Walmart AI: 100x Productivity... | AviaryAI Newsletter
Walmart's disclosure that AI has enabled 100x productivity improvements in certain operational functions is a landmark data point -- and for credit union operations leaders, it raises an urgent question: where in your institution could AI deliver comparable gains? The answer depends on where your team spends the most time on routine, repeatable tasks -- and for most credit unions, member outreach, document handling, and staff knowledge retrieval top that list.
Walmart's AI Supercharges Productivity 100x
Walmart CEO Doug McMillon revealed that their use of generative AI in managing the product catalog has dramatically boosted efficiency. The AI-driven process improved over 850 million data points, a task that would have required 100 times more employees to complete in the same timeframe without AI assistance.
So what?
This staggering productivity gain demonstrates AI's potential to revolutionize operations, particularly in data-heavy tasks like loan processing, compliance reporting, or member data management. The key is identifying repetitive, time-consuming processes where AI could free up staff to focus on high-value, member-facing activities.
Read the full story here
Gemini Live: Google's AI Assistant Gets a Voice
Google has launched Gemini Live, a voice chat mode for Gemini Advanced subscribers. This new feature allows for natural, conversational interactions with the AI assistant, including the ability to interrupt and resume conversations. With 10 new voices to choose from and plans for iOS support and more languages, Gemini Live aims to make AI interactions more seamless and personalized.
So what?
This marks a significant step towards truly integrated AI voice assistants in our daily lives. We're moving closer to a future where these AI voice agents can handle complex, multi-step tasks across various aspects, all interfaced by speaking a few words. Financial institutions can already leverage similar technology with AviaryAI’s outbound voice agents. Learn how to provide automated proactive service with AI voice here.
FCC Proposes AI Voice Disclosure Rule for Robocalls
The FCC recently proposed new regulations to combat AI-generated robocalls from bad actors. The plan would require AI voices to disclose their artificial nature at the start of calls and texts, aiming to protect consumers from deceptive practices and potential fraud.
So what?
Those partnering with forward-thinking providers like AviaryAI are already ahead of the curve, as our AI agents proactively identify themselves. This compliance-first approach not only aligns with upcoming regulations but also builds trust with members, increasing engagement and effectiveness of AI-powered communication channels.
THE AI EXECUTIVE'S HANDBOOK
One simplified GenAI concept per week to build your AI Acumen
Mitigating AI Risk with New MIT Database
The new AI Risk Repository created by MIT is a powerful tool that can help you identify, understand, and manage these risks effectively. Let's break down how you can use this resource to enhance your AI strategy.
What is the AI Risk Repository?
The AI Risk Repository is a comprehensive database and classification system for AI risks. It categorizes risks based on their causes and the domains they affect. With over 777 risks identified, it provides a clear roadmap to navigate the complex landscape of AI-related threats.
How can this help you?
- Understand the big picture with breakdowns on the cause of each risk and their domains they impact (discrimination, privacy, malicious use, etc.)
- Assess and mitigate risks relevant to your business using the filters available
- Prioritize your efforts by determining which risks are most pressing and how they manifest in your industry.
- Develop better policies and strategies by leveraging the insights provided in the repository
How to use the Repository:
If you're short on time, start with the Plain Language Summary on page 3 of the Repository document. It provides a quick overview of what's inside.
- Access the Repository here: https://airisk.mit.edu/
- Identify relevant risks:
- Use the Domain Taxonomy to locate risks specific to financial institutions such as privacy breaches, misinformation, and cyberattacks.
- Narrow risks down by using the Causality filter to understand the variations of this risk from different sources.
- Analyze Relevant Risks:
- For each identified risk, read the detailed descriptions to understand the specific issues and scenarios.
- Asses the potential impact of these risks on your institution’s operations.
- Develop Mitigation Strategies:
- Create or update internal policies to address identified risks. For example, revise data handling procedures to mitigate privacy threats
- Deploy technical and procedural safeguards based on the new/updated policies.
By incorporating the AI Risk Repository into your strategy, you can minimize risks and leverage AI more effectively, ensuring your institution remains ahead in this rapidly evolving technological landscape.
Training Data
Training data is like a big collection of examples that we use to teach a computer how to do something. Just like you might learn by looking at lots of solved math problems, an AI learns by looking at tons of data. This data could be pictures, text, or any other kind of information related to what we want the AI to do. The computer studies this data to find patterns and rules, which it then uses to make decisions or create new things on its own. The more good-quality training data we give it, the better the AI usually becomes at its job.
CFPB Takes Aim at AI Customer Service:
Is Your Institution Ready?
The Consumer Financial Protection Bureau (CFPB) is set to release new rules regarding the use of chatbots on company websites as part of the Biden administration's "Time is Money" initiative. This move aims to address consumer frustrations with excessive paperwork, long hold times, and general aggravation in customer service interactions. While the CFPB has expressed concerns about chatbot limitations, industry experts argue that complaints specifically related to chatbots represent a tiny fraction of overall customer service issues.
So What?
If your institution is already using AI chatbots, it's time to take a hard look at your setup. First, review your disclosure practices – are you clearly telling members when they're talking to a bot? Next, assess how seamlessly members can switch from the bot to a human. If it's not a smooth transition, fix it now. Audit your chatbot's responses to ensure they're accurate and up-to-date with your latest policies and products.
Finally, consider forming a small team to stay on top of these upcoming regulations. Being proactive now could save you headaches later and might even give you a leg up on larger, slower-moving competitors. The goal isn't just compliance – it's using these changes to build a better, more trustworthy digital experience for your members.
Frequently Asked Questions
Is Walmart's 100x productivity claim realistic for credit unions?
For specific, well-defined tasks -- like processing member inquiries that follow predictable patterns -- AI can deliver productivity multiples of 10x-100x. Across an entire institution, realistic productivity gains from AI are typically 20-40% in targeted areas. The Walmart example applies to a narrow, optimized use case.
What is MIT's AI risk framework and how does it apply to financial institutions?
MIT's AI risk framework provides a structured approach to evaluating the risks of AI deployment -- including bias, reliability, privacy, and accountability. Credit unions can use it to build internal governance documentation that satisfies NCUA vendor management requirements.
What credit union tasks are most suitable for AI automation?
The highest-ROI AI automation targets for credit unions are outbound member communication (welcome calls, card activation, payment reminders), staff knowledge retrieval from policy documents, and member onboarding follow-up -- all areas where AviaryAI's platform delivers measurable results.
