AI decision automation: Where it works, and where it doesn't work

1640629881 AI decision automation Where it works and where it doesnt

Some companies use AI for end-to-end decisions, but not all decisions can be made without human intervention. Here are some real-world issues.

Image: iStock / Blue Planet Studio

As artificial intelligence (AI) ascends to the market, the question remains as to how far to be trusted when it comes to the "last mile," the final decision of which follow the analyzes and recommendations provided by AI.

In medicine, AI and crunch analysis through data reviews and scientific analysis to come up with a series of recommendations for difficult diagnosis, but the final decision will be made by the experienced medical professional.

SEE: TechRepublic Premium editorial calendar: IT policies, checklists, tools, and search for downloads (Premium TechRepublic)

In the loan approval process, automated decision-making software reviews third-party application and data to confirm a loan decision, but this is done by the subscriber or the loan supervisor. the final decision.

“Not all decisions in organizations can be completely automated, and some of these require human intervention,” said Arash Aghlara, CEO of Flexrule, which produces automation software. conclusions. "Decision-making automation should allow situations where fully automated decisions are not possible due to doubts, uncertainties and so on regarding the decisions. Instead, these require the intervention and intervention of land experts.


    Where AI can make decisions for an organization

    However, if a company wants to use an end-to-end automated AI decision-making process, there are areas of operation where AI can work. These are areas where the ironclad rule is established that groups are comfortable with them and have no opportunity for exceptions.


    • In your internal budgeting and budget management processes, you can automate the levels for line item approval based on dollar sum. Software can encourage you when amounts exceed $ 50,000 or $ 100,000 and must be approved by the CFO or the CEO.
    • In sales promotion and performance, based on dollar amount and past customer support, automated product discounts can be provided.
    • A piece of factory equipment can be custom designed based on performance and environmental requirements in a specific facility without the need for intervention salesman or sales support engineer or engineer.

    SEE: How AI can help encourage the use of renewable energy (TechRepublic)

    READ  Here's how the 2022 tax brackets work

    All of these examples have one common thread: They work on a very detailed logic that the industry applies and has little chance of varying from one case to the next. In these ironclad cases, it is possible to implement end-to-end automation without human intervention.

    Taking automation a step further

    There are, of course, companies that want to push the envelope and have made a strategic commitment to go ahead with end - to - end automation despite the fact that automation may be wrong at times.

    An example is Bestow, a life insurance company that made the decision to go end-to-end AI decisions on life insurance, without any human intervention.

    “By leaving the decision to the AI ​​and its algorithms, we believe we have solved one of the key challenges when human subscribers make those decisions, "said Ben Hsieh, director of product development. "In human decisions, there can be inconsistency and bias."

    SEE: Can AI replace human decisions? Most companies say no, but it can help (TechRepublic)

    Hsieh understands that there are times when algorithms and data models need to be updated and when human intervention would be nice, but the distance to the company's automated subscription market and customer satisfaction levels. making that risk worth it.

    So what does all this tell companies in general? That you can make a strategic decision to go “all AI” into your decision, as long as you understand the risks and that it is worth taking the risk.

    For most companies, however, a hybrid approach to machine “last mile” decisions works best.

    "We use machine learning to automatically decide who should make the pricing decision - the vendor or the model," said Yael Karlinsky-Shichor, professor of marketing at Northeastern University . "What we are finding is that there is a hybrid structure that allows the model to price most of the values ​​that come into the company but that allows the experienced retailer those issues that are more specific or out of the ordinary can be achieved even better. "

    See also

    Related Posts

    Deja una respuesta

    Tu dirección de correo electrónico no será publicada. Los campos obligatorios están marcados con *


    We use cookies to ensure that we give the best user experience on our website. If you continue to use this site we will assume that you agree. More information