The rise of artificial intelligence makes for sensational headlines.
And while intelligent software agents are not new, we are clearly entering a new wave of innovation around bots and related software that has the potential to impact our personal lives, our business lives and business operations in general.
The Rise of the Bots: Why Now?
The current blossoming of machine-learning is driven by both technical and environmental factors, and influenced by business and consumer-oriented perspectives.
Businesses struggle with the corporate brain drain caused by the continuous churn of employees, and increasingly turn to technology to capture and share information, to somehow retain and capture the knowledge that powers organizational processes. At the same time, businesses face pressure to scale effectively in the global digital environment, to cope with the need to perform actions or make decisions repetitively at high speed, without error.
Our expectations as individuals also play a part. Interacting with tools like Siri and Cortana makes us more comfortable engaging with technology using natural language, and we increasingly accept — or expect — recommendations from technology, such as website recommendations.
What’s more, advances in a number of technologies — developing in parallel — make this blossoming possible. Our ability to collect and process vast datasets from all aspects of our business and home-life environments helps bot applications improve their understanding of natural language, while the more pervasive availability of APIs makes it easier to connect to applications.
Add to this the democratization of development and the increasing availability of (often free) bot development and natural language processing toolkits, and the foundations are fully in place.
What’s New in Modern Bots?
Today’s new crop of bots and agents deliver materially different value in the work environment than they did five years ago — though based on the same foundational technology.
These learning systems are moving out of the back-office and infrastructure automation jobs, and into the flow of work — engaging directly with the customers and employees who use an organization’s services.
Examples abound: from chatbots tasked with taking care of the most common or simply-resolved support requests, to virtual assistants that respond to voice-activated commands and undertake more sophisticated activities on the user’s behalf (such as search-on-steroids, calendar scheduling, or even integrating with richer layers of service applications to send payments, place orders, book travel, etc.).
Bots (and similar agents) now learn from past experiences, which means they no longer have to follow rigid flowcharts, unable to deviate from a script (however elaborate) without human intervention. Now every interaction and data point is a learning opportunity that contributes to a bot’s level of intelligence about the tasks at hand. And it can put that better understanding of context and possibilities to effective use by suggesting alternatives, where its predecessors would have foundered without human help.
For example: IBM Watson, operating within IBM’s Verse email and calendaring application, can now suggest what meeting to attend in the event of a calendar clash based on your role, project responsibilities and past interactions. If you trust it enough, bots like this can schedule your appointments and liaise with meeting organizers — who may be bots themselves — without your involvement.
We’re also seeing organizations deploying bots to augment human-driven work, rather than provide full-on automation. An intelligent agent acts as an ‘expert on tap’ to drive work outcomes more quickly, when it’s impractical (because of a dispersed workforce) or uneconomical (because of scale) to rely purely on human experts to get things done.
What's Your Business Strategy?
Approaches to automation or augmentation will differ, depending on the task or the work scenario.
A comprehensive automation strategy for your organization will involve a number of factors for each scenario, including: the desired outcome (is it about increasing work volume, increasing work accuracy, increasing value-add, and so on); the cost of automation / augmentation compared to the cost of adding more people (which might depend on your operating model); cultural factors (which might vary by region or country); and, of course, existing levels of automation.