asking questions
PHOTO: Emily Morter

IT research firm Nucleus Research published a new report looking at current uses of artificial intelligence (AI) in the enterprise. Compiled by analyst Daniel Elman, the report, "A Realistic Take on the State of Artificial Intelligence Today" (subscription required) looked at where AI was in the enterprise at the end of 2019.

The Need For Clean Data

The report acknowledges one of the factors holding businesses back from introducing machine learning systems: the massive amounts of data needed to accurately train the system. While this is fine in a research setting where the dataset can be painstakingly curated, labeled, and studied before being fed to the model for training, but in a business environment, data is often incomplete, irregular, or too siloed to efficiently aggregate in the necessary quantities.

When considering whether or not AI and machine learning (ML) should be brought into the enterprise, it cannot be stressed enough that companies looking to implement machine learning capabilities for their business must already have a mature data culture with institutional knowledge and practices for data collection, preparation, storage, and governance, Elman told CMSWire. “Vendors can sell out-of-the-box models trained on general third-party data for simple applications like sentiment detection in email, lead and opportunity scoring, or employee turnover, but for machine learning that is fully optimized for a specific process, years’ worth of historical data is often required,” he said.

Another significant consideration is the compute cost. These models require thousands of iterations of mathematical operations on high-dimensional tensors — for large models or continuous training. Businesses should be aware that the compute demands are heavy and need to account for the increased demand.

The result is that for customers looking to implement AI/machine learning for their enterprise, it’s important to understand the reality of the technology and look at their own internal data culture — weighing the compute costs associated with this technology — before jumping headfirst into the AI abyss.

Related Article: Why Artificial Intelligence Will Create More Jobs Than it Destroys

AI and Business Efficiency

Anthony Macciola, chief innovation officer at Milpitas, Calif-based ABBYY, agreed. He pointed to a recent Accenture survey, which indicated that 75% of executives said their companies might go out of business in the next five years if they are unable to scale AI. “The reasons for embracing AI and investment in automation are simple — more business efficiency and increased profits,” he said. Companies need AI technologies to tackle critical pain points in their content-centric operations, such as manual sorting and classification of documents, manual data extraction from documents, inadequate compliance with security/privacy regulations, and poor data, errors and inaccuracy of information.

Organizations that deployed AI technology specific to digitizing, automating, and optimizing document workflows reported reducing costs by over 35%. They also reduced the amount of time spent on document-related tasks, and perhaps the most compelling benefit was that they reduced errors by 52%.

Other reasons enterprises benefit by using AI for content intelligence include increased employee productivity, increased customer satisfaction, and improved responsiveness to customers, as well as new product or revenue opportunities, increased visibility and/or accountability, or increased customer engagement.

AI Benefits

The benefits of automation are too attractive to be ignored and as businesses transition away from paper into the digital world, AI will provide some useful assistance at simplifying several repetitive workflows and menial tasks, Mario Grunitz co-founder of Amsterdam-based WeAreBrain, said. He pointed out that, in the enterprise context, the benefits can be summarized as follows:

  • Greater overall efficiency
  • Improved accuracy (no/less human error)
  • Easier compliance
  • Decreased (repetitive) workload
  • Streamlined onboarding and training
  • Not to forget, this technology promises to be cost-effective

AI will allow enterprises to do more with less, freeing up time to grow their business and improve areas such as customer service. Think about accounts payable, procurement, booking, credit checks, salary processing, tax reports and time-consuming audits. These are all good candidates to introduce some AI-powered assistants. Outside of this there are three use cases were AI is well established:

1. Building a Data Culture

Jesse Rio Russell is president of Madison, Wis.-based Big Picture Research and Consulting and a data scientist specializing in helping clients work in the border area between the computational and technological side with the human and values side. “What we tell our clients is that they shouldn't focus on any one approach, or any one platform, or any one new technology. Instead, they should focus on building a culture of data. A culture of data where everyone in the enterprise can speak a common language around how language gets used in the company, a language for how date intersects with their own work, and a language for how data helps the enterprise serve its customers better. Once this kind of culture of data exists, then a company can start thinking about what technology to invest in, what platform to use, what questions to answer,” he said.

2. Predictive AI

Mario Ciabarra is CEO of Quantum Metric. He said that enterprises need to gather data and utilize new capabilities of AI to improve and personalize the experience that customers have with a brand at every level. But automation and AI for the sake of it won’t get companies where they want to be unless they approach the technology as an opportunity to also help align internal processes. “The predictive aspect of AI technology is absolutely crucial in helping companies prioritize their projects cohesively, and we can expect to see it become advanced enough to continuously optimize operations inside and out,” he said.

A use case would be a financial service provider dealing with a high churn rate. Predictive analytics would allow them to identify the customers that are having difficulty making payments or transfers, quantify the impact to the business, and surface the reasoning behind the issue without the team needing to spend days searching for the answer. With this look into the consumer's mind, businesses can pre-emptively align on digital priorities. Whether utilized by retailers, banks, or airlines, predictive tech can create empathy with what the customer will experience. Companies that cannot keep up will see themselves falling behind their competitors, losing revenue, as well as customer loyalty.”

Related Article: Is Augmented Artificial Intelligence Already Disrupting Artificial Intelligence?

3. Mobile Messaging

Beerud Sheth, the founder of Santa Clara, Calif-based Upwork, told us that businesses don't realize how powerful and transformative mobile messaging can be for their them. Most businesses focus their digital efforts on their websites and, maybe, their apps. However, consumer behavior, led by the millennial generation, is starting to shift to mobile messaging. They're impatient with long-duration voice calls or long-form web pages — they like short-form posts, tweets and messages. Successful businesses will learn to adapt to evolving consumer preferences.

Businesses should modify their systems to send rich, engaging, short messages containing actionable buttons. They should setup chatbots to handle customer queries and enable 2-way conversations with consumers. They should integrate their catalogues, booking systems and payments with messaging. Virtually all their customer touchpoints will eventually be through mobile messaging — more so than voice, web or app.

Every CIO of an organization of any reasonable size is spending time thinking about Digital Transformation. It is inescapable, thrust upon us by media hype as well as the very deep and real issues of increasing complexity and fast-growing pools and lakes of big data. In this era of accelerating complexity, prior methods to manage this data are quickly proving to be insufficient, Sheth said.

Fortunately, new tools like artificial intelligence are emerging as powerful aids to augment past methods and help automate some of these daunting tasks. “We are clearly past whether enterprises need this technology. What they need to consider now is how to achieve the right balance between human and machine efforts to optimize outcomes,” he said.