7 Questions to Form an AI/Gen AI Strategic Journey Map

I argued that technology-enabled transformation through AI/Gen AI is a journey rather than a sprint .

What makes up the journey-map?

Instead of providing a list of specific recommendations, it may be more helpful to pose a series of questions to guide the development of a strategic journey map, considering the unique context and history of each organization and industry.

Question 1: Who are we?

Unless you are a startup, your organization has a valuable identity shaped by its history and culture.

I own an Ovation guitar. Originally, Ovation was an aerospace company that developed a deep knowledge of reducing vibration in aerospace. For guitars, they used this knowledge to amplify vibration instead. It's the same underlying identity, just applied differently. It's often more effective to seek out a more suitable application than to redefine the core identity.

At Integreon , we developed valuable AI/Gen AI-based solutions. Our sales team sometimes wonders if we are becoming a software company. My answer is no. I love software and had worked at a software company as the Chief Operating Officer (COO). However, I recognize our core identity and culture of Integreon is a high-quality legal and business creative services solution company. We are advancing our tech-enabled innovation as part of our cultural values.

Question 2: How do we product manage our AI/Gen AI solutions?

As part of product management practice, articulating products in specific applications is essential to making the vision and strategy tangible. Instead of the classic roadmaps (unless you're building a physical rocket with physical dependencies), many product managers employ strategic product spider charts and agile OKR (objectives and key results) product management practices. This approach involves connecting the vision and goals to the initiatives, releases and features more flexibly and adaptively.

Question 3: How do we understand the market evolution deeply and continuously?

I often receive inquiries about AI and general AI solutions in the market. Evaluating each solution in this manner can lead to a superficial comparison. This approach may result in impressive PowerPoint presentations, but it usually does not answer the fundamental questions of why and how it matters.

The alternative approach is to carefully analyze what is important and why by integrating customer feedback with the help of sales and marketing leaders, along with inputs from the delivery team. The insights from sales and delivery experts, combined with direct customer input, usually create the best scorecard for conducting a market analysis of available solutions. Once the market scan is complete, we can then move on to define our market position and approach to developing and marketing our products and solutions.

Question 4: How do we acquire the AI/Gen AI skillset in a market of severe talent shortage?

Successfully implementing AI/Gen AI solutions requires skilled developers and operators. Our CEO, Subroto Mukerji , once gave an analogy for this. He compared it to retraining coachmen to become drivers of modern cars.

Learning occurs through both classroom instruction and hands-on, on-the-job training. Typically, a few hours of awareness sessions are not enough. We need to consider more deeply how to cultivate a significant number of individuals to reach advanced and expert levels so that those at the foundational level can progress and drive the organization forward.

Question 5: How do we enhance risk management practices?

AI and Generative AI can introduce new risks. Enhancement of risk management practices includes developing and improving policies, strengthening internal audits, seeking and obtaining external audits and certifications, managing third-party risks, and upgrading to new technologies to automate and create a more resilient risk environment.

Question 6: How do we go about Buy, Partner, and Build decisions?

With Gen AI, customers usually are not looking for generic LLMs (large language models). They want the industry-specific LLM. Why? In our experience, while Gen AI LLMs from Google, FaceBook, or Open AI, Claude or Mitra can be acceptable for personal use, for professional use with domain-specific questions, the accuracy and relevance usually starts around 50-60%. A tuning and training phase is required to improve the accuracy to above 85 - 95% or higher and to integrate Gen AI into an end-to-end process to ensure the end outcome is on par with or better than the human process.

Horizontal solutions that set industry norms lend themselves to buying. However, if you want to be a first mover, like in our case around a few service offerings, where we have deep domain knowledge, it makes sense to build, to lead the market. A partnership can accelerate growth and increase the overall value for customers and partners.

How you make these buy, partner, and build decisions is essential to sustaining the AI/Gen AI journey.

Question 7: What is our cloud strategy?

Even if you pursue an onpremise LLM strategy, as your deployment grows, you will need access to cloud GPUs to scale economically. You would need comprehensive security architecture to protect your data. Among the big hyper-scalers, you may also need to decide who is primary and secondary to concentrate your focus.

When formulating a cloud strategy, you need to consider cloud hosting options, information security standards, management approaches, and carefully plan networking architecture for hybrid cloud solutions.

These are the essential questions that must be answered with execution sustained to win this AI/Gen AI transformation race. It is much more than RAG (Retrieval Augmented Generation) or LLM.