DeepSeek’s AI Breakthroughs Transforming Multiple Industries ?
This year, DeepSeek has made waves in the global stock market by announcing the development of an advanced artificial intelligence system for a surprisingly low investment of $6 million. While some industry experts have raised questions about the surprisingly low expenditure related to developing these AI systems, there is a consensus that DeepSeek has significantly slashed operational costs associated with running robust AI models. By providing its technology for free, the company has initiated a shift in the AI landscape.
The Insurance Revolution with Roadzen ?
Global companies are evaluating how DeepSeek’s AI solutions could reshape their financial frameworks and operational practices. For instance, Roadzen, a public company on Nasdaq, is attempting to revolutionize the auto insurance industry leveraging artificial intelligence. Rohan Malhotra, CEO of Roadzen, highlights that their AI offering enables clients in the insurance space to cut down the time taken to solve 80% of minor accident claims from six weeks to just two minutes.
Subscribe to our Social Media for Exclusive Crypto News and Insights 24/7!
- Insurance claims often involve sensitive information.
- Predicting total expenses accurately for insurance can be complex.
Due to these factors, Roadzen previously limited itself to using a select few intricate AI models known for their precision. These included offerings from well-known AI companies like OpenAI, Anthropic, and Meta. However, everything changed with the introduction of DeepSeek’s R1 model.
Accuracy is Paramount ?
Malhotra, who possesses a Master’s in Robotics from Carnegie Mellon University, cited that the quality of outputs generated by DeepSeek’s R1 matches up with OpenAI’s most advanced model, the o1. Additionally, DeepSeek’s solutions offer cost advantages for Roadzen. For example, the company processed over 600,000 insurance claims over a recent three-month period, which requires significant token usage. A token represents the smallest data unit used by an AI model - approximately 750 words equate to 1,000 tokens.
- Using OpenAI’s model, the quarterly AI costs amounted to $36,455.
- In comparison, DeepSeek’s technology reduced that to just $17,012.
This represented a reduction of 50% in typical AI costs per claim, showing how significant financial savings can be achieved through utilizing DeepSeek’s model. Furthermore, additional expenses for adjusting or training AI models can also be drastically reduced when using DeepSeek, from over $21,000 with OpenAI down to just above $10,000 with DeepSeek.
The Open-Source Advantage ?
Industry experts have pointed out that DeepSeek’s landmark decision to offer its reasoning model as open-source makes it particularly appealing. This contrasts sharply with existing open-source options like Meta’s Llama. Arli Charles Mujkic, the CEO of Ooda AI-an AI platform based in Sweden-mentioned that they integrated DeepSeek’s technology immediately upon its release. Ooda AI operates a digital marketplace allowing users to select the most effective AI model for specific tasks.
- Ooda AI generates revenue through subscriptions, usage fees, and enterprise contracts.
- Mujkic noted that DeepSeek’s V3 model operates about 20% better than Meta’s previous iterations.
With a client roster that includes major health insurers, Ooda AI anticipates that employing DeepSeek’s models will considerably diminish costs while boosting overall revenue, since savings will proportionately benefit their customers as well.
Unprecedented Cost Reductions ?
Neal K. Shah, who leads CareYaya, a service dedicated to assisting clients with health insurance claims, shared his enthusiasm for DeepSeek’s offerings. Shah revealed that the integration of DeepSeek’s technology allowed his firm to reduce costs related to appealing health insurance claim denials by nearly 90%. With previous vendors, costs stood at almost 12 cents per appeal, now dramatically falling to just 2 cents through DeepSeek’s capabilities.
Shifting Dynamics in AI Costs ?
Despite substantial declines in AI service pricing over the last two years, companies are not envisioning a similar decrease in the cost for end users. According to Malhotra from Roadzen, AI costs presently constitute only a small fraction of the approximately $150 charged for each claim in developed markets. Most investments focus on areas like research and development, in addition to connecting traditional systems with advanced AI solutions.
Looking forward, reduced expenses in AI technology could facilitate automation within emerging markets, where labor costs currently remain competitive against AI solutions. Malhotra summarized this shift, stating that once inference costs decrease sufficiently, global deployment becomes feasible.
Sources
Source 1, Source 2, Source 3, Source 4, Source 5, Source 6
