To bot or not to bot: Understanding A.I.’s role in the enterprise

Over the past few months, the internet has been buzzing about bots.

Microsoft, Facebook and other major players are opening up their A.I. platforms for developers, and these toolkits are an exciting and important step towards democratizing access to A.I.

[aditude-amp id="flyingcarpet" targeting='{"env":"staging","page_type":"article","post_id":2006739,"post_type":"guest","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"bots,","session":"D"}']

For the enterprise, however, the bot frenzy has accelerated a challenge that executives have been facing for the past couple of years. Many enterprise companies understand that they need an A.I. strategy and that the technology will be deployed throughout their business. Yet the challenge for them is where to actually begin.

Transformational tech

These companies understand that A.I. is a transformational integration for their business and that it will eventually touch everything from their customer service, their analytics and business intelligence, sales and CRM, and even internal knowledge management and HCM.

Realizing that the strategic decisions around their A.I. deployments will have impacts across their business for years to come, enterprise executives are rightly not taking the decision lightly. They have taken a wait-and-see attitude.

However, those decisions have now been forced to the front of the agenda. If bots are the new apps, as both Microsoft and Facebook have declared, then it’s reasonable to expect that soon we’ll be facing a veritable Cambrian explosion of bots and bot stores, just as we’ve seen with apps over the past decade. Before long we’ll be accessing bot stores with our smartphones. Our new apps will exploit natural language processors, tone analyzers, and other intelligent functions collectively gathered under the umbrella of artificial intelligence.

If you’re in a position to determine where A.I. fits into the future of your enterprise organization, any enthusiasm you may have for a post-app era is likely now tempered by a good deal of bot anxiety and confusion. As the enterprise rushes to adopt A.I. to remain relevant in a world increasingly powered by machine learning and cognitive computing, bots may provide a temptingly accessible and low-risk first step.

Bot buyer beware

Bots will undoubtedly be deployed to handle any number of tasks that can and should be automated. Bots are the perfect solution for small transactional tasks that ought to place a low cognitive burden on human and computer alike.

This isn’t anything new, but we’re about to be inundated by a world of bots. In fact, most of our initial interactions with companies will likely be through some form of bot. This means that companies need to develop a strategy and framework to understand when bots will be the effective solution within their organization and when they need a more deeply integrated intelligent system.

Bots will be an excellent way for many companies to begin exploring A.I. and its role for their business. Unlike clunky in-app and push notifications, bots and messaging apps meet the user where they already are. They are also a great way to get started with conversational user experiences and build a business case in the process. And perhaps most enticingly, they promise to be pretty easy to build and maintain.

[aditude-amp id="medium1" targeting='{"env":"staging","page_type":"article","post_id":2006739,"post_type":"guest","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"bots,","session":"D"}']

Yet bots and messaging apps are what we might call “hard-coded” intelligence, which means that the use cases they serve are predetermined and pre-scripted. Bots and messaging apps are very brittle for this reason. They get very finicky when you ask them to step outside their comfort zone.

Many businesses will be drawn in by the allure of how quickly they can create a bot, but they’ll quickly come to want more. Developers love the concept of the end-user experience but ultimately require a greater level of intelligence to deliver it. Bots can’t tell us when they are good or not, for example. They are not particularly self-analytical, leaving open the question of whether they are satisfying business objectives.

Bots will also present a scaling challenge as the enterprise dives more deeply into A.I. and relies more heavily on automated, bot-deployed tasks and communications. Unless an organization is prepared to keep deploying an increasing army of bots for the specific tasks needed (and maintaining those bots with upgrades across the full as array as needed), then they will quickly find themselves in need of a more robust intelligent system for which bots are but one end-point of access and communication.

‘One or more’ rule of enterprise bots

In thinking about whether bots will be an effective solution for your enterprise needs, it’s helpful to consider what I refer to as the “One or More Rule.” This framework can help guide decision makers on when bots are the right kind of automated intelligence for a task, and when more intelligence is required.

[aditude-amp id="medium2" targeting='{"env":"staging","page_type":"article","post_id":2006739,"post_type":"guest","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"bots,","session":"D"}']

More than one utterance:

Bots and messaging apps are call-and-response mechanisms. They receive a prompt and return a timely, mostly uniform response. If and when your use cases begin to require a real conversation, meaning that the discussion naturally wants to continue with perhaps a clarification or a subsequent choice by the user, it’s time for more integrated solutions. Bots and messaging apps aren’t good at back-and-forth; true contextual conversation, supported by domain expertise, requires a deeper commitment to AI development than generic NLP interactions.

More than one step:

Bots and messaging apps don’t run a sophisticated underlying domain and language model that would allow them to execute complex tasks on a user’s behalf. Making travel arrangements is the most common example that people use, but from insurance qualification lookups to financial advice, customer service inquiries, and more, bots need help when your business process has more than one destination or more than one acceptable outcome.

[aditude-amp id="medium3" targeting='{"env":"staging","page_type":"article","post_id":2006739,"post_type":"guest","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"bots,","session":"D"}']

More than one system:

Bots and messaging apps don’t necessarily have the smarts to compile information from more than one system of record. They run well on a single database or feed to which they are highly attuned. They don’t do well with different data structures and/or different web services, the combination of which inevitably throws them for a loop. For the enterprise, with many systems of record that their A.I. must call upon to achieve effective resolution and deliver the right information, a more integrated intelligent system will likely be needed.

More than one domain:

Bots and messaging apps similarly fail when asked to cross domains. It is much easier to model a single vertical with a finite number of statistically prevalent commands than it is to teach a computer how to traverse multiple fields of knowledge. Bots and messaging apps are, for this reason, best used for single-serve, single-environment needs.

[aditude-amp id="medium4" targeting='{"env":"staging","page_type":"article","post_id":2006739,"post_type":"guest","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"bots,","session":"D"}']

 

Bots will undoubtedly play an important role for the enterprise in the coming years, and become a more familiar form of interaction for customers. For companies looking to begin exploring A.I. employments, bots will likely be an excellent starting point as they begin to understand their needs and the full capabilities of how A.I. can be leveraged across their organization. But it’s also important in this nascent moment of enterprise A.I. that decision makers are thinking strategically and with the long-term vision of how intelligent systems will be integrated fully across and transform their business.