Thank you for Subscribing to CIO Applications Europe Weekly Brief
Artificial Intelligence in Coherence with Users' Expectations
By Doron Reuter, Chief Product Owner, ING Wholesale Banking Advanced Analytics
Exhaustive efforts in cognitive disciplines such as machine learning have helped empower today’s gadgets with the capabilities to arrive at decisions which humans would have difficulty rationally optimizing in their heads. Thus, AI-powered machines can be employed for complex, abstract strategy games such as Go and they can effortlessly defeat even the best human players. These interesting examples best prove the ability of AI in influencing the world drastically in the days to come. The psyche of the average human, in general, today is centered around how most tasks can be accomplished with minimal effort and time. Thus, AI is here to stay and influence human lives for a long time. There will of course be decisions and actions at which humans excel in comparison to machines. Probably those decisions that need a emotional interaction for e.g. psychotherapy or sales, or thought processes that require a high level of creativity and intellectual flexibility e.g. breakthrough research. Probably those actions that require a certain human finesse for e.g. ballet dancing
A futuristic possibility across most retail outlets is walking in and out of the store with a basket full of products you need without the need to take a basket, lift a product from the shelf, scan each product, pack each product and take out your wallet to pay. The future Supermarket experience will be an immersive personlised experience as you walk around the store and see specials, recipes, predictions of what you might need and what is cheap now on your phone and click. You pick up an auto filled, scanned and paid for bag at the exit. Eliminating the need for attending humans, improving the experience and limiting waste.
The genius of AI is to leverage the benefits of massive amounts of new types of data, powerful & cheap computational power and machine learning into algorithms that can within milliseconds predict in a given situation what the next best action or decision is without having ever seen that exact scenario before
Artificial Intelligence in Businesses
Apart from making applications such as games more enticing, AI is already playing a pivotal role in empowering businesses. The best instance is AI-powered automated telephonic customer service today that assists clients based on their location, sparing them from the exhaustive exercise of traversing a long tree of options, a huge leap compared to the 90s. Enterprises should ideally accommodate AI practices that suit the tastes of clients and are also aligned with their own organizational goals. The amalgamation of both perspectives is the key for firms looking to propel by exploiting the benefits of Artificial Intelligence. Consolidation of the myriad best AI practices suggested by industry veterans and application of the ideal combination is an activity decision makers should do to steer ahead of industry peers.
The best practice to adopt any technology discipline ideally is to prioritize the user, respect the privacy of his data and not seem to be intrusive. Virtualization can also prove to be a recipe to destroy the trust of clients if not practiced ethically. Many enterprises try to woo users by offering AI-powered personalization, which is usually a double-edged sword. While some users might be enticed by personalization, a few others might fret over breach of private data. For instance, if the target audience thinks it’s undesirable for users to get unnecessary notifications that offer little benefits, the purpose of utilizing cutting edge disciplines such as AI is defeated.
If you first try to organize and structure all your data before developing any AI then in five years time you will be behind the curve. Instead, enterprises should ideally develop a funnel of specific high impact use cases and then bring together the data for those use cases and start.. One of the biggest challenges plaguing most enterprises today is which AI cases to prioritize because there are so many use case options and AI is a broad topic with many different underlying technologies and methodologies. However, prioritization and focus are key to getting momentum as high impact enterprise use cases take time and resources to start.. Organizations should ideally zero-in on a selection criteria that are most effective for their company.
For example, if you are a bank and you create value primarily by running prudent credit processes requiring a high level of analyst work using a variety of unstructured data sources like rating reports, annual reports, industry reports, financial reports etc. then use cases that use OCR technology to extract data points and Natural Language Processing to understanding phrases in the documents will be use cases of high impact and your selection criteria should reflect this. However, if you are am Ecommerce player who sells and distributes goods bought online from a network of warehouses then use cases of high impact will likely be online product recommenders and other marketing cases that increase conversion or supply chain optimization cases. A cognitive application that works wonders for one enterprise, won’t necessarily improve the prospects of another.
Tailored to the Audience’s Psyche
The adoption of AI needs to be aligned with the needs of an organization’s audience for it to be relevant in the future. Most importantly, the mindset of today’s millenials is an area most decision makers need to inspect and accordingly adopt ideal AI practices. Students today, for instance, exchange and utilize information in a completely different manner as compared to a decade ago where most pupils relied on pen and paper. Thus, the expectations of the possibilities from technology are bound to mount steadily in the days to come. Organizations shouldn’t merely devise AI solutions that would suffice for the current but also justifies the broader picture of keeping pace with rapid technological evolution and customer demands.