As organizations implement generative AI solutions, intellectual property (IP) is emerging as a clear differentiator. While basic GenAI capabilities are becoming widely available, sustainable competitive advantage comes from integrating unique IP to create solutions that deliver distinctive value.
This integration is reshaping how organizations approach GenAI implementation. In the insights industry, where generic AI tools often lack depth, combining GenAI's processing ability with rich proprietary data and domain expertise enables solutions that deliver meaningful competitive advantages.
Understanding the AI value chain
To understand how organizations can effectively implement this approach, we’ve identified three core components that drive value in AI solutions across global organizations:
- Users: Business professionals seeking insight to develop new approaches and maximize AI system effectiveness.
- GenAI Agents: AI systems that interpret and process user requests. While rapidly evolving, these systems alone offer limited differentiation as they are available to everyone and trained on common data.
- Specialized Skills: The real differentiator is where data, data science expertise, and IP frameworks combine to create unique capabilities.
The third component delivers the value. While GenAI provides language understanding and reasoning, our specialized skills, grounded in proprietary data, models, and frameworks, create a distinctive advantage. We’ve achieved exactly this with Kantar’s AI Assistant (KaiA), where GenAI's language and reasoning capabilities combine with our extensive datasets and specialized analytics to deliver uniquely valuable insights.
The power of custom-trained models
Therefore, the key to unlocking GenAI's potential lies in customizing language models with organization-specific data and vocabulary. This process, known as domain adaptation, lets AI systems understand and generate insights using industry-specific terminology, frameworks, and context that generic models might misinterpret or overlook.
Specifically, these are examples of GenAI ready IP assets that companies own:
- Data Assets: Historical transaction data, customer interaction logs, brand and advertising data and performance metrics that can help train AI models.
- Knowledge Bases: Documented expertise, methodologies, and best practices that can be transformed into AI-readable formats.
- Algorithmic IP: Existing proprietary algorithms and decision frameworks that can operate on some of the data assets and be integrated with GenAI.
- Industry-Specific Taxonomies: Specialist vocabulary and classification systems that can enhance AI comprehension.
Layer-by-layer integration: Building the AI organization
Building effective GenAI solutions involves combining IP across three key layers of skills abstraction: Foundational, Analytic, and Reasoning.
The foundation layer enhances core AI, Machine Learning (ML), or GenAI capabilities through domain-specific training for foundational problems. For example, in the consumer insights industry, these would include developing custom algorithms to model survey responses, or search and social data in natural language, building an algorithmic toolkit to analyze and interpret video advertisements, or generating synthetic and “digital twin” data.
The analytics layer uses these foundational elements (often in combination) to create products that satisfy particular business objectives. An example is LINK AI, our creative effectiveness product that is built on the world's largest database of ad testing comprising over 260,000 tests and 35 million human interactions. This model leverages several foundational algorithms for video and audio analysis and natural language processing along with a custom neural network to predict an ad’s effectiveness accurately and in close to real-time. LIFT ROI is another example in our business – this provides a very granular understanding of the impact of media spend on sales and brand, while explicitly capturing the quality of creative assets.
And finally, enterprise grade applications can also involve a third layer of abstraction – the reasoning layer. Think of this as a way for business users to interact with the analytics and foundational layers in an easy and interactive manner. Business problems do not fit cleanly into silos; they involve identifying, connecting, and executing multiple skills and capabilities from the analytics and foundational layers. The solution is an AI agent that is capable of reasoning, planning, and executing this workflow.
Our AI assistant KaiA showcases this hierarchy of skills. While it harnesses large language models for natural language processing and for chain-of-thought reasoning, its real strength lies in its connection to our proprietary datasets, knowledge base and analytic protocols – in other words, our analytic and foundational layers. This enables our clients to receive precise, actionable insights that outperform generic AI solutions, and thereby optimize their brand performance.
Measuring the success of proprietary IP and GenAI
Although our recent Marketing Trends 2025 report states that 36% of marketers don’t think they or their teams have the skills required, we are seeing organizations that implement IP-driven AI solutions achieve measurable results. These include enhanced campaign performance, improved customer retention through predictive analytics, better marketing ROI via optimized media spend, and premium pricing through strong emotive clarity.
Our own experience in this regard is highly encouraging. LINK AI enhances multimedia campaign performance by up to 30%. Needscope AI allows brands to achieve greater emotional clarity across their touchpoints – our research confirms that brands with strong emotive clarity achieve significantly greater differentiation and are 1.5 times more likely to command premium pricing. LIFT ROI has delivered, for example, a 20%+ increase in ROI and a 40%+ increase in profits due to optimized media mix.
In summary, forward-thinking brands are carefully positioning themselves to create novel solutions that deliver sustained competitive advantage. These solutions always involve a strong suite of proprietary IP.
To learn more about implementing IP-driven AI solutions, contact us to discuss how we can realize your organization's objectives.