Fisher

Brad Fisher

Partner, Data & Analytics Lead, KPMG (US)

+1 212-909-5498
Traci Gusher-Thomas

Traci Gusher-Thomas

Principal, Data & Analytics and AI, KPMG (US)

+1 267-256-1617