Deploy AI Apps Across the Enterprise
In the previous article in this series, we discussed Making Your Data and Your Enterprise Ready for AI.
Once your data and your organization are ready for AI, there is a tremendous amount of value that can be delivered through Infusing AI throughout your enterprise. AI has been perceived to be on a hype curve for years, but through our thousands of engagements with clients, we’ve proven that there is real value to clients and enterprises. AI can be deployed in so many ways across an organization that it is easy to lose sight of the most impactful use cases, the areas that deliver the quickest ROI, and the use cases that allow you to continue to Infuse AI in your enterprise. Throughout this article, we will explore the top three AI use cases that we see in enterprises, the consequences of the current state, and how modernizing can benefit your organization and your clients. These top three use cases focus on AI for Customer Care, AI for Planning and Forecasting, and AI for Managing Risk.
AI for Customer Care
Nearly every client that I speak to has a goal to increase responsiveness to their clients, while at the same time increasing their NPS. Most of these clients have explored basic chatbots to help reduce the burden on their agents. Still, most of these organizations have also stopped short of deploying across all avenues of interaction, and almost all have only delivered the ability to ask simple, short questions. Additionally, most of the organizations that have implemented AI solutions for Customer Service are still struggling because they have implemented solutions that require an extremely high level of technical skill to use. There is a strong need to transform the buyer journey or “experience,” which in turn provides more information for enterprises to make better business decisions, while also providing a personalized experience. Applying AI systems only in pockets of the business, and using tools that need high levels of technical expertise, ultimately restricts companies from realizing the full value of AI.
Current State and Related Consequences
Currently, organizations are struggling with high volumes of incoming calls, high turnover in their call center workforce, and siloed approaches to handling customer interactions. Due to these issues, organizations are looking to modernize the way they engage with clients, ultimately resulting in higher NPS (Net Promoter Score). 75% of organizations want to leverage customer service bots, but most do so in a siloed approach, if at all. Half of the clients believe that the agents that they are interacting with do not understand their needs. Clients want to engage in their time, with their methods. Gone are the times when clients will choose to call into a call center for resolution; clients want to interact through SMS, Social Media, Email, etc.
Due to these issues, clients are choosing to leave their existing providers. They move to ones with more modernized experiences; this is incredibly important, as 91% of unsatisfied customers will not return to a business. This dissatisfaction is continually occurring in current organizations, as 50% of customer service calls go unanswered or require escalations to get answers. All of these deficiencies equate to massive wastes in spending and vast amounts of lost revenue. For example:
· Acquiring a new customer can cost five times more than retaining an existing customer.
· Increasing customer retention by 5% can increase profits from 25–95%
· The success rate of selling to a customer you already have is 60–70%.
· In contrast, the success rate of selling to a new customer is 5–20%.
Clients want to get answers through their preferred communication channel, they want answers immediately, and they want to be able to get answers to complex questions that new agents or limited chatbots can’t answer. All of this is required to create better experiences, which ultimately result in reduced costs and increased revenues.
To be successful in delivering on the future state, an enterprise requires:
1. A deeply integrated, consistently excellent, highly personalized user experience across all channels, including chat, voice, email, and social media.
2. The ability to call center agents to leverage AI solutions internally to assist with finding the correct answers, faster from complex business documents.
3. A fully integrated AI platform that anyone in the organization can use to build and train AI capabilities (not just developers or data scientists).
4. Integration into a broader Data and AI platform to deliver a more integrated experience for clients.
The optimal future state is one in which organizations allow for the integration and seamless handoff of client interactions across all channels so that clients can pick up where they left off. Where clients get correct answers rapidly, and that allows anyone in the organization to build AI applications, rather than limiting it to just Data Scientists and application developers.
AI for Planning and Forecasting
Organizations are dealing with increased market volatility, competitive pressures, and ever-changing customer demands. Not to mention the changing global landscape due to the impact from global pandemics, natural disasters, and economic hardships, which in turn are impacting supply chains, business operations, and the way organizations deliver products and services. As a result, the use cases that we see most impacted are scenario planning, resource planning, and financial and operational reporting. Error-prone planning processes inhibit the ability to view business performance across individual business units and the organization as a whole, which makes it challenging to respond to change when needed. As the business environment changes, the role of finance is changing too. It’s becoming more forward-looking, focusing on guidance, evaluating opportunities for growth, and predicting future performance. Integration is key to streamlined planning, budgeting, and forecasting. Operations, sales, marketing, human resources, and other departments and disciplines all need fast, flexible planning and analysis. And all of them can use the same tools to provide insight and manage performance, while also leveraging additional AI modeling capabilities. When people in one part of the organization see how their decisions affect other parts of the organization, all of the activities will be better coordinated and drive better results. Integrated planning ensures that all parts of the organization are connected, and planning is streamlined.
Current State and Related Consequences
The current state is one in which there are Manual, disconnected planning processes that result in using spreadsheets for planning. Planning processes are time-consuming, error-prone, inflexible, and siloed between departments. Disparate planning processes create analytics silos and data inconsistencies; as a result, a holistic view of business performance across the organization is challenging to achieve. Highly skilled Finance professionals spend a majority of their time building plans and forecasts, and not enough time analyzing data to create recommendations to improve business performance. Data is often aggregated and updated manually, leading to data integrity and accuracy concerns. Organizations cannot link operational tactics with financial plans to ensure alignment across the organization. Planning is hard, which is why many companies only do this exercise once a year or once a quarter. Today’s “Just in Time” processes require organizations to have the ability to deliver planning models rapidly to meet ever-changing market demands. For example, Amazon executes intraday planning and has evolved to be able to do this every 15 minutes using IBM Planning. There is also a lack of ability to perform granular analysis and deliver more substantial insight into trends and root-cause analysis of performance gaps. Organizations are generally unable to become data-driven and forward-thinking. They are mostly relegated to manual processes and antiquated tools.
To be successful in delivering on the future state, an enterprise requires:
1. A single source of truth to ensure confidence in data, plan, and forecast accuracy.
2. Acceleration– organizations need to automate slow, spreadsheet-based planning, budgeting, and forecasting to derive insights faster.
3. Collaboration- organizations need to synchronize processes and data across the organization from finance to operations to create more accurate, integrated plans.
4. Agility- organizations need to ability to create effective plans, forecasts, and reports that can easily adjust to changing internal and external conditions.
5. An integrated planning process to streamline and connect all parts of the business to provide a comprehensive view of business performance.
AI for Managing Risk
All industries must monitor and mitigate risk, especially those highly regulated, such as banks, insurance companies, and institutions within the energy, utilities, and telecom industries. As the Risk executive, one needs to understand and manage the business’s risk appetite, and tolerance. It is vital, therefore, to have a high confidence level that your risk departments ‘know what they need to know’ and have the right processes, control and metrics in place to protect the business and allow it to propel forward. AI can “read and interpret” changes to regulatory compliance regulation, and both help organizations make sense of the changes and also determine if their organization is compliant with changes. Institutions are wrestling with constant uncertainty and increasing compliance requirements. Manual processes inhibit the ability to gain a complete view of risk exposures across all the individual business entities that make up the enterprise, making it difficult to respond to change when needed. Risk executives want to shift their operations from a “what happened” operating model to a “what’s coming” pro-active approach. The infusion of AI into these applications allows organizations to see patterns and trends in real-time and create or adjust risk models accordingly, not “after the fact.”
One where everyone in the institution realizes they play a part in managing risk. To establish this, timely, relevant, and predictive information needs to be shared by all levels of the business. Organizations need to move to a proactive approach to managing risk, rather than backward-looking.
Current State and Related Consequences
Currently, organizations have a lack of a holistic end-to-end view of risk and compliance for the enterprise. Standardizing across the multiple siloed GRC and supporting systems produces incomplete pictures and incomplete answers. Along with that, regulatory compliance monitoring processes are primarily manual, time-consuming, and prone to error, which puts a significant burden on the Risk department. Additionally, interpretations are also generally inconsistent. Finally, there is an inability to readily demonstrate, with confidence, that models and risk decisions have been correctly created, validated, and used appropriately. Due to this, business users are not able to identify multiple instances of risk nor view reliable history to get to a root cause of an incident. Also, data is often combined and manipulated manually, leading to a concern of data integrity and trust in resulting analytics and conclusions.
Siloed sources, manual processes, and inability to audit decisions create an environment of uncertainty and levies a heavy burden on personnel in the risk department.
To be successful in delivering on the future state, an enterprise requires:
1. A standard risk library and single data model that ensures consistent holistic views of risk and compliance that is scalable to tens of thousands of users. Advanced analytics, visualizations, AI evolves risk views from What’s Happened” to “What’s Coming.”
2. Out of the box visualizations, and personalized workspaces to improve risk management and productivity.
3. Collaboration between business and IT (visual design using UI and embedded workflow) and accelerate using pre-built content (configurable, customizable).
Ultimately, organizations are moving towards deploying AI across their enterprise. However, to be successful, organizations need to make their data and enterprise-ready for AI. Once they have done that, solutions and technologies must be implemented with a non-siloed approach, that allows all skill levels in the organization to participate, and assurance that these solutions and technologies integrate seamlessly with existing Data systems and processes.
Co-Authored by Chris Zobler (Christopher Zobler)
John Emmert leads Global Sales and Strategy for IBM Information Architecture within IBM Data and AI. Lives in Raleigh, NC. He is married to his wife Sarah, and has 3 boys, Jack (7), Liam (7), Cormac (1).
