Onderstaande vijf taken laten zien hoe AI-modellen kunnen helpen bij het verbeteren van marketing en bedrijfsstrategieën. Ze zijn handig voor studenten marketing, marketeers, productmanagers en bedrijfsstrategen die betere inzichten willen krijgen en betere beslissingen willen nemen. Met deze technieken kunnen professionals verder kijken dan alleen de oppervlakte om verborgen redenen te ontdekken, kansen op de markt te vinden en de ervaring van gebruikers te verbeteren.
1. User Motivation Mapping
- Identify Sources of User Feedback: Start by gathering feedback from a variety of online sources. This includes platforms like G2, Capterra, Reddit, ProductHunt, and TrustRadius, which are specifically mentioned. You can also include any other relevant sources where users discuss your product or brand. The key is to collect a broad range of opinions and discussions.
- Use a Reasoning Model: Choose an advanced AI reasoning model such as DeepSeek or ChatGPT. These models are better at understanding the underlying reasons behind user feedback compared to standard models.
- Input the Feedback: Provide the AI model with the collected user feedback. This might involve copying and pasting text from the various platforms into the AI tool.
- Instruct the Model: Give the AI clear instructions. Ask it to analyse the feedback and uncover the deeper motivations behind the discussed features and pain points. Specifically, ask it to identify two or three user archetypes or personas and point out any disconnections. For example, you could ask it to “analyse the most recent user discussions and reviews about Slack from G2, Capterra, Reddit, ProductHunt, and TrustRadius”. Then, ask it to “uncover the deeper motivations behind the top discussed features and pain points” and to “identify two to three user archetypes or personas and any disconnections that it identifies”.
- Analyse the Output: The AI model should provide a detailed explanation of its thinking process. It should go beyond surface-level analysis and propose the hidden psychological drivers. For example, the model might suggest that users who like organised channels may actually desire more control or a sense of belonging. Similarly, users looking for search functionalities might be seeking reliability and security for retrieving older messages.
- Adjust and Refine: Use the insights from the model to fine-tune your marketing message or product development. The goal is to align your brand’s messaging with the true motivations of your users, which the AI has identified.
2. Market Impact Analysis
- Identify the Market Event: Recognise a significant event that could affect the market. This might include a new product launch by a competitor, like OpenAI’s Deep Research. Or, it could be a sudden change in the industry or technology.
- Use a Reasoning Model with Search: Choose a reasoning model, such as DeepSeek and ensure that both thinking and search modes are activated. The search function allows the AI to gather the most up-to-date information, while the reasoning capabilities help to provide deeper analysis.
- Input the Event Details: Provide the AI with details of the market event from various sources. This includes information from the competitor’s blog, social media, and reputable tech news.
- Instruct the Model: Ask the AI to analyse the event and consider the immediate changes that might occur in the market. For example, ask it what immediate changes would be triggered for the AI research product market. Ask it to predict which customer segment will shift the choice first, and what quick wins opportunities you can capitalise on.
- Review the Output: Pay close attention to the AI’s chain of thought details. The model should provide insights into potential market shifts, such as an acceleration of AI agent development. It should identify the customer segments most likely to shift, such as enterprise researchers, and pinpoint opportunities like addressing accuracy and real time fact checking.
- Apply Findings: Use this analysis to inform your strategic planning and decision-making. The aim is to understand how market changes may affect your position and how to take advantage of new openings. For a more well rounded response, use another model like o3 or perplexity combined with reasoning models, as o3 has the ability to scrape forum discussions.
3. Customer Journey Analysis
- Collect Journey Page Data: Gather relevant data from each stage of the customer journey. This includes screenshots or data from the category page, product page, shopping cart page, and payment page of your website.
- Choose an AI Model: Select an appropriate AI model. ChatGPT o1 is mentioned as a good option when using screenshots, as other models may have technical issues with file uploads. However, you can use other models when not uploading files.
- Upload Data: Provide the AI model with the gathered data. This could be the screenshots of each page, along with any important background information. For example, explain to the AI the target customers and the search query used.
- Instruct the Model: Give clear directions to the AI. Ask it to analyse the pages and consider the primary questions a customer might have before proceeding, any missing trust signals, and potential confusion points on each page.
- Evaluate the Analysis: The AI model should give a detailed breakdown of customer questions, potential weaknesses, missing trust signals, and confusion points. For example, it might find that the product page is missing dimensions, or a return policy.
- Improve the Journey: Use the insights gained to address any issues and improve the overall customer experience. The goal is to make the journey as smooth and intuitive as possible, based on the AI’s assessment. You can make the analysis more thorough by uploading additional data like funnel statistics and persona details.
4. Competitive Positioning Analysis
- Gather Competitor Information: Collect information from your competitor’s website, including their homepage, feature page, pricing page, and case studies. Use a URL-to-text tool to convert the webpage content into pasteable text.
- Use a Reasoning Model: Select an AI reasoning model, like DeepSeek or o3.
- Input the Information: Paste the competitor’s website copy into the AI model.
- Instruct the Model: Ask the AI to analyse the copy, and uncover your competitor’s core beliefs and assumptions about their target customers. Request that it identifies their beliefs about customer needs, why they believe customers need the solution now, and how they expect customer needs to change.
- Analyse the Output: Review the AI’s interpretation of the competitor’s core beliefs and expectations. The AI may identify that a competitor believes customers need a unified platform.
- Identify Gaps and Opportunities: Ask the AI to identify specific user needs that are not addressed by the competitor, and any missing connections. For example, the AI might suggest that the competitor is not addressing needs such as role-specific AI, or customizable AI models.
- Position your Brand: Use this analysis to better position your brand, identify points of differentiation, and fill the identified gaps. The aim is to refine your messaging and product offerings to take advantage of the areas where your competitor is lacking.
5. User Experience Test
- Identify a Specific Page or Element: Select a product page or landing page on your website that you want to optimise.
- Use a Reasoning Model with Search: Use an AI reasoning model like ChatGPT o3 with search functions enabled. (ChatGPT uses this by default).
- Provide Context to the Model: Give the AI the URL for the specific page you want it to analyse, along with context about the target customer. For example, you could specify that the target customer is a family with kids.
- Instruct the Model: Ask the AI to assess the page and think from the customer’s perspective. Request that it identifies potential concerns that the customer may have. Then ask it to check whether the page elements address those concerns, and to propose three A/B tests to improve the user experience.
- Review the Analysis: Examine the AI’s analysis to see the potential customer scenarios, like concerns about spills and stains. Check the identified shortcomings, such as missing family-focused messaging, and the proposed A/B tests.
- Implement Testing: Use the recommendations to implement A/B tests to improve the user experience. This might include testing different messaging, including family-focused content, or adding an FAQ section to address potential customer concerns. You can make the response more insightful by including additional data like persona behaviour, and heatmaps.
By following these detailed steps, you can effectively use AI reasoning models to enhance your marketing strategies and improve your business outcomes.
Source: a YouTube video transcript from the channel “Grace Leung”. The video is titled “The AI Hype is a Distraction: What Actually Matters (+ Best Ways to use ChatGPT o3 & DeepSeek)“. I *highly* encourage you to follow Ms. Leung.