Whither AI for Enterprises?

With the recent practical advancements in AI and the recent recognition of the field with a couple of Nobel prizes in 2024, investments in the application of AI technology in our daily lives are sky-high. Furthermore, as enterprises aim to adopt all kinds of technologies with “AI” embedded in them, it behooves organizations of all shapes and sizes - to establish a strategy for exploring and adopting AI technologies. Advances in the field are yet to come, and questions linger - is AI ready for prime time - if so, when, how, and more. This brief essay aims to set AI in context for enterprise leaders and provide some guidelines to consider as they continue their AI adoption journeys.

As a first step, it is important to establish what “AI” is and what it can do in the larger scheme of things. A good point of departure is to understand this in the context of the earlier article in this forum titled - “Thoughts, Tools and Techniques,” - which, in the bigger picture, highlights the lifecycle of any technology and associated knowledge creation lifecycle. So firstly, “AI” is a “meta-technology” - a way of discovering and doing things repeatably. “AI” can come in many forms - software, hardware, network, small, big, and more.  Now, the hubbub around AI has been because - Does it enable “better” thoughts? Does it enable one to build “better” tools? Does it enable one to discover “better” techniques? Or is AI a tool for improving efficiencies in the “thought” generation process, tool building process, or executing a pre-defined technique efficiently?

We believe the reason for this is that “AI” as technology aims to do what is/was usually considered a highly “human” activity - creativity, thought, discovery, and more - basically knowledge “creation”. The issue of efficiencies has been addressed in the context of automation. Automation with thought to handle “new” things is a different ball game.  Furthermore, the anthropomorphization of AI and hype has added to the overall confusion and misunderstanding.

“AI” enables creativity in content generation, as seen in textual, image, and video content generation. AI enables better tooling, as illustrated in developing tools for coding, data analysis, and more. AI also allows better “techniques” as it learns by example - such as training robotic manipulation, the robots of Boston Dynamics, and more - from humans executing those activities.

The recent prototypes and proofs-of-concept illustrate the art of the possible, but much remains to be done to make these AI-tech-based systems reliable enough to incorporate into your revenue-generating activities - whatever the core competencies of an enterprise are. There are a couple of reasons for this - 1)  going from prototype to a reliable technology - requires major investment in AI tech development and all the auxiliary systems and processes around it - without well-defined performance metrics and goals - this seems like a moving target. From the prototype (that you see with current AI platforms) to production, the remaining 80% of the journey is conservatively speaking. 2) Envisioning your organization - and its people, customers, and suppliers- how will that evolve -. This is not easy to do as most of us have really not thought through all the implications of AI technology in one’s own organization. We still need people/humans to interact and manage these AI-enabled systems, and this requires a completely new level of “technical/AI literacy” - as humans and machines interact in the real world. Experiences in the next few decades will guide us in this process. However, the pressing issue for most Enterprise leaders is - what should I do with AI in my organization in the next year, 3 years, and 5 years or longer?

Towards this end, we propose some basic guidelines (Some of it may be counter-intuitive)  -

1. Firstly, review what your organization’s core competency (could be multiple)  is - the revenue-generating part and assess if introducing any AI-enabled tech is going to move the needle on your core offerings - or is it useful in making all the other regular business functions more efficient. A good example is if you are a restaurant offering a specialized cuisine - there is no point in replacing the chef. You may invest in AI technology to make better ingredients and more. However, all other restaurant management activities may be AI-enabled - but you will source those technologies from other vendors rather than invest in developing them in-house. Implementing AI technology requires you to standardize many aspects of your offerings - such as the food in quick service restaurants -  but one prefers a custom-cooked hamburger any day compared to the one from a fast food restaurant. So, to conclude, AI tech that you invest in-house should enable you to push the needle on your core offering - if not - wait for other things to evolve.
2. Invest in your people - if tech is core to your offering - and AI is relevant (from 1 above), then get the best people to own the problem and build your viewpoint. Embed it in your offering and validate it with your customers. Evaluate if it meets your current needs and enables future growth - new features, new techniques, and more. This is not easy to do - usually projects of this kind are in the 3-5 year range. One-year projects are highly exploratory just to assess the state of the art.
3. If you are doing 2) above, budget for investing for a longer horizon - As the AI tech stack evolves continuously, it is imperative to be patient - as your teams learn the capabilities of the technology, assess its relevance, and more. One has to realize that one is starting a new startup - in-house, and it needs to be managed accordingly. Each is a moonshot to get the level of quality and reliability one seeks.
4. Projects that are relevant to your core offerings versus the auxiliary ones need to be run independently - goals are different, and success criteria and complexities are different.

We hope some of the above musings help you organize your plans for an AI-driven future.

Much remains to be learned as the AI journey continues. In the near future, we will share more of what we have learned and learned in this exciting space. If you have comments and questions, please contact us at editor@wdf.ai.