Exploring AI for Your Company? Where to Start

AI advancements are making headlines daily. It seems AI is everywhere, and you might feel a fear of missing out. As a business executive, understanding the opportunities and threats AI presents to your company is crucial. But where do you begin?

With everyone talking about AI—especially after the rise of ChatGPT and conversational language models—enthusiasm is high. However, these discussions don’t help answer the immediate question you face: “How do I explore AI for my company?” This post aims to help.

As a boutique technology consulting firm specializing in AI businesses, products, and solutions, we categorize each new customer or project into one of three groups. Each category has unique challenges and requires a tailored approach.

Our project allocation is straightforward and based on two simple questions:

• Is the problem known?
• Is the solution known?

Consulting: When Both Problem and Solution Are Unknown

When a client is unsure about both the problem they want to solve and how AI can help, we categorize the project under consulting. A typical request might be:

“We are very interested in AI but have no idea how to apply it to our business.”

At this stage, it’s important to understand how AI differs from other technologies. Like databases, cloud computing, or mobile devices, AI can help automate processes, reduce costs, speed up time to market, and improve the bottom line. However, AI offers a unique advantage: it can enhance your competitive edge. Properly designed and implemented, AI technologies can make your products harder for competitors to copy.

So, you need to decide whether you’re aiming to improve the bottom line, gain a competitive advantage, or both.

If your goal is to improve the bottom line and you are less concerned about competitive advantage, you should consider buying rather than building. Using third-party APIs and off-the-shelf AI products can offer faster time to market and reduced costs.

For these projects, consultants can help you:

• Audit your current processes and solutions
• Define the AI business case
• Understand the current AI product landscape and shortlist candidate solutions
• Select the best development partner
• Plan and support the implementation project

If you’re interested in gaining a competitive advantage, it’s crucial to design AI features around your product’s perceived value and user journeys. Modern AI relies on machine learning; models, including large language models (LLMs), improve with more training data. You need to create feedback loops in your business:

1. User Engagement Generates Data: The more users engage with your product, the more data you collect.
2. Models Improve with Data: Increased data leads to better models.
3. Better Models Enhance the Product: Improved models make your product more attractive, bringing in more users.

By building your product around this principle, you create a barrier for competitors. Consultants can assist you to:

• Identify opportunities to build a competitive advantage
• Review your user journeys
• Design AI models and training approaches
• Develop AI feedback loops, monitoring, and retraining strategies

Solution Design: When the Problem Is Known but the Solution Is Unknown

When a client knows the problem but is unsure of the solution, we categorize the project under Solution Design. A typical request might be:

“We would like to build an AI legal assistant for labor law across European Union countries.”

If you are ideating a new AI product, it may be worthwhile to validate and refine your value proposition with the market. Consultants can help you run the entire customer development process:

• Defining initial value proposition hypotheses
• Recruiting potential users
• Conducting customer development interviews
• Synthesizing, refining, and iterating

If you are confident in your product design and product-market fit, the next step is to design the product itself—as both a user journey and a technical solution. Consultants can help you:

• Design user journeys so your product captures the right data
• Select machine learning models that fit your needs and limitations
• Architect the technical solution using modern cloud platforms
• Plan the implementation and deployment project and select the right development partner

Implementation: When Both Problem and Solution Are Known

When both the problem and the solution are clear, we categorize the project under Implementation. A typical request might be:

“We need to build a system that automatically configures the delivery of marketing messages. We are using AWS Bedrock and Anthropic Claude. The system will also use LangChain for the LLM part.”

In this scenario, success hinges on effective project planning and execution. For AI projects, it’s crucial to embrace an iterative approach, incorporate user feedback, and have a strong and detailed risk management plan.

Consultants can assist you by:

• Auditing the current project stage
• Planning and overseeing project execution
• Joining your team for hands-on work

Projects often move from one category to another. For example, we might start in the Consulting stage by designing several potential AI business cases for competitive advantage. After prioritizing one or two, we proceed to the Solution Design stage to develop detailed product designs and system architectures. Finally, we move on to the Implementation stage to develop and launch the solution.

Having a structured approach to managing emerging technologies like AI is crucial for ensuring project success and return on investment. We developed this framework from our own experiences—learning from both mistakes and successes—and it is now an integral part of our daily work.

Related Articles

• LLM Practitioner’s Guide: Gemma, a Game-Changing Multilingual LLM
• Why Do You Need an AI Strategy? AI as a Competitive Advantage
• Four AI Designs That Don’t Work