Beyond Guesswork: Appier Research Teaches AI to Recognize Its Limits and Choose the Right Reasoning Approach

Enabling AI to identify information gaps and choose the right reasoning language for more reliable enterprise Agentic AI

SINGAPORE, Sept. 8, 2026 /PRNewswire/ -- Appier (TSE: 4180), an AI-native company delivering Agentic AI as a Service (AaaS), continues to advance AI innovation and research while expanding the frontiers of marketing and advertising technology. As Agentic AI becomes increasingly embedded in core enterprise operations, Appier's AI Research team has published two papers examining how large language models (LLMs) can recognize when retrieved information does not support a valid answer and how selecting the appropriate reasoning language can better serve users across linguistic backgrounds — setting a more rigorous benchmark for the trustworthiness and global deployment of enterprise AI.

Both capabilities are increasingly important in real-world business scenarios. Consider an e-commerce service agent: if a shopper asks about a product not covered by an existing return policy, an AI that applies rules for a similar product could provide misleading information and create disputes. Likewise, a game studio expanding overseas may rely on AI to understand local players, but linguistic nuances can shape market preferences and risks. If the AI does not reason in the local language, it may overlook insights that a native speaker would recognize. To support enterprise decision-making effectively, Agentic AI must be able to manage uncertainty and select the reasoning approach and language best suited to each task.

Recognizing Information Gaps: Teaching AI to Answer Honestly When Data Falls Short

To complete tasks autonomously, Agentic AI typically retrieves information from enterprise knowledge bases, documents, or external sources before reasoning and responding. Retrieval-Augmented Generation (RAG) is a common architecture for this process. However, when retrieval fails to provide sufficient information, the model's ability to recognize that gap directly affects the reliability of any subsequent decision.

In "None of the Above, Less of the Right: Parallel Patterns between Humans and LLMs on Multi-Choice Questions Answering," Appier's AI Research team used "None of the Above" (NA) options to simulate scenarios in which no valid answer was available. The team tested 28 leading LLMs of varying sizes and found that model accuracy fell by 30% to 50% when "none of the above" was the correct response. The findings show that even when models possess relevant knowledge, they tend to select suboptimal or incorrect options rather than proactively flagging insufficient information.

This capacity to "know what you don't know" matters most in tasks like business ethics, where models must assess several plausible options holistically, a more complex task than solving problems with directly verifiable answers, such as mathematics.

Appier's AI Research team applied two training methods—Supervised Fine-Tuning (SFT) and Direct Preference Optimization (DPO)—to teach models to recognize "none of the above" scenarios. SFT trains models using correct examples, while DPO exposes them to both correct and incorrect responses so they can learn to distinguish between them. DPO improved model accuracy in identifying questions with no correct answer by nearly 30 percentage points, showing that targeted training can strengthen this capability.

However, "none of the above" is not suitable for every question. It is most effective when answers are clearly defined and the options are mutually independent, suggesting that an AI's ability to withhold an answer should be trained for specific tasks. In enterprise knowledge scenarios, this means improving retrieval accuracy while also introducing a checkpoint to confirm whether sufficient information is available before the AI acts. If the information is insufficient, the system can search again or escalate to a human.

Choosing the Right Reasoning Language for Each Task

Beyond managing uncertainty, Agentic AI also needs to select the right reasoning language for each task. Appier's research paper, "Language Matters: How Do Multilingual Input and Reasoning Paths Affect Large Reasoning Models?" found that the language used by a large reasoning model (LRM) can directly affect its logical reasoning, safety judgments, and cultural understanding. Even when prompted in another language, models often default to high-resource languages such as English for reasoning. For some models, the reasoning language differed from the response language in over 90% of cases.

Using a "text prefilling" technique—setting an opening phrase that prompts the model to reason in a specified language—Appier's AI Research team examined how language choice affected task performance. High-resource languages such as English generally produced stronger results in mathematics and knowledge-based tasks. For tasks requiring cultural understanding, however, reasoning in the local language captured local context more effectively. Safety testing also showed that local-language reasoning was more effective at identifying harmful or illegal queries.

In other words, the optimal reasoning language should be chosen dynamically based on the nature of the task. This research opens the door to "reasoning-language routing": beyond selecting the right model, tools, and data, Agentic AI could, in the future, dynamically select the most suitable reasoning language based on task type, market, and cultural context, while continuing to respond in the user's preferred language.

Beyond Answer Accuracy: Recognizing Limits and Choosing the Right Approach

"These two papers redefine the standard for evaluating AI. As AI moves from answering questions to making autonomous decisions, measuring whether a model produces the correct answer is no longer enough. We must also assess whether it can recognize insufficient information, adjust its actions accordingly, and select the reasoning approach best suited to each task," said Chih Han Yu, CEO and Co-founder of Appier. "These capabilities will help Agentic AI evolve from simply executing instructions into a reliable decision-making system capable of navigating real-world complexity. Through sustained foundational research, Appier aims to turn these critical questions into measurable and improvable AI capabilities, enabling enterprises across markets and languages to adopt Agentic AI with greater confidence."

Looking ahead, Appier's AI Research team will continue advancing research into large language models and Agentic AI while exploring applications across the company's three product lines—Ad Cloud, Personalization Cloud, and Data Cloud. The goal is to enable AI not only to act autonomously but also to make more reliable judgments based on the available information, task requirements, and market context—helping enterprises turn Agentic AI into scalable, tangible business value.

About Appier

Appier (TSE: 4180) is an AI-native Agentic AI as a Service (AaaS) company that empowers business decision-making with cutting-edge AdTech and MarTech solutions. Founded in 2012 with the vision of "Making AI Easy by making software intelligent," Appier endeavors to help businesses turn AI into ROI with its Ad Cloud, Personalization Cloud, and Data Cloud solutions. Now Appier has 17 offices across APAC, the US and EMEA, and is listed on the Tokyo Stock Exchange. Visit www.appier.com for more company information, and visit ir.appier.com/en/ for more IR information.