After people: machines handle the heavy lifting
We may be closer than ever to the possibility of autonomous enterprises, meaning startup founders or enterprise team leaders can launch and maintain business functions employing digital and artificial intelligence with little or even no staff. However, innovators need to proceed with caution, as there remain many elements of businesses that need the human touch — from humans who understand what needs to be done. Moreover, can AI and automation be completely trusted?
An autonomous enterprise can be defined as an organization that applies AI and automation to engagement, servicing, and operations to achieve what is essentially a self-driving business.
A recent study of 602 executives, published by HFS in April 2023, shows that “enterprises are embarking on a wide range of initiatives that are the building blocks and driving forces behind the autonomous enterprise.” The three leading strategic initiatives of today’s business leaders are to imporve automation of processes and data, leverage emerging technology, and imporve their ability to respond to market shifts.
Separately, a study of 600 executives from Pega shows momentum toward self-driving businesses, with 58% expecting to define themselves as an autonomous enterprise within the next 10 years. At least 15% say they are already at this stage today, and 36% project they will reach this point five years from now. Three quarters (73%) of respondents said they already have some sort of plan to start becoming an autonomous enterprise, and 96% expect to be there 10 years from now.
However, launching and sustaining a completely autonomous or self-driving enterprise may be a pipe dream, due to technological limitations and organizational issues. The autonomous enterprise “is a cool idea but the complexities of IT, the combination of legacy and cloud-native services, and the fact that most business-critical services for many enterprises still run on legacy systems including mainframes can make this difficult,” says Andy Thurai, principal analyst with Constellation Research.
With legacy services, “it is extremely difficult to scale the services up and down as needed,” Thurai continues. “With cloud-native services, it is somewhat easier to build an elastic IT to support a business need, though extremely complex scripting would be needed to make this happen as the solutions are not mature yet. This will require a lot more work and may not offer much of an ROI at the current time.”
If designing a fully autonomous enterprise, the first order of business is “to be able to predict the needs and demand accurately in any functional area,” Thurai states. “If a customer calls with a specific question, the system should be able to predict with reasonable accuracy the reason for the customer call, or an IT system should be able to predict the demand forecasting based on a certain that might occur based on historical patterns. Once that is done, then the second necessity of executing that to achieve the desired results such as an elastic IT system scaling up or down based on real-time demand. It might take years, if not decades, to achieve that.”
Tools such as generative AI are also accelerating adoption of automation across business functions and services. “There is a rising need to customize these models to enterprise workflows,” according to Saloni Patil, project lead for digital at Zinnov, speaking at a recent webinar. An example would be customizing a marketing campaign to a business's product portfolio.
Developing customized applications and systems is key, and, as part of this trend, there has been “an upsurge in launching open source data set libraries and LLMs,” says Patil. “This is to help developers generate bias-free algorithms and customize them on their own data sets. There was an inherent need to AI to prove and moderate the content that was being generated by AI, to flag inappropriate content and avoid misinformation.”
HFS researchers sound a note of caution around supporting infrastructure. “End-to-end automation of an organization presents a persistent challenge, with processes and departments running in siloes and making siloed technology choices,” they state. “Organizations don’t give business functions enough scope to drive better internal alignment toward building end-to-end solutions that are fundamental to the autonomous enterprise.”
For these reasons, “true autonomy is a ways away,” Thurai emphasizes. “Enterprise IT is very siloed, and each function CRM, ERP, and supply chains, are all working as their own islands with most parts with many inefficiencies built in. In certain areas, such as business processes, IT operations, and customer service support. We are only making progress incrementally.”
Trust and effective governance are the keys to success in any autonomous enterprise, the HFS team emphasizes.
