In this edition of our “Four Questions with” series we chat with the company’s CEO and Co-founder, Jeremy Jawish about all things agentic AI. We talk about how embracing agentic AI is a natural extension of Shift’s pioneering legacy in insurance AI; the opportunities and agentic AI represents for Shift and its clients; why the future is now when it comes to agentic adoption; and how agentic AI maps to Shift’s IDN initiative.
Shift’s approach to agentic AI is an extension of what has made the company unique since its founding; taking an AI-native approach to solving some of the insurance industry’s biggest challenges. When we began this journey, that meant embedding deep insurance domain expertise into production-grade machine learning and predictive AI to help insurers better fight fraud. We then extended our products to address subrogation, underwriting, fraud, waste and abuse in healthcare and other critical processes. As large language models evolved, we integrated generative AI capabilities into our solutions to give insurers even more ways to use AI to the benefit of their businesses and their customers.
With our latest evolution, which began when we introduced Shift Claims in September of 2025, we've productized and deployed scalable, secure AI agents that execute complex insurance workflows such as intake, investigations and liability determination, among others. And they do this with great accuracy, all while preserving the human-in-the-loop checkpoints, governance, and integration rigor that our customers require. These agents leverage our existing models, scoring engines, APIs and insurance intelligence layer, so customers gain productivity and quality improvements without disruptive rip-and-replace projects.
Under the hood we’re investing in orchestration, monitoring, and advancing our insurance intelligence layer, framing our agent reasoning in business reality. This strategy results in practical, agent-led capabilities that are built on proven domain models, protected by operational controls, and designed to fit into insurers’ IT stacks with clear ROI and explainability.
Making agents the core of our offerings represents a number of opportunities for both our customers and Shift. Agents can help automate complex end-to-end processes, shorten cycle times, continuously gather evidence, and free experts to focus on exceptions. That translates into higher throughput, reduced claims costs, a better customer experience, and new operationally feasible products. For Shift, an “agent first” strategy lets us productize deeper capabilities and deliver differentiated services.
For Insurers, the real opportunity for agents is focused on the workforce. Too often, AI in general, and agentic AI more specifically, is feared as a job killer. The reality is something far different. The insurance industry is facing a widely recognized talent gap. Agentic AI can be the bridge. It automates the mundane and repetitive, allowing employees to instead focus on high value tasks that impact the business and its policyholders. Agents empower those with less experience to benefit from the collective knowledge of the organization to make the best decisions possible. And customers can rely on their claims being settled quickly, accurately, fairly, and perhaps most importantly, with empathy.
Three important forces have converged that make agentic AI viable for insurers. These are model capability, proven business use cases, and operational pressure. Foundation models are at a point where they can sustain multi-step reasoning and tool use, and their capabilities continue to improve at a rapid pace. Carriers have demonstrated repeatable wins in claims and fraud where gains are measurable. And finally, business realities including cost pressures, labor shortages, and customer expectations all make automation a strategic imperative.
In addition, many insurers also have mature data pipelines, APIs, and digital workflows that make safe agent deployment feasible. That said, readiness doesn’t remove complexity. Successful adoption requires disciplined change management strategies, robust guardrails, and partners who can demonstrate both technical and insurance expertise. Where those elements exist, agentic solutions can deliver clear, fast payback on well-chosen workflows.
In so many ways, continued advancement of IDN directly, and positively, impacts what our agents are capable of. AI always performs best when it has access to the best data, and agentic AI is no different. The one thing that I would add about IDN is the need for agents to access tools to complete their tasks. These tools can be enterprise software, APIs, browsers, or other services. IDN not only provides data, but can also act as a tool the agent can access to complete its task.
IDN members have access to an unprecedented amount of cross-carrier claims data. This data can be used to trigger alerts, deliver insights, and provide a comprehensive view of how a claimant, provider, or network is impacting both individual or multiple insurers. An agent’s ability to access the underlying data, as well as the alerts or insights that result from analysis of that data is incredibly powerful. We see continued investment in IDN as directly benefiting our agentic strategy.