Key Insights at a Glance

• AI creates impact in insurtech when it improves real business decisions, not when it’s deployed as a standalone tool.

• The biggest early mistake is starting with technology instead of defining clear outcome-driven use cases.

• Real value comes from combining AI reasoning with automation and orchestration to execute end-to-end processes.

• Human oversight, governance, and strong organizational readiness remain essential for scaling AI responsibly.

Introduction 

Artificial intelligence is no longer a far-off promise for insurance. Today’s leaders are wrestling with AI not because it’s novel, but because it actually works. From underwriting to customer service, the latest advances are unlocking automation and insight pipelines that were previously science fiction.

Still, the difference between hype and impact matters. In insurtech – and especially embedded insurance – success isn’t defined by which tools you adopt. It’s defined by how you embed AI into real decisions that shape customer outcomes and business metrics.

In this article, we unpack why AI is taking off in insurance, what most companies get wrong at first, and how a practical strategy rooted in outcomes can drive real value. 

1. AI Is Already Moving From Pilots to Production

Recent research from McKinsey underscores that AI has officially entered insurance industry operations. According to The Future of AI in the Insurance Industry, only a subset of carriers has captured outsized value from AI so far, but those that do show enterprise-wide strategy and deep integration outperform others. Moving beyond pilot projects toward scaled AI is what unlocks competitive advantage.

This isn’t just about adding one chatbot or automation script. McKinsey reports that AI leaders achieve significantly better results in underwriting, claims, pricing, and customer service because they integrate AI into core workflows, not isolated tasks. 

Meanwhile, industry automation research emphasizes another theme: AI is most powerful when it is connected with automation and orchestration. Reports like The State of Automation in Insurance argue that advanced AI – what many call agentic automation – goes beyond recommendation to execution within complex, multi-step workflows. 

Together, these studies suggest that merely experimenting with AI isn’t enough. The companies that get value are those that think about AI and automation as parts of broader decision frameworks, not separate tools.

2. The First Mistake: Starting With Tools Instead of Decisions

When leadership teams first discuss AI, the questions often sound familiar:

• “Where can we use AI?”

• “Should we add a chatbot into claims?”

• “Can we automate fraud detection next?”

This is understandable as a starting point but they often lead to scattered pilots that deliver minor wins and no long-term momentum.

McKinsey’s research highlights that scattered, un-prioritized pilots rarely create competitive advantage. Leaders who succeed define clear strategic domains where AI will be deployed, such as underwriting, customer engagement, or claims workflows. 

This “tool first” approach also limits AI’s impact because it treats AI as an add-on, rather than as a foundation for reshaping how decisions are made and executed. Successful AI initiatives start with identifying business areas that are AI-friendly – with defined governance, workflows and clear decision authorities, not the technology itself.

3. The Real Power Comes at the Intersection of Automation + AI

One of the most important clarifications in the automation literature is that AI alone has limits. Traditional automation or RPA (robotic process automation) provides operational leverage by moving work without deep reasoning. AI provides reasoning. But real work happens when reasoning is connected to execution.

This idea is central to the concept of agentic automation – systems that don’t just suggest actions but can perform work within governance boundaries, orchestrating across systems and workflows, as highlighted in UiPath’s State of Automation in Insurance report.

What this means for insurtech companies:

• AI integrated with orchestration platforms can automate multi-step processes – such as evaluation, approval, communication, and updating of records – in ways that reduce manual handoffs.

• Decision improvement becomes measurable through outcomes like conversion rate, claims cycle time, or eligibility accuracy.

• Agentic systems preserve human oversight while scaling repetitive decision patterns, making governance and auditability essential from day one.

• Organizational readiness also plays a key role: strong prioritization, reusable data foundations, and agile teams help AI initiatives move from experimentation to measurable business value faster.

This combination – automation, AI reasoning, and organizational alignment – is what moves AI from smart insights into strategic business impact.

4. Humans Still Matter: AI Augments, Not Replaces

A persistent misconception is that AI will replace human judgment entirely. In fact, both McKinsey and automation research point to a “human-plus-AI” model where humans retain oversight, especially for edge cases and governance.

AI excels at patterns, scale, and consistency. Humans excel at judgment, context nuance, and accountability. Together, they create a hybrid workflow that delivers outcomes that neither could achieve alone.

Importantly, governance frameworks and auditability are not optional. As agentic systems gain autonomy, transparent controls are critical for compliance and trust. 

5. Result Come From Focusing on Decisions That Matter

So where should AI start in insurance?

Success, seen in industry research and real-world deployments, comes from anchoring AI on fundamental decision points that shape customer experience and economic outcomes. However, in a B2B embedded insurance context, not all domains should be tackled at the same time. Prioritization matters.

Start with: AI-friendly workflows
For embedded platforms, claims triage and service engagement are often the strongest starting points. These areas sit closest to the end user experience and allow teams to deliver measurable improvements quickly without deeply restructuring carrier-side risk models.

• Claims triage and routing: automating the initial screening and sorting of claims while escalating complex cases.

• Service engagement: delivering proactive, contextually relevant interactions that reduce churn and support partners at scale.

Advance to: Deeper risk and commercial optimization
Once governance, data flows, and orchestration are mature, AI can expand into more complex domains that require tighter carrier alignment and regulatory oversight:

• Underwriting accuracy and speed: reducing time to issue and expanding risk appetite with confidence.

• Pricing personalization: adapting offers in real time based on behavior, risk signals, and context.

McKinsey’s work highlights that the companies showing the most value from AI are those that treat it as part of end-to-end domain transformation, not isolated technology sandboxes.

In embedded insurance, every one of these decisions lives inside someone else’s product experience. That’s an opportunity. But it also raises the stakes. Misaligned AI outputs can directly affect conversion, trust, and revenue because there’s no separate insurance screen to cushion the impact.

6. What Differentiates AI Winners in Insurtech

When we look across carriers and platforms that are gaining advantage, a few themes stand out:

• They define strategic AI domains and link them to business metrics.

• They integrate AI with automation and orchestration to drive measurable impact.

• They treat AI as part of process redesign, not technology adjunct.

• They build and invest in governance from day one.

For embedded insurance platforms – where AI decisions live in checkout flows, policy issuance, and claims pathways – excellence in these areas turns technology investment into real revenue and customer trust.

Conclusion: The First AI Decision Is Organizational

AI in insurance is not a question of if anymore. It’s a question of how.

The companies that extract value from AI aren’t just experimenting; they are reimagining decisions, workflows, and outcomes. They move beyond demos, minor pilots, and top level diktats.  They invest in integrated, enterprise AI strategies that align with business goals.

For insurtech – and embedded insurance – this means starting not with tools, but with intent and outcomes. What decision will AI improve first? How will that decision scale? What organizational changes support that transformation?

Answering these questions is where the real value of using AI starts to materialize.