Taffrail

How do we use a rules engine to create expert and transparent advice? In our last post, we went over the surprising reason why digital brokers won’t answer your questions and how Regulation Best Interest permits brokers to do just that. To tackle this misunderstanding and clear the way for digital brokers to give clients clear and helpful answers to their investing questions, Taffrail made a rules engine designed to provide personalized financial advice to self-directed investors. Rather than having to sift through logic puzzles on the IRS website or the generic educational content currently offered by brokers, our advice provides a holistic overview of action steps a self-directed investor can take. Taffrail’s expert system can answer client questions with accuracy and transparency. To ensure self-directed investors are only asked questions by the system relevant to the advice they need, we have a team with a deep understanding of the financial markets, investment strategies, tax laws, and privacy. Questions like the one we mentioned in our last post, “What type of IRA should I use?” can be answered based on what makes the most sense for a user rather than solely what they may be eligible for.

Adding Real Intelligence to Artificial Intelligence Generative AI is excellent at guessing a user's intent. The user would prefer that their AI chatbot didn't also guess the answer. For the thousands of financial questions where experts know the correct answer, wouldn't it be better to give that answer to the user? An easy solution is to train your Generative AI model with a rules engine like Taffrail's. That way, when your AI chatbot interprets the user's intent, it can rely on an expert system that will always give a precise and accurate answer.

As we wrap up 2023 at Taffrail, we express our gratitude to our clients and partners. Thank you for your support and collaboration throughout the year. It has been wonderful to work with you, and we look forward to continuing to bring investors trustworthy and personalized digital advice. From all of us to all of you, have a joyful holiday season! 

How to turn Generative AI into precision AI Generative AI would be even more valuable in Wealth Management if we could trust it. Clients and Investors want answers to their financial questions. Unfortunately, GenAI doesn’t know when it gives the wrong answer, and Wealth Management firms and their compliance departments don’t like giving clients incorrect responses. Generative AI is in the machine learning branch of AI. As the name implies, and as NVIDIA’s stock price confirms, a GenAI Large Language Model (LLM) needs thousands of machines utilizing the latest technology to learn all the patterns of text that it can find. An LLM doesn’t really know what it is saying, though. It doesn't understand any concepts and excels at bluffing. It will say 2 + 2 = 4 because it read it somewhere, but it doesn't know math. If GenAI is machine learning, it would be logical that there would be another type of AI where you can teach the machine to know what to say. That type of AI is called a “rules-based system” or an “expert system.” As the names imply, experts program the rules in a rules-based system to give correct answers every time. These systems work well in answering questions with known answers, like “2+ 2 = ?” or “Should I open a Roth or a Traditional IRA?”. Generative AI is astonishing at figuring out what the user is asking, that is, determining their intent. When a GenAI model learns what question a user is asking, it could answer with a hidden and seamless handoff to an expert system. Wealth Management firms already use this technique when they respond to a chatbot question with a chart, table, or document. When you use generative AI to determine the user’s intent but use another system to answer, you can be confident that the answer is correct. At Taffrail, we developed an expert system that can answer thousands of client questions with transparent, deterministic, and provable answers. Our rules make it so that your generative AI systems can give trusted and compliant answers on topics that, until now, would have been considered out of bounds. #generativeAI #chatgpt #wealthmanagement #fintech

AI doesn’t have to be the enemy of perfection. The aphorism that “perfect is the enemy of good” makes sense for most systems, but probably not for financial advice. When an investor asks ChatGPT a question that has a definitively right or wrong answer, everyone involved would prefer the investor gets the right answer, regardless of how plausible or confident the wrong answer might sound. Generative AI is the newest type of AI that has captured our imagination with its ability to generate human-sounding answers quickly and cheaply. Unfortunately, Large Language Models often give the wrong answer—some reports indicate these “hallucinations” happen 29% of the time.  A rules-based system is another type of AI that uses pre-programmed rules to always give correct answers. The only disadvantage of a rules-based system, or an “expert system,” is that until now, it has been prohibitively expensive to program all the rules. In Wealth Management, Taffrail is proving it is less expensive to write rules in an expert system to give the correct answer than it is to filter the prompts of generative AI to avoid giving the wrong answer. The solution is straightforward. Firms can train their generative AI to use Taffrail’s expert system to give personalized and accurate answers where possible. Taffrail has a library of thousands of reliable and transparent rules crafted by humans who have a deep understanding of the financial markets and investment strategies, tax laws, and privacy. Get in touch today—we’d love to give you a demo. #generativeAI #chatgpt #wealthmanagement #fintech

Generative AI is not a solution for automated financial advice. AI systems like ChatGPT, Bing AI, or Google Bard, are incapable of providing both factually correct and personalized answers to complex financial questions. Since AI works by guessing the next words in a sequence, it doesn't have pre-scripted or predetermined logic, based on factual sources, built-in. Not to mention, ChatGPT struggles with math and failed the CPA exam just two weeks ago. Taffrail is the leading provider of personalized digital advice for the financial brokerage industry. The advice provided by Taffrail is risk-free and always correct, without the hallucinations or inaccuracy of generative AI. #ai #fintech https://lnkd.in/einrQNJA