How to survive the looming AI avalanche of enterprise automation

The latest advances in AI have brought us to the brink of a massive upsurge in enterprise automation. But will this deliver the promised benefits? AI is still subject to all the same caveats that apply to any new technology. It will take longer than we expect to make a huge difference — per Bill Gates' adage that we overestimate what we can achieve in one year but underestimate what can be done in ten. In the first few years, people will apply the technology to speed up what they already do rather than use it for true innovation — the horseless carriage syndrome. As a consequence of these two factors, much of the investment in AI in the next few years will be wasted. But just like John Wanamaker's spend on advertising, the difficulty for enterprises will be knowing which investments will turn out to have been in vain.

One thing's certain. AI is making it dramatically easier for people to automate the work they do. It's now possible to create an agent to perform a task much faster and more easily than ever before. Whereas in the past, building an automation meant mapping out every possible instruction and carefully linking each of them to potential actions, generative AI cuts out a lot of that manual effort. It is able to understand everyday language and automatically figure out how to work with data and functions it finds in the underlying system to deliver what the user is asking for. This means that developers can build automations far more quickly than before, and in many cases non-developers can use no-code tooling to build automations themselves without even having to involve developers. It sounds like a huge step forward, but is a rapid roll-out of new automation necessarily going to be a good thing?

At first sight it seems very welcome. When learning how enterprises have been adopting AI over the past few months, one thing that's struck me is that many of the examples are familiar automation challenges. Rather than enabling new automations, it seems that AI is mostly used to make it easier and quicker to introduce automations that were already long overdue. This leads to a huge amount of optimism around AI. First of all, it brings new hope of clearing up the longstanding backlog of pent-up demand from people across the enterprise to just be able to get stuff done faster and with less faff. Secondly, it creates an overwhelming impression that AI is delivering an immediate and impactful return on investment. This is somewhat misleading. These are automations that always were needed, and could have been delivered earlier if they'd been prioritized, but were denied the resources, budget or motivation. AI simply lowered the bar to getting them done. The ROI comes not because of AI per se, but as a result of the automation of previously manual processes. Nevertheless, AI swoops in at the last moment and gets all the credit.

All of this further fuels the hype around AI, encouraging enterprises to unleash the technology as rapidly as possible and accelerate the pace of automation. Trouble is, however overdue these automations may be, many of them will quickly become redundant as the technology moves forward. The most obvious route today may not turn out to be the best choice in hindsight. Unleashing an avalanche of automation without proper forethought or co-ordination will have unpredictable and disruptive results. The enterprise landscape will be much changed, but not necessarily for the better.

Suboptimal processes

The biggest danger comes from automating processes that were already suboptimal. In the past, I've often questioned the utility of transforming paper forms into manually completed PDFs that are digitally signed, when the system can automatically present an action for approval and has already validated their identity when they signed in. Similarly, there's little value today in helping employees write better emails when a better process would eliminate the need to send an email in the first place. Helping people draft reports more efficiently risks adding to a mountain of reports that no-one ever reads or acts upon. Before automating an existing process, the first question should be, do we even need this process given the technology and connectivity we now have available?

This risk is compounded once automation is democratized through the dissemination of no-code tools. With everyone suddenly able to automate their own processes, there will soon be multiple automations across the enterprise, each doing essentially the same thing but in very different ways — the vast majority of which will be hugely inefficient. Developers often talk about the concept of technical debt, where successive enhancements and modifications are layered on top of each other and over time lead to inefficiencies and potential conflicts. The same pattern exists in process debt — layering new automations on top of each other without revisiting the underlying processes may achieve little more than speeding up wasted effort. In both cases, it's important to take stock every so often and work out where the edifice needs to be rebuilt to streamline operations.

Ultimately, enterprises need to consider radically new ways of doing things that are enabled by the technology. For example, spend management vendor Coupa is currently looking to build an AI-powered supplier collaboration platform that will bypass the need to exchange traditional RFPs, purchase orders and invoices because its autonomous agents will be able to negotiate and complete those transactions. Salvatore Lombardo, Chief Product and Technology Officer at Coupa, told me:

If you use AI and just ask the question, 'What can I automate tomorrow?', you will gain a little bit of value out of it. But this is not the real disruption. No, the real disruption is combining it with the idea of collaboration, inventing collaboration objects which know each other [and] have the data set, which they can call, discuss, and create things with each other because they're intelligent agents as bots, talking to each other, so to speak.

Choosing a careful path

This is just one example out of many innovative approaches that are currently under development, but which will not come to market in the immediate future. Enterprises therefore need to choose a careful path in their adoption of AI. On the one hand, the pace of change is terrifying. As I recently noted when considering the challenges Microsoft face in bringing AI to its customers at a time when the technology is doubling in performance every six months:

People tend to underestimate the compounding effect of scaling at that pace. Fail to act and your competitor will have a 10x advantage in less than two years, a 50x advantage in less than three, and a 100x advantage six months later. But if you make the wrong choice, you'll not only fall behind just as much, but lose your investment too.

Timing is everything. There's certainly some low-hanging fruit that enterprises should seize immediately — overdue automations that now become affordable or feasible thanks to the new capabilities that AI brings, and which deliver significant business value. But a headlong rush to trigger an avalanche of automation risks burying the organization in a mass of contradictory, wasteful processes that defeat the ultimate object of cutting costs or increasing output.

One of the hallmarks of the long-term impact of any new technology is the advent of standardized approaches — consider the emergence of automated configurations scripts and containerization as cloud computing grew to industrial scale, or how enterprise SaaS coalesced around configurable business processes that came ready-to-use. When considering how enterprises should harness AI, my expectation is that the full impact will only come after the adoption of a new wave of standardization in how long-established enterprise processes are carried out — standardization that, like Coupa's proposed supplier collaboration platform, cuts across enterprise boundaries to establish common data models to enable more powerful automations. It's time to start keeping an eye out for these nascent standards as they begin to emerge, but in the meantime maintain careful oversight of AI-powered automation within every organization to ensure that it isn't creating wasted duplication of effort instead of meaningful business outcomes.