AI iQ: Asking the Right Questions in AI Deployments

Introducing AI IQ, a new column from the founder and CEO of Qantm AI and former chief AI officer of IBM

Seth Dobrin

November 14, 2023

6 Min Read
Credit: Qantm AI

The most common mistake organizations make when evaluating emerging technologies like generative AI is to focus on the capabilities of the technology rather than on how it can support overall business goals and strategies. It is easy to get caught up in the hype and potential of any new shiny object, such as generative AI, without considering what specific use cases are relevant given the company's current challenges and opportunities.

The allure of generative AI is understandable − the promise of being able to interact in natural language with an AI system and having it do high-value tasks such as building creative content like blogs and images; turning your disaster of a shared drive into institutional knowledge; creating software programs using natural language; providing better human interactions with customers, employees, and partners and so much more.

However, these technical capabilities alone are not enough to magically make organizations more effective and efficient. For any technology implementations to truly drive value, companies must start with the core elements of their business strategy and identify where technology can create strategic advantage – and AI and generative AI will inevitably be part of the solution. 

Strategy before technology

Begin by clearly defining the organization's objectives, priorities and desired outcomes for the commitments made to your shareholders or constituents. Start at the top − with revenue, profit and other financial obligations − and then define the specific outcomes that will drive them for your situation. Are you running hot on spending and need to reduce operating costs so you can reach your profit target? Do you need more revenue goals and need to drive more business? Are your customers leaving you so fast that it affects your top and bottom lines while increasing customer acquisition costs?

You must ask these questions honestly and clearly before you move on to the myriad use cases where you can apply AI and generative AI. Once the key goals are established, examine each business process and function to determine where and how AI could drive meaningful progress toward those aims. Look beyond the hype and take a pragmatic approach grounded in business needs rather than technological possibility.

This strategic alignment ensures that any technology project, including those in generative AI, has relevant scope and clear measures for success from the start. It helps secure buy-in across the organization and also aids in change management. With a solid business context, AI initiatives are positioned as enablers for achieving strategic goals that benefit the entire company. AI becomes a powerful means to execute strategy, not just an end in itself.

While generative AI holds tremendous promise, realizing its full potential starts with strategy, not technology. Once an organization has clarity on its objectives and priorities, it can assess how and where to leverage AI to drive impact and value. Technology enables and powers strategy, not the other way around. Keep this principle at the forefront when integrating emerging technologies like generative AI into your business.

Understand your business objectives 

Once an organization has committed to a strategic approach to generative AI adoption, the next critical step is to turn the high-level business outcomes into strategic decisions that define specific use cases. Too often, companies go after AI solutions without comprehensively analyzing where the problems and opportunities truly lie within their current business processes and functions. They know AI is important but have yet to diagnose where it can drive the most impact.

Begin by asking probing questions to identify priorities and desired outcomes across all facets of the business:

  • What are our key sources of revenue, and how could AI make these more predictable and repeatable? 

  • Where are our most significant costs, and how could AI drive efficiency? 

  • What pain points detract from the customer experience, and how could AI ease or enhance interactions?

  • What processes are cumbersome or inconsistent, and could AI introduce automation or consistency?

  • What data-driven insights could propel our innovation efforts if analyzed at scale?

  • What core capability can differentiate you from your competitors?

As this internal analysis identifies objectives and needs, avoid falling into the trap of applying AI just because it is currently popular. Only some business processes require AI to improve. Look for the most significant and highest-value opportunities where AI's capabilities are likely to be implemented and are well-matched to your organization's needs. 

Next, dig into priorities such as customer experience, evaluating each touchpoint and interaction. Map the customer journey to pinpoint frustrating pain points like long wait times. Assess how humans currently handle tasks and make decisions to highlight areas for improvement or consistency. Look for ways to enhance rather than replace human capabilities and judgment. The goal is not to automate every function but to determine where AI could augment human performance and decision-making.

This thorough internal analysis and mapping to objectives helps build an AI roadmap tailored to your organization and grounded in real potential value. You can then explore relevant AI methods like machine learning, natural language processing and computer vision to address the identified needs. The technology supports the strategy rather than drives it.

The message is clear − know your business objectives and needs first, then explore how AI can address them. Do not let the hype push you into AI for AI's sake. Stay focused on how it can create strategic advantage, given your organization's unique goals and constraints. This pragmatic approach is the key to realizing AI's total value.

Evaluate AI capabilities in context

Once an organization has a clear understanding of its business objectives and needs, the next step is to evaluate specific AI and generative AI capabilities in the context of those goals. It is essential to have a realistic and nuanced perspective on what current AI technologies can and cannot do before determining how they may apply.

There are undoubtedly impressive feats of AI today in areas like computer vision, speech recognition and natural language processing. However, these techniques also have significant limitations. For example, machine learning models require massive training datasets and are prone to bias and opacity. Tasks like comprehending casual speech or perceiving 3D spaces remain challenging for AI. We are far from artificial general intelligence.

When assessing the ability of AI to help achieve a business goal, key factors to examine include the following:

  • Data availability - AI relies on quality training data relevant to the task. If internal data is insufficient, costs of acquisition and labeling must be considered.

  • Model accuracy - No model is 100% accurate. Carefully evaluate performance metrics and minimum acceptable thresholds.

  • Interpretability - Can the AI be understood and errors diagnosed? For many applications like loan approvals, interpretability is critical.

  • Security and compliance - AI brings risks that must be addressed, especially in regulated industries like finance and health care.

  • Impact on stakeholders - Assess the effects of AI on customers, employees, suppliers and partners. Avoid inequitable outcomes.

  • Technical infrastructure - AI requires high-performance computing capabilities, which can get expensive.

  • Maintainability - Building and retraining models requires specialized skills that must be developed or acquired.

By realistically appraising both the pros and cons of AI for a specific use case, organizations can make informed adoption decisions and mitigate risks. Moving forward without considering limitations often leads to suboptimal results and disillusionment. 

The core message is not to expect miracles from AI. Evaluate capabilities in the context of your goals and constraints. Take an eyes-wide-open perspective, neither over-hyping nor dismissing potential value. This pragmatic, nuanced approach enables successfully leveraging AI where it can drive real strategic impact.

About the Author

Seth Dobrin

Seth Dobrin is the founder and CEO of Qantm AI and the former chief AI officer at IBM.

Keep up with the ever-evolving AI landscape
Unlock exclusive AI content by subscribing to our newsletter!!

You May Also Like