PlanYear AI Newsletter - December 2024

And if you’re wondering, no I did not use AI to write this. (Out of principle, this newsletter is human powered) 

Working at an AI company in 2024 has been quite the journey. Each week I open LinkedIn only to be greeted by yet another piece of breaking news. Anthropic releases Claude 3. Google drops Gemini Advanced. Meta makes Llama 3 open source. And now, OpenAI dropped a full version of o, their latest model capable of reasoning.

So, in our latest newsletter, I wanted to take a step back and look at how 2024 went in the world of AI. How did AI develop? What does it mean for brokerage? What should you keep an eye out for in 2025?

How AI developed in 2024: The technology finally delivered on its early promises. 2024 marked a significant evolution in generative AI applications. Multi-modal experiences in AI, where systems combine text, voice, images, and video, came to market. The increasing sophistication of these generative AI models provided companies with entirely new efficiencies and capabilities.

As AI systems evolved, their interactions with humans transformed dramatically. Simple text exchanges gave way to more natural conversations that mirror human communication patterns. A pivotal advancement was the expansion of context windows, which refer to AI's capacity to hold information in its working memory. These expanded windows enabled AI models to analyze significantly more data simultaneously.

This expansion matters because it unlocks more complex capabilities. Modern AI systems can now process entire books, maintain coherent discussions across multiple topics, and create comprehensive long-form content.

The scale of this advancement is striking: in 2024, OpenAI expanded its commercial model's context window to 128,000 tokens from 4,096 - a 31-fold increase.

This cognitive leap laid the foundation for multimodal AI, where systems seamlessly integrate text, voice, images, and video. Models like GPT-4V and Claude 3 now analyze visual data with remarkable precision, whether describing complex images or engaging in natural discussions about visual content. This breakthrough spans industries - from architects analyzing blueprints to medical professionals examining diagnostic imagery.

The AI startup landscape also continued evolving in 2024, with new startups emerging and existing platforms incorporating AI features. Many people now use AI without realizing it - through text summarization, video captioning, and intelligent code completion in their everyday tools. This subtle integration of AI into our digital experience represents a quiet revolution in human-computer interaction, a trend likely to accelerate into 2025.

Favorite prompt technique for 2024:

Here's a practical technique you can start using right now to help improve your output from LLM’s: ask the AI to help you craft better prompts. 

The approach is straightforward - I use this template:

“Knowing everything you know about how LLMs work internally - write me a prompt to give an LLM to accomplish the following: (insert what you’re trying to accomplish in your own words here). Imagine you’re a top 1% expert in your field.”

I then use this AI-crafted prompt in a fresh chat, which consistently seems to give better results. Feel free to try it out for yourself. 

Stats on AI I found interesting from 2024:

• A Harvard study found that management consultants who incorporated AI tools into their work were more productive, completing tasks 25.1% more quickly, an average of 12.2% more tasks in total, and with over 40% higher quality compared to a control group
• According to McKinsey's latest Global Survey on AI, 65 percent of respondents report that their organizations are regularly using generative AI in 2024, nearly doubling the percentage from the previous survey just ten months ago.
• The adoption of AI in corporate settings has doubled since 2020, with 50% of surveyed organizations incorporating AI into at least one business function in 2024.

And with all this impact, governments worldwide took notice. Specifically, 2024 marked a huge increase in efforts to regulate AI, with initiatives like President Biden's executive order on AI and calls for international cooperation to manage AI risks. The state department even commissioned a report warning of potential "extinction-level" threats from advanced AI systems. Head spinning stuff.

What this could mean for brokerage: the single biggest AI-related piece of news related to brokerage was the development of multimodal AI systems. Simply put, AI is now capable of processing quotes from carriers in any formatting. PDF, Excel, tables, and images are now able to be processed by LLM’s.

The implications of this development are almost limitless. 

No longer are firms chained to the carriers via API’s. The playing field has been leveled. If a broker now has the efficiency to spread any carrier quote in any format nearly instantaneously, it becomes easier to provide groups with a greater number of alternatives. This increased competition provides greater optionality and may even result in cost savings for employers and employees. Historically, brokers have been limited in their ability to quote all available carriers due to time constraints, leading to a restricted set of options for employers and less competition between carriers (read: higher rates). By leveraging AI, brokers can efficiently present quotes from a wider range of insurance carriers within their markets. This expanded carrier selection creates more competition, compelling insurance providers to offer more competitive prices and better terms to win business, even in larger market segments.

AI's Impact on Employee Experience

Recent research from Lighthouse Research & Advisory hints at another potential impact of AI in the benefits space. According to their Chief Research Officer Ben Eubanks, AI has the potential to improve health outcomes, increase employee engagement and retention, and revolutionize the way employees negotiate total rewards. 

The implications for benefits engagement are particularly interesting. AI is becoming sophisticated enough to predict when employees might need specific benefits, even before they realize it themselves. For instance, AI systems can analyze claims data to identify pre-diabetic employees and facilitate early interventions, potentially saving both lives and healthcare costs.

This predictive capability extends to employee retention. AI tools can now monitor workplace patterns - from meeting loads to performance signals – and alert managers about potential issues before they escalate. 

Perhaps most intriguingly, AI is beginning to reshape how employees negotiate their total reward package. Eubanks notes that employees actually prefer discussing career mobility and promotions with AI agents rather than human managers initially. This shift in preference signals a broader trend: AI isn't just about processing efficiency – it's becoming an integral part of the employee experience ecosystem.

What should you keep an eye out for in 2025:

I feel incredibly lucky to be working with a company that’s pushing this technology forward in an industry that’s ripe for innovation. If you can ingest any carrier quote in any format, and then integrate that same data throughout your entire renewal process, what else is possible? 

PlanYear held a webinar last week on our 2025 roadmap. Click here to explore the direction generative AI is heading to shape the employee benefits landscape in 2025.

As we see it, the future of employee benefits brokerage may be transformed by AI in several key areas:

• Actuarial analysis has the chance to become more precise and efficient. AI models have the potential to process vast amounts of historical claims data to predict future risk patterns and optimize premium structures.
• Transparency into claims may increase dramatically. AI-powered systems can be designed to instantly read, categorize, and analyze ongoing claims in whatever format the carrier sends them in. This has the potential to reduce the workload for the broker, increase visibility for the employer and identify opportunities for intervention to mitigate high cost claims.
• Billing reconciliation and consolidation may continue to evolve as well. AI can automatically ingest bills, reducing processing times and human error while identifying billing discrepancies that otherwise drive up costs for the employer.
• Compliance management can be streamlined with AI-generated workflows that adapt to changing regulations. These smart systems can create customized compliance checklists, deadline reminders, and required documentation based on each client's specific plan structure and regulatory environment.

The beauty of AI is that it allows for both the flexibility of data input as well as data output. You no longer need an API to ingest data, and you have the flexibility to manipulate said data in any way you can imagine. (Stay tuned, it's going to be an exciting year ahead at PlanYear).

Perhaps most exciting is the emergence of AI agents – specialized AI systems that can work autonomously to accomplish specific tasks. Imagine having a dedicated insurance operations team powered by AI that works 24/7: one agent continuously monitoring policy changes and compliance requirements, another handling routine client inquiries and updates, while a third focuses on data analysis and risk assessment. These agents could coordinate with each other, escalate complex issues to human experts when needed, and learn from each interaction to become more effective over time.

From the perspective of AI specifically, there’s tons of speculation as to where this technology will go from here. I found this article from Nature to be particularly interesting. Yes, AI companies are actively working towards developing machines with human-level intelligence. Do I think this will happen personally? Most likely not in the near future. According to the article, many researchers believe that a different technology beyond LLMs will be necessary to achieve human-level AI.

We’ll have to wait to find out. One thing is for certain - It’s going to be interesting.

Want to learn more about the PlanYear Benefits Platform? Contact us now to learn how you can quickly modernize the employee benefits experience with PlanYear.