AI/Automation
AI for the Rest of the Enterprise
You wake up to Alexa’s soothing voice wishing you good morning. Although she can’t make your coffee, the ever-present assistant guides you to the nearest cafe to grab a cup. Your phone unlocks through facial recognition, and personal emails are already organized.
From the start of each day, artificial intelligence (AI) fills and guides our personal lives, even for the simplest tasks. But that ubiquity often changes once we get to work. A KPMG survey of 30 of the world’s largest companies found that nearly one-third used AI for selected functions. However, only 17% deployed technology at scale.
Many organizations and startups have created amazing AI algorithms and techniques that can solve specific business problems. That narrow focus is valuable, but companies struggle with whether they can make the technology scalable. Even if they can expand AI enterprisewide, executives must grapple with serious issues, such as finding the right data to train the models, ensuring data privacy, ethical concerns, and reliability.
AI isn’t just a technology tool
Companies plan to expand their use of AI (Figure 1), but first must create a structured approach to discovering, developing, and democratizing the technology throughout the enterprise. Success lies in identifying the right opportunities to unlock business value, rather than thinking of it only as a technology tool with limited uses.
Figure 1. Organizations are significantly accelerating their adoption of AI and machine learning applications.
Below are the steps required to expand AI companywide, instead of relegating it to a few isolated corners:
- Maturity assessment framework — Companies need to understand their AI and automation maturity across multiple dimensions, including the current AI and automation strategy, data strategy, organizational readiness, and technology readiness.
- Process discovery — AI is most valuable when companies define, map, and analyze their processes end to end — from the customers’ perspective. This includes process discovery using value stream mapping, day-in-the-life analysis, and process mining (using tools such as Celonis, Minit, AssistEdge Discover, and Digitran).
- Data-based discovery — Organizations can use their legacy systems to curate insights and learning models. Businesses should use demand-capacity correlation analysis, event alert correlation and problem analytics, service assurance, and social media analytics. The result is greater insight into customers’ needs.
- Co-innovate with partners — Companies can discover new opportunities by seeking outside assistance with business priorities. These collaborations can help conduct structured interviews with important stakeholders, review key performance indicators, and set up design thinking workshops. Infosys has successfully used its Living Labs to innovate, experiment, and co-create in situations such as these.
How to embed AI
There are different ways that AI can improve business operations while seamlessly integrating with existing systems. When implemented well, the resulting applications and services can operate with greater efficiency and offer better customer experience.
Companies can use AI operations (AIOps) to incorporate preventive maintenance and self-healing capabilities into application and infrastructure management. AIOps can enable organizations to detect application failure even before it occurs, thereby giving support teams enough time to fix the problem or to self-heal without business disruption. This can be accomplished through the use of bot factories that produce armies of digital workers: sensing bots, analytical bots, and action bots.
To manage business operations, companies can use conversational AI and process bots that combine to create virtual agents. These significantly improve efficiency and increase customer satisfaction by reviewing customer histories, asking relevant questions, and providing fast responses. For example, Infosys’ AI-enabled Cortex 2 call center platform can analyze customer data, history, and call patterns, and then prioritize tickets. The result is reduced call handling times.
These efforts are not as daunting as they sound. Application development now requires less technical knowledge than what many organizations expect. There are multiple low-code and no-code platforms available that allow engineers to create new applications in minutes. These provide out-of-the-box AI-assisted design and development, machine learning-based code generation, and mixed reality interfaces with a screenless user interface for a more immersive experience.
Tools such as the Infosys Digital Foundry can create templates for selecting security, technologies, environment provisioning, architecture and design patterns, and for setting up source control with integrated DevSecOps. Testing and validation also benefits from AI, using QAOps, spidering AI, and AI-based data generation.
Modernization suites can ingest legacy codes and noncode artifacts to derive meaningful insights and thereby ease integration of legacy systems with the latest application suites. Organizations need not view new applications as a burden or an overhead, but instead can view them as a means to modernize their existing systems.
How can we democratize AI at scale?
AI offers clear benefits but requires work on the part of the organization that embraces these technologies. Companies that switch to an AI-enabled world need to do so in a way that doesn’t hinder their legacy systems. Businesses also need to manage their risks and create an environment that allows them to learn, apply, and take AI to scale. Here are ways to move in that direction:
- Industrywide and business function use cases give companies a greater understanding of real-life implementations and illustrate how to emulate these successes. Reusable bots and digital workers also help accelerate adoption of AI and automation and return value faster.
- Applied AI cloud, curated with services in each domain, can be applied across solutions, including technology services for vision, speech, text, language, and more.
- An ecosystem of startups and partners can help with the adoption of AI platforms and solutions, such as methodologies for handling data privacy, bias, and trustworthiness.
- Businesses need a structured approach to change management to nudge end-user behavior toward applied AI. Organizations should also reskill their workforces to prioritize AI.
Andrew Ng, AI luminary and co-founder of Coursera, described AI as the “new electricity.” In Ng’s words, “Just as electricity transformed almost everything 100 years ago, today I actually have a hard time thinking of an industry that I don’t think AI will transform in the next several years.”
With AI being so pervasive in our lives now, organizations cannot risk ignoring the technology or failing to understand its broad uses — future-proofing technology, managing risk, and delivering new business solutions. Applied AI is the necessary way forward for companies that want to provide a superior experience for clients and employees.
