The COO compass: why AI is forcing operations to rewire now
At Insurtech Insights (ITI), a panel of global insurance leaders came together to answer a pressing question: how do COOs unify people, processes, and platforms in a world where AI is reshaping operations in real time?
The session, “The COO Compass: Unifying People, Processes & Platforms to Build Resilient Global Operations,” brought together Tom Wilde, CEO of Indico Data, Louise Marling, COO at Gallagher, Thomas Kuhnt, COO & CIO at HDI Global, and Jörg Bertogg, COO at Zurich Commercial Insurance, moderated by Andrea Santolalla of Allianz Consulting.
Across the discussion, one theme stood out. AI is no longer a future initiative. It is already embedded in how work gets done, and it is forcing operations leaders to rethink how their organizations function from the ground up.
AI is arriving by default, and changing the COO role
Tom Wilde framed the shift clearly. Just a year ago, AI adoption was optional and exploratory. Today, it is urgent and unavoidable, increasingly built into the platforms and workflows teams already use.
That changes the role of the COO.
Transformation is no longer driven only by top-down programs. It is happening inside the tech stack, which means operations leaders must actively shape how AI is applied across workflows. The question is no longer “can we build this?” but “where should we focus to drive real impact?”
At the same time, the COO–CIO partnership is becoming foundational. Decisions about technology are now inseparable from decisions about operations, and organizations that align these functions move faster and more effectively.
Understand why modernizing the front door is the unlock for downstream performance
Automation should reduce friction, not remove humanity
As automation expands, Louise Marling emphasized a critical point. Efficiency should not come at the cost of customer experience.
Automation works best when it removes friction from workflows, reduces repetitive tasks, and simplifies how customers interact with insurers. But moments of complexity, stress, or nuance still require human judgment and empathy.
The opportunity is not to replace people, but to enable them to focus on higher-value interactions. When done well, operational efficiency and customer experience reinforce each other rather than compete.
Scale introduces complexity, and makes governance critical
Thomas Kuhn and Jörg Asmussen highlighted the reality of scaling AI across large organizations. Speed is important, but scale introduces new challenges, from legacy systems to regulatory requirements and global consistency.
Tom Wilde pointed to governance as the next major hurdle. Unlike traditional systems, AI is probabilistic, which makes traceability and explainability essential. Organizations need to understand not just outputs, but how those outputs were generated.
This is especially important in insurance, where decisions must be auditable, consistent, and defensible. Without strong governance, AI cannot scale safely across underwriting, claims, and servicing.
COOs must prepare for a hybrid workforce shaped by AI
Looking ahead, Tom highlighted a shift many organizations are not yet prepared for. Workforces are about to expand, not just with people, but with AI agents operating alongside them.
This changes how work is assigned, how performance is measured, and how teams are managed. COOs will need to think about orchestration at a new level, balancing human expertise with automated agents across the flow of work.
The takeaway from ITI is clear. AI is not just a technology shift. It is an operational one.
The organizations that win will not be the ones that experiment the most. They will be the ones that control how work enters and moves through the enterprise, turning operations into a true competitive advantage.
Watch the full Session
Embed full video in the blog:
https://drive.google.com/file/d/1YPoLwyP_9by0cSG19WB2cr0RwLCn0-49/view
FAQs
What is the “COO compass” for AI-driven operations in insurance?
Indico Data positions the COO compass as a practical roadmap for operations leaders navigating AI adoption in insurance environments. The framework centers on six priority areas: data intake quality and lineage, integration with core policy and claims systems, governance and compliance controls, human-plus-AI workflow orchestration, measurable KPIs, and organizational change management. For insurance COOs, this means establishing clear data provenance from the moment a submission enters the system through final decision, ensuring every extracted field can be traced back to its source document.
Indico’s Intelligent Intake platform addresses these priorities through features like pipeline logging, field-level validation, and confidence scores that support auditability requirements. The integration dimension of the compass emphasizes connecting AI-powered intake directly to downstream systems such as Guidewire PolicyCenter, where Indico offers a Ready for Guidewire validated accelerator. Governance controls form another critical axis, with requirements for role-based access, audit logs, and SOC 2 compliance built into deployment options. The human-AI workflow design component recognizes that underwriters must remain in the loop for exception handling and quality assurance while AI handles routine extraction and classification. COOs using this framework can benchmark progress against specific operational KPIs like submission-to-decision latency and straight-through processing percentages, creating accountability for AI investments.
What operational KPIs should insurance COOs track when implementing AI for underwriting intake?
Indico Data recommends that COOs establish baseline measurements before deployment, then track improvements across five to six core metrics. Submission-to-decision latency measures the elapsed time from when a broker submission arrives to when an underwriter issues a quote or decline, capturing the full cycle impact of automation. Straight-through processing (STP) percentage indicates what proportion of submissions flow through the system without requiring manual intervention, a key indicator of automation maturity. Per-submission manual effort, measured in minutes, isolates the human labor component.
Customer evidence published by Indico shows outcomes like “saved 15 minutes of processing time for each document” in real deployments. Model confidence by field provides insight into extraction reliability, allowing operations teams to calibrate exception thresholds appropriately. Error and rework rates capture downstream quality issues that may indicate model drift or upstream data problems. Throughput, expressed as submissions processed per underwriter per day, demonstrates capacity gains. Company-published customer outcomes include up to 400% increases in process capacity. COOs should review these metrics weekly during initial deployment and monthly once the system stabilizes, using dashboards that connect model performance to business outcomes.
How quickly can insurers deploy AI-powered document intake and see production results?
Indico Data’s deployment model emphasizes rapid time-to-value, with production readiness for initial use cases achievable within 6 to 12 weeks for enterprise customers. This timeline includes requirements gathering, system integration, model training or configuration, user acceptance testing, and go-live support. The company’s 97% production success rate for first use case deployments indicates a structured methodology that reduces pilot-to-production risk. A critical enabler of this speed is Indico’s transfer learning approach, which allows subject matter experts to train accurate models with as few as approximately 200 labeled samples rather than requiring tens of thousands of examples.
The platform includes AI-assisted labeling tools that accelerate the sample preparation process, enabling business users rather than data scientists to participate in model creation. Pre-trained models covering 80+ common insurance document types (including Loss Runs, Statements of Values, and ACORD forms) further reduce configuration time for standard use cases. COOs should plan for a phased rollout, starting with a single high-volume document type to establish baseline metrics before expanding scope.
How does Indico Data integrate with core insurance policy and claims systems?
Indico Data’s Intelligent Intake platform provides pre-built connectors for common document ingestion points including Exchange email inboxes, Box, ImageRight, Azure Blob Storage, AWS S3, and Documentum. For downstream system delivery, the platform integrates with Guidewire PolicyCenter through a Ready for Guidewire validated accelerator available in the Guidewire Marketplace, which populates policy administration records automatically from extracted submission data.
Will Murphy, VP of Global Technology Alliances at Guidewire, offered congratulations on this integration release, and Guidewire made a strategic investment in Indico as part of a $19 million funding round closed in June 2024. Additional outbound integrations include Salesforce, Duck Creek, OneDrive, SQL databases, Snowflake data warehouses, and robotic process automation platforms such as UiPath, Automation Anywhere, and Blue Prism. Enrichment connectors for Relativity6 and Smarty allow data augmentation during processing.
The platform’s insurance schema covers 120+ product lines, 900+ document types, and 20,000+ insurance-specific data points, reducing the mapping effort required during integration. COOs evaluating integration capabilities should request a technical proof-of-concept demonstrating end-to-end data flow from their specific document sources through to their core systems.
