Transitioning from Dialogue to Implementation: Four Critical Insights on Generative AI

​Generative AI: From Obscurity to Business Driver
The concept of generative AI was virtually unknown just a year ago. However, the launch of OpenAI's ChatGPT changed the game. With over 100 million users within a staggering 64 days, ChatGPT sparked a major shift in conversations. Business teams and their IT counterparts went from "what is it?" to exploring how to leverage this technology. The impact wasn't limited to internal discussions at Microsoft. Customer-facing teams also saw a surge in interest, with generative AI becoming a prominent topic in conversations and major projects.
This initial excitement has evolved.  Microsoft, through its customer interactions, has identified valuable lessons to help businesses maximize the return on investment (ROI) from generative AI. The article delves into four key areas to consider when exploring this technology within your organization.


1. Define Your Problem, Not Just Play with the Tech
Don't get caught up in the novelty of generative AI. Resist the urge to simply experiment without a clear purpose. Instead, focus on the specific business problems you're trying to solve. Clearly define what "success" looks like for these problems. This will help you "paint the vision" and guide your use of generative AI.
Identify and Prioritize Use Cases
Start by brainstorming potential applications, or use cases, where generative AI could be beneficial. Evaluating these use cases more closely might reveal alternative solutions for some. You might also find that certain use cases share similarities, allowing you to group them for a more efficient approach.
Prioritization is Key: Impact vs. Effort
A crucial step is prioritizing which use cases to tackle first. Consider the potential business impact of each use case against the effort required to implement it. This helps you focus on areas that deliver the most value for your investment. Microsoft uses a simple yet effective method: an "Impact vs. Effort" matrix. Each potential use case is placed on the matrix based on its estimated effort (low or high) and its anticipated business impact (low or high). This visual approach helps prioritize projects effectively.​

Prioritization is Key: Impact vs. Effort
A crucial step is prioritizing which use cases to tackle first. Consider the potential business impact of each use case against the effort required to implement it. This helps you focus on areas that deliver the most value for your investment. Microsoft uses a simple yet effective method: an "Impact vs. Effort" matrix. Each potential use case is placed on the matrix based on its estimated effort (low or high) and its anticipated business impact (low or high). This visual approach helps prioritize projects effectively.
Assessing Impact and Effort
When evaluating impact, consider factors beyond just immediate benefits. Think about how generative AI could lead to faster completion times, improved decision-making, and cost savings for your business. However, also consider the timeframe for realizing these benefits.
Effort, on the other hand, refers to the time and money required to implement the use case. Be mindful of the different systems and personnel involved. This will help you realistically assess the resources needed to achieve the desired outcome.
Defining Success with Measurable Goals
Once you've identified your target use case(s), it's vital to define clear success metrics. What does "good" look like in this context?  Establish quantifiable measures based on time, effort, or cost savings to determine the effectiveness of your generative AI implementation. The more specific and measurable your goals, the better you can track progress and gauge success.
2. Start Internally: Test and Refine Before Launch
While it might seem counterintuitive for a customer-focused organization, starting with internal testing can be highly beneficial. Generative AI is a new technology, and piloting it within your company offers several advantages.
Firstly, it allows you to test the technology and refine your approach in a controlled environment. Your own internal teams act as a "friendly audience" for experimentation. This helps identify potential challenges and refine your use of generative AI before deploying it to external customers.
Secondly, internal testing lays the groundwork for a comprehensive generative AI roadmap. This roadmap can be structured using a "Crawl-Walk-Run" approach, with three distinct phases:
Phase 1: Crawl - Building the Foundation
The crawl phase focuses on internal use cases with "human-in-the-loop" reviews of generated content. This is a safe and controlled way to gain experience. Here are some potential use cases for retail and consumer goods companies:

• Summarization and Analysis: Utilize generative AI to summarize call center interactions or customer feedback.
• Categorization: Organize internal knowledge management systems or streamline communication processes.
• Content Generation: Create product descriptions based on existing data or generate product images for your e-commerce platform.

Stage 2: Progression
Scenarios involving direct engagement with staff and clientele
At this stage, there could be oversight over numerous scenarios by humans, each aimed at furnishing employees or clients with insights for more sound decision-making. Examples in the retail and consumer goods domain might encompass:

• Aggregation and evaluation—compiling summaries of product critiques, descriptions, and features. CarMax's method of generating their second-hand vehicle listings—which has saved a decade-plus in content creation efforts—stands as a notable instance.
• Organization—examining trends on social media, conducting sentiment evaluations, or suggesting product design enhancements.
• Creation of content—crafting customized user experiences or marketing materials.

Stage 3: Acceleration
Enhanced automation
This concluding stage emphasizes novel product introductions or automation intended for direct interaction with your clients. In the context of retail or consumer goods, examples might involve:

• Aggregation and evaluation—automation of call centers and contextual automation of processes interfacing with customers.
• Organization—streamlining the generation of product descriptions on your e-commerce site by integrating customer feedback.
• Creation of content—automating the production of tailored marketing and branding materials.

3. Lead with Business-Oriented Solutions
Focusing on use cases with an emphasis on their potential impact and required effort ensures that your efforts are concentrated on solving problems that offer tangible value to your organization. Visualizing your objectives and what you aim to accomplish is critical in this process.
Continuous feedback from business units is crucial. As with any initiative where technology is a major component, the most effective projects are those where there is a strong partnership between the technological and business sides, both aligned towards addressing a specific organizational issue.
For technology-driven projects, including those involving generative AI, it’s vital to consider the synergy between people, processes, and technology. Thus, it's essential to pay attention to each aspect. For instance, in your efforts related to specific use cases, it's important to:
MICROSOFT RESPONSIBLE AI
Discover our insights
Identify the business processes linked to your use case. Comprehend what triggers these processes and understand the interactions with various systems and stakeholders.
Clarify the roles of individuals involved in the process and recognize who will be affected by the successful implementation and utilization of your use case in the operational setting.
Determine how involved individuals will engage with the process—clarify the timing and methods for notifying them about necessary interactions or approvals. This aspect gains importance when considering the principles of responsible AI, specifically accountability.
Examine your use case through the lens of responsible AI. Utilizing tools such as the Microsoft Responsible Impact Assessment can streamline this aspect.
Give thorough consideration to change management. Knowing the individuals impacted is key to facilitating this process. The novelty of generative AI and the extensive coverage it receives in the media mean there are sensitivities to navigate. Applying and respecting Microsoft’s responsible AI principles, including fairness, reliability and safety, privacy and security, inclusiveness, transparency, and accountability, is beneficial in addressing these challenges.

4. Prepare Your Team for Rapid Adaptation
The enthusiasm for generative AI's potential is well-founded—with McKinsey estimating that it could contribute $2.6 trillion to $4.4 trillion annually to the global economy. These figures are particularly noteworthy, especially when considering that the GDP of the United Kingdom stood at $3.1 trillion in 2021. Therefore, leveraging this technology for innovation is crucial.
This requires fostering skill development within your organization to quickly implement use cases. Cultivating your team's skills and capabilities is essential in this process.
There may be a temptation to adopt a 'fast follower' strategy, where you quickly emulate a competitor's generative AI-powered solution that benefits their customers, possibly improving upon it.
This strategy is viable, but it demands that in-house skills and expertise be developed over time to effectively compete. Focusing on the use cases you have identified and are innovating upon is among the most effective methods to develop the necessary in-house expertise.
ChatGPT's achievement of reaching 100 million users within just 64 days, establishing it as the quickest-growing consumer app in history, underscores the potential for generative AI to be the fastest-expanding technology segment ever witnessed.
As the realm of possibilities and advantages continues to rapidly evolve, attentively considering these aspects will enable your organization to learn and effectively leverage the capabilities of generative AI. This strategic approach not only enhances your business's value but also positions you as a top ERP system provider or one of the leading ERP suppliers, adept at navigating the swiftly changing technological landscape.