Marketing Strategies in the AI Era - Management Perspectives and Practical Steps to Build a Growth Engine

Customers are now using AI tools like ChatGPT to compare products and services. This signifies a "change in customer touchpoints." While discussions around the introduction of AI into advertising operations and content creation have generally focused on "efficiency," marketing in the AI era must consider both "operational efficiency" and "changes in customer touchpoints."

This article will primarily discuss the "changes in customer touchpoints" brought about by the emergence of AI search. This includes not only efficiency in Google Ads and Meta Ads but also the advancement of personalized ad delivery.

The Rise of AI Search

In marketing strategies of the AI era, the normalization of AI searches like ChatGPT and Gemini has accelerated, making AI search measures (LLMO, AIO, AISEO, GEO) another key focus. The starting point is to organize business goals and company challenges, and to select areas to tackle while considering expected effects, execution burdens, and available data.

This article introduces four evaluation axes for reviewing strategies, changes in purchasing behavior, risks of failure during implementation, and four execution steps. For individual AI initiatives, you can also check out examples of AI marketing and implementation methods.

What is Different About Marketing Strategies in the AI Era?

In short, the difference from traditional marketing is that marketing strategies have become more complex.

Previously, search rankings and various advertising metrics were set as KPIs, allowing for a broadening of personas. However, in an era where AI search is commonplace, touchpoints have shifted to AI, search demand/variations have diversified, and we have entered a time when search intentions cannot be fully summarized.

 

It is essential not to make the use of AI itself the goal but to redesign roles to improve customer decision-making and business outcomes. According to a survey conducted by Deloitte Digital in 2024, 29% of brands have implemented generative AI in their marketing operations. 【Reference】

Four Evaluation Axes to Assess Your Current Situation

We will look at ① Response to AIO, ② Real-time Personalization, ③ N1 Customer Understanding, and ④ Continuous PDCA. These are four perspectives to find areas for improvement.

1. Response to AIO: Is Your Information Accurately Communicated in AI Search?

AIO (AI Optimization): An initiative to design information so that company information is accurately cited and recommended in AI searches like ChatGPT and Google AI Overviews.

In traditional SEO, keywords and backlinks were the main targets for measures. In AIO, we add the perspective of how well the questions (prompts) of search users in AI search align with our company content.

For example, responding to conditional questions like "Are there any recommended coffee makers under 8,000 yen for the latest model?" from "Recommended coffee makers."

• Does your company name or service name appear in the responses of ChatGPT, Gemini, or Google AI Overviews?

• Is there primary information on your company site regarding prices, conditions, features, etc.?

• Does the main text directly answer the expected customer search prompts?

Structured data (like Schema.org) is effective for SEO, but it is not a sufficient condition for being cited in AI search. First, it is essential to prepare content that answers users' questions and verifiable primary information.

For specific definitions and countermeasures for LLMO and differences from SEO, please refer to another article.

2. Real-time Personalization: Are You Improving Recommendations?

Real-time personalization: A method of presenting information and products tailored to each user by utilizing purchase history, behavioral data, and recent search intentions.

This cannot be judged solely by "whether you are collecting data." We need to see if the three stages of acquisition, recommendation, and verification are connected. It is not only us who analyze and improve using data generated by AI. For those running ads, it is also applicable to consider that the internal AI of platforms (Google and Meta) optimizes ad displays.

• Data Acquisition: How frequently can you obtain browsing history, search history, purchase history, and inquiry content?

• Recommendation Logic: Are past data and recent dialogues/search intentions linked to the characteristics of products or services?

• Verification and Improvement: Can you confirm reactions to clicks and purchases and provide feedback on recommendations?

It is essential to review recommendations based on the frequency of data updates and to handle personal information appropriately.

3. N1 Analysis: Do You Understand Customer Psychology That Is Not Reflected in Numbers?

N1 analysisA method that delves into the background, hesitations, and deciding factors of a specific customer from awareness to comparison to purchase in chronological order.

AI excels at pattern extraction from large data sets. On the other hand, feelings like "why did it become necessary?" or "what were the confusions?" need to be interpreted by humans from statements made during interviews, business discussions, or inquiry records.

It is important not to mix the facts spoken by customers with the interpretations and hypotheses of the analyst. After confirming the relationship with actual behavior, this information should be reflected in advertising expressions, creative concepts, and the brand's worldview.

4. PDCA: Can You Verify AI Outputs and Continue to Improve?

PDCA (Plan-Do-Check-Act): A cycle of setting objectives and KPIs, executing measures, checking results, and making improvements.

Taking AI search measures as an example, we will confirm that the following five stages are connected in a continuous flow.

1. Understanding the Current Situation: Quantifying exposure status in AI search.

2. Selecting Search Prompts: Identifying question phrases that customers are likely to use.

3. Strategy Design: Deciding the target AI and the context in which exposure is desired.

4. Content Creation: Organizing primary information and improving pages to match questions.

5. Tracking Improvement: Monthly confirmation of citation status by search prompt and AI engine.

AI can also experience hallucinations (the phenomenon of generating responses that differ from facts). The final fact-checking and adjustments should be handled by humans. Items to confirm before starting AI search measures are summarized in the LLMO Current Status Diagnosis Checklist.

How Will Customer Purchasing Behavior Change with the Expansion of AI Search and SNS?

Information gathering is no longer just about "opening sites one by one from search results." Customers make judgments while navigating through SNS, word-of-mouth, videos, comparison sites, and AI searches. A linear purchasing funnel alone cannot capture this influence.

Reconstructing the Customer Journey with Multiple Touchpoints in Mind

Customers are considering products and services while moving between multiple information sources such as AI searches, SNS, word-of-mouth, and comparison sites. Therefore, it is crucial for companies to understand the flow of information gathering from awareness to purchase, not just a single touchpoint.

This way of organizing the connections between information touchpoints and their influence on purchasing decisions is called an "influence map." Specifically, we will confirm the following three points.

①Understanding Touchpoints: Confirming which information sources customers are using, such as AI search, SNS, and word-of-mouth.

② Understanding Questions at Each Consideration Stage: Organizing what customers want to know at each stage of awareness, comparison, and purchase.

③Analyzing Influences on Purchase: Confirming which information touchpoints are involved in comparison and purchase decisions.

Queue Corporation divides the customer's consideration stages into TOF (Awareness) → MOF (Comparison) → BOF (Purchase) and organizes natural language questions that customers are likely to input into AI searches at each stage.

Marketers Transition from Executors to Strategic Decision Makers

As AI can handle A/B testing, bid adjustments, and routine copy creation, humans are required to focus on strategy design. The three necessary abilities are the ability to consider return on investment, cross-departmental coordination skills, and literacy to verify AI outputs.

• Return on Investment Thinking: Confirming ROI and deciding to scale from small successes.

• Producer-like Coordination Skills: Facilitating consensus across departments and handling uncertain decision-making.

• AI Verification and Improvement Skills: Not taking outputs at face value but reflecting them against customer understanding and brand strategy.

Marketing is an activity linked to management decisions. 　

Therefore, in the marketing domain, it is necessary to emphasize not only the skills to operate AI tools but also the ability to make investment decisions, coordinate across departments, and properly evaluate AI outputs.

Three Failure Risks to Check Before Implementing AI Tools

While AI implementation can be expected to yield benefits, there are burdens that may be discovered after operations begin.

The three failure risks to check before implementing AI tools are "over-reliance on data," "cost overruns," and "privacy and ethics."

Are You Losing Sight of Customer Intent by Relying Too Much on Data?

If recommendations are made solely based on past behaviors, they may feel imposed on the individual. Confusions and frustrations that are not reflected in numbers should be supplemented with interviews and inquiry logs. Continuously ask, "Does this recommendation truly solve the problem?"

Have You Estimated Not Only Initial Costs but Also Operational and Personnel Costs?

The costs of AI tools are not limited to initial implementation fees. There are also costs for operations, organizing training data, and securing specialized personnel.

• Avoid company-wide implementation and start with measurable areas, such as optimizing email distribution or recommending a single product category.

• Decide on KPIs and the investment recovery period before implementation, defining "success."

• If specialized personnel are lacking, compare the costs of external partners with in-house production.

The cost trends and contract considerations specific to AI search measures are detailed in LLMO Countermeasures Cost and Pricing Structure.

Can You Properly Manage Personal Data and AI Recommendations?

As the accuracy of personalization increases, the amount of personal data collected also rises. In addition to compliance with personal information protection laws and GDPR (EU General Data Protection Regulation), unify within the company the purpose of use and consent, audits for unfair recommendations based on attributes, and policies for retention periods and deletion.

Four Steps to Execute AI Marketing Strategies

After inspecting the evaluation axes, determine the connection between challenges and outcomes, and conduct small-scale validations. Here, we will introduce four steps centered on AI search measures, but advertising, production, and analysis should also be judged using the same "goal → responsible party → verification" framework.

Step 1: Clarify Challenges Through Current Situation Analysis

Visualize which of your marketing measures are functioning and which are lacking in the AI search environment.

• Does your company name or service name appear in the answers of major AIs?

• In what context are competitors being recommended?

• Are there existing SEO measures that lead to AI search traffic?

• To what extent do the expected search prompts align with your company content?

If you want to know the current situation, please try the free AI Search Current Situation Analysis Report.

 

Step 2: Set KGI, KPI, and Achievement Deadlines

 KGI (Key Goal Indicator) measures the final outcome, while KPI (Key Performance Indicator) measures the process to achieve that outcome.  

Examples of goals include "Doubling inquiries from AI searches within six months" as a KGI, and "Achieving a company mention rate of over 50% among 20 major search prompts" and "Increasing monthly site traffic from AI searches by 30%."

Also, confirm whether there are continuous mentions across multiple AI engines and search prompts. For comparing AI search measurement tools, you can refer to LLMO Tools for Measuring Citations and Brand Recommendations.

Step 3: Separate Responsibilities Between AI and Humans to Execute Measures

For content creation, customer analysis, ad operations, and AI search measures, separate the processes supported by AI from human judgments.

 

When utilizing specialized tools, proceed with the four cycles of "Diagnosis → Design → Improvement → Monitoring," checking citations and mentions by AI. Specific methods for creating and revising AI searches are covered in LLMO Countermeasures Implementation Methods.

Step 4: Connect Effect Measurement to Next Improvements

After execution, confirm the changes by search prompt and AI engine monthly, and decide whether to continue or review.

• AI Citation Rate: The proportion of your company being cited or mentioned in responses among the set questions, as treated by the original article.

• Exposure Stability Rate: An indicator of whether mentions for the same question are temporary or ongoing.

• Contribution to CV: Whether traffic from AI searches led to inquiries or purchases.

AI mentions and visits to the website or CV are separate phenomena. Record the numbers without confusion, and humans will monitor to ensure that AI is not misrepresenting your company.

What Kind of Support Does Queue Corporation Provide?

If you have challenges such as not being mentioned in AI searches or only competitors being recommended, external support may be an option. On the other hand, companies with an in-house LLM engineering team that handles large language models (LLM) and can complete everything from analysis to revisions may find the need for external support relatively low.

Queue Corporation offers the AI search countermeasure service " umoren.ai." The achievements mentioned in the text include support for over 100 companies, an AI citation improvement rate of +460%, a 4.4 times improvement in CV from AI search traffic, and a customer satisfaction rate of 98%.

In support, we design with an awareness of the process by which AI reads, compares, and adopts information based on insights related to LLM engineering, such as RAG (Retrieval-Augmented Generation), Embedding (vector embedding), and Tokenizer (tokenizer).

Frequently Asked Questions About Marketing Strategies in the AI Era

Q. Where should I start reviewing the strategic framework?

Start by reviewing purchasing behavior, focusing on what customers look at, the order in which they compare, and what influences their decisions. Since AI searches, SNS, and word-of-mouth overlap, it is important not to rely solely on existing purchasing funnels. If necessary, also identify the questions customers might pose to AI.

Q. What is the difference between SEO, ad operations, and AIO?

The primary optimization targets differ. SEO focuses on visibility in search results, ad operations on click-through rates (CTR) and conversion rates (CVR), while AIO focuses on how information is handled in AI responses. AIO involves designing information with an awareness of the meaning and intent of questions (Cosine Similarity: a measure of similarity between vectors). The coexistence strategy of SEO and LLMO is also a useful reference.

Q. Where should I start if there are no specialized personnel in-house?

Before making large investments, visualize the current situation within measurable limits and start from areas where there is potential for results. For AI search, you can check for challenges using the free current situation analysis diagnosis. The approach of confirming small successes before establishing a system leads to better cost and risk management.

Conclusion: Define Roles for AI and Humans and Continue to Improve

Combining AI's strengths in data processing and automation with human responsibilities in customer understanding, brand building, and investment decisions is essential, as is achieving a state where your company is recommended in AI searches. First, confirm business goals and customer behaviors, and identify areas for improvement.

Then, considering implementation costs and privacy, cycle through the four steps of current situation analysis → goal setting → measure execution → effect measurement in small increments. AI search measures are one of the options to pursue when deemed necessary for your company's customer touchpoints.