Adopting Practical AI Across the Enterprise

Enterprise companies are relentlessly aiming for better market positioning, higher revenue, lower cost, optimized operation, higher productivity, better customer experience, deeper insight, and sustainable growth. Applying strategic thinking and systematic approach on adopting enterprise AI can help them break through. By Yingwu Gao

In 2020, artificial intelligence (AI) is moving from experimentation to adoption in the enterprise. According to a recent survey by IBM, enterprise AI is catching on faster than expected. About 45 percent of senior executives of large companies surveyed by IBM said they have adopted AI. But what about the rest who have not? I believe the key to making enterprise AI break through is for businesses to align a strategy with the practical intelligence they found fit, then align with AI technology.

AI for Enterprise

The majority of businesses executives I talk with are aware of the importance of AI and its potential. They want to know how to break through with its adoption. They agree on envisioning AI as going beyond a success with one use case or project. They want to scale AI to the organizational level to benefit the entire business functions, including sales and marketing, service and support, supply chain and logistics, information technology and production, finance, human resources, and so on. But they also realize barriers to AI enterprise adoption exist. In fact, 85 percent of AI projects will not deliver for CIOs. Businesses are encountering many obstacles toward achieving enterprise AI. They include:

Indecision: where and how do they get started? When do they build, buy, or partner? Businesses lack good use cases that speak to their business challenges.
A culture of fear: a fear of taking risks; fear of financial impact of initial investment; fear of competition; and uncertainty about market dynamics.
Data: lack a unified and simplified data infrastructure and ecosystem; lack enough good-quality data; lack right resources and tools; underestimate the complexity of data pipeline management and model training.
Platform: lack productive platform or there is no platform to develop and deliver solutions end to end, in time and under budget.
Expertise: there exists a gap of skills and understanding between domain expert and data scientists.
Security and compliance concerns.

So how do enterprises overcome barriers and break through?

Breaking through towards Enterprise AI

1. Understand the Main Areas of Improvement in Your Business

Those areas for improvement typically exist at a fundamental level. They include:

Core business needs: growth, innovation, etc.
Production: launching product and service lines.
People: making people more productive and valuable.
Operations: reducing costs, streamlining processes, etc.

The important thing to do here is to identify your top challenges for these fundamental areas of the business.

2. Define the Line of Business You Want to Transform First

Which lines of business (LOBs) are the sources of the main areas for improvement you identified above? Which LOBs constitute your core strengths? Where might AI offer the most significant improvement? Such as:

Sales and marketing: AI can help increase revenue, attract customers and users, get into new markets or increase existing market share, and improve customer relationships.
Service and support: AI can scale and elevate your offerings, improve customer satisfaction, retain customers, improve business health, and get insight on product or service performance.
Supply chain and logistics: AI can create sustainable business networks and healthy business-to-business and business-to-consumer ecosystems, optimize your transaction and goods flow, and improve your competitive strength on pricing, cost, and quality.
IT: AI can improve IT functions and operations, keep the competitive edge on your product and service offering, increase productivity, and optimize information sharing and knowledge sharing.

Depending on the industry, you might put more weight on addressing some LOBs over the others. In any event, select the LOBs and key processes that support them, and then develop one or more use cases to improve those LOBs. When you see success from it, you will have an easier time expanding the scope of enterprise AI.

3. Choose the Type of intelligence you want to Improve with AI

As a sequel to a recent blog post “Purpose of intelligence” in the context of AI, the intelligence is key to make a difference. Enterprise AI could adopt these forms of practical intelligence to support your business:

Content intelligence

The business consumes, distributes, and stores enterprise information as content. Examples include business emails; sales contracts; marketing campaigns and advertisement; service logs; and support tickets. AI is applied to get the insight from the information, such as spam and risk management, pattern matching, and smart classification.

Communication intelligence

Communication intelligence is about engaging people in real time using AI technology such as voice, text, and virtual assistants; helping people ask the right questions and reply with the right answers; enhancing personalized experience in customer engagement and employee internal communication; supporting customer churn analysis; and interactive customer support.

Transaction intelligence

In financial or merchant transactions, AI can enhance personalized recommendations, fraud detection and prevention. AI can forecast stock trading; predict demand based on inventory and stock; automate self-checkout or self-booking; and support order tracking and fulfillment, among other functions.

Operation and planning intelligence

For example, AI can support logistics and supply chain management and demand forecasting; inventory tracking and management; product planning, predictive maintenance and repair planning; risk and safety management; transportation and fleet management; network and route optimization; and warehouse robotics.

Object intelligence

With the growing physical objects everywhere, AI is used to capture and detect the situation of the objects themselves and the surrounding environment, such as production line monitoring; keep machines up and running; perform visual inspection for defects; perform identification of parts and commodity products; track moving objects; and do safety and security management for buildings, roads, and cities.

4. Get Help from an Advisor

Many AI initiatives stay at strategy level but never move on to execution. You may need to consider getting advice from a trusted counselor who can address critical issues such as:

Actually creating your strategy to support enterprise AI. Writing about a strategy in a blog post is one thing. Creating a strategy is another.
Evaluating the main areas of improvement apposite business use cases, and identifying the right type of intelligence to apply in practice.
Launching a pilot project focusing on your few chosen use cases or business processes within the intent LOB.
Working with IT to understand your data ecosystem; what type of data is flowing in your system in and out; whether and how the data should be captured and leveraged; the data volume.

Data complexity should not be underestimated. You need right platform and right expertise to manage the end-to-end data pipeline. Once you are successful on the pilot project, you can expand to other areas easily and quickly.

Pact.AI Can Help

If you want to accelerate business growth with AI, to gain intelligence from data and turn data into valuable asset, to augment user experience and bring true value to employees and customers, or to establish your AI vision and drive intelligence innovation, please contact us to learn more.

About the Author:

Yingwu Gao is VP of Product Engineering and AI Practice, heading the enterprise product engineering and innovation including AI, Data Science, and Cloud. Her team plays a vital role in defining and building newer market relevant products and services, such as Enterprise AI, Cloud AI solutions, and Pact.AI platform innovation.