At a recent ITC London session, Tom Wilde, CEO of Indico Data, joined Arslan Hannani, Chief Innovation Officer at Aviva, for a candid discussion about the state of insurance modernization.
The conversation did not focus on the latest AI hype cycle. It focused on something more fundamental: why so many modernization efforts have failed to materially improve operational performance.
Despite billions invested in policy administration systems, claims cores, workbenches, analytics platforms, and now GenAI, insurers are still wrestling with slow cycle times, constrained capacity, and stubborn expense ratios.
The reason is structural.
Modernization efforts have largely ignored the one layer that governs everything else: how work enters the enterprise.
The breakdown before the workflow
Wilde described the core issue as architectural. More than 90% of inbound insurance work still arrives as unstructured submissions, loss runs, spreadsheets, broker emails, claims packets, and attachments. None of it is standardized. Much of it is incomplete. All of it requires manual effort before a core system can act.
Instead of systems governing work, people do.
Operations teams act as the integration layer between inboxes and core platforms. They extract data, validate fields, cross-check attachments, and manually route work to the right systems and teams. That upstream friction constrains throughput, breaks automation, and erodes underwriting and claims capacity.
Workflows assume clean inputs. Insurance rarely provides them.
As outlined in Indico’s strategic point of view, this is a flow-of-work problem that sits before every workflow, every system, and every decision.
Hannani emphasized that this upstream friction impacts everything downstream, from speed-to-quote to claims cycle time to regulatory defensibility. When intake is inconsistent, automation breaks. When data is incomplete, AI stalls. When routing is manual, scale is constrained.
Why downstream modernization hasn’t delivered
Carriers did not modernize the wrong systems. They modernized around the problem instead of through it.
When incomplete submissions land in a workbench, underwriters spend significant time reconstructing data instead of evaluating risk. When intake varies by region or broker, automation breaks. When AI models ingest inconsistent inputs, they amplify errors instead of creating insight.
No amount of downstream sophistication can compensate for upstream chaos.
Modernization must start with the operational layer that governs how unstructured work is ingested, enriched, validated, and routed before it reaches core systems.
What changes when the front door is fixed
When inbound work is standardized, enriched with insurance-specific context, validated for completeness, and intelligently routed before it hits downstream systems, everything else performs as intended.
Policy rules execute correctly. Workbenches receive complete, decision-ready data. Automation scales. AI becomes dependable because it operates on structured inputs. Cycle times compress. Capacity expands without adding headcount.
This is the shift behind Indico’s positioning as the Intake and Orchestration Platform for insurance, the operations layer that modernizes how work enters and moves through the enterprise.
Indico ingests messy submissions, emails, and claims packets; enriches them with insurance-specific context; validates completeness; and orchestrates work to the right systems and teams. It prepares work before it reaches core platforms so underwriting, claims, and servicing teams can focus on decision-making instead of document assembly.
See how Indico keeps underwriting work in motion. >>>
Insurers are rebuilding the front door because the old operating model no longer works
Both Wilde and Hannani pointed to mounting pressure on the operating model.
Labor no longer scales in a headcount-driven model. Submission variability continues to increase. Regulatory requirements are expanding. Boards expect measurable AI impact.
But as Wilde noted, AI is a multiplier of quality, not a substitute for structure. You cannot prompt-engineer your way out of inconsistent intake .
The takeaway from ITC London was not about chasing the next wave of technology. It was about fixing the foundation.
The next decade of insurance will be defined by who controls how work enters the enterprise. Carriers that rebuild the front door will unlock the performance they expected from modernization all along. Those that don’t will continue to invest in downstream tools without addressing the constraint that limits them.
FAQs
How does Indico Data’s agentic AI differ from traditional document processing?
Indico Data’s Agentic Decisioning Platform represents a departure from conventional intelligent document processing by deploying AI agents that ingest, enrich, and decide rather than simply extract data fields. The platform’s Agent Studio, Agent Builder, and Agent Workflow Canvas provide infrastructure for creating, tuning, and deploying insurance-specific agents that can be exported, imported, and promoted across environments.
Traditional document processing tools extract information and pass it to humans for interpretation, while Indico’s Next Best Action Agents provide appetite scoring, routing recommendations, and triage decisions directly. The platform employs a hybrid AI architecture combining discriminative, generative, and heuristic models, using what Indico describes as “the right technology for the right task”. This architecture supports outputs that include field-level provenance with token bounding boxes, page sources, model versions, and timestamped human edits for auditability. The practical impact shows in production results: a $50 billion carrier cut underwriting and renewal processing time from 7 days to 15 minutes using the platform. Guidewire’s Head of Corporate Development Jay Grayson noted that “Indico’s deep experience makes them a great fit for Guidewire’s global customer footprint,” validating the platform’s enterprise readiness.
What data enrichment capabilities does Indico Data provide for insurance workflows?
Indico Data’s Embedded Data Enrichment Agents provide native access to proprietary and third-party datasets including location intelligence, policy history, property characteristics, and risk scores directly within ingestion, triage, FNOL, and policy servicing workflows. These agents normalize extracted values such as occupancy codes, total insured values, and deductibles while linking related documents to create complete submission packages. The enrichment capability addresses a friction point in traditional workflows where underwriters must manually access multiple external systems to gather context for decisions.
The platform’s out-of-the-box coverage spans 120+ product lines with training on 20,000+ insurance-specific data points, enabling enrichment logic tuned to insurance terminology and structures. This embedded approach reduces the cost and complexity of maintaining separate data vendor relationships for each enrichment need. The enrichment layer feeds into the platform’s Next Best Action Agents, which use the combined extracted and enriched data to generate appetite assessments and routing recommendations without requiring manual data assembly.
How does Indico Data integrate with Guidewire and other core insurance systems?
Indico Data maintains a strategic partnership with Guidewire, including a strategic investment from Guidewire as part of a $19 million funding round. The Ready for Guidewire accelerators provide pre-configured field mappings and workflow hooks for PolicyCenter and ClaimCenter, enabling direct population of core system fields from extracted and enriched submission data. The integration architecture supports both synchronous straight-through processing for high-confidence submissions and asynchronous flows that route exceptions to human reviewers before committing data to the core system.
Beyond Guidewire, the platform exposes REST API and GraphQL interfaces for submissions and results, with SDKs available for Python, Java, and C# development teams. Webhook and SNS push notifications alert downstream systems when processing completes, enabling event-driven architectures. The platform’s deployment flexibility includes cloud, private-cloud, isolated tenant, and on-premises options with end-to-end encryption, role-based access, and detailed audit logs. Indico Data claims SOC 2 and HIPAA attestations along with a 99.99% SLA for platform availability. This integration approach allows carriers to maintain their core system investments while adding intelligent intake capabilities without replacing existing infrastructure.
What results have insurers achieved with Indico Data’s platform?
Indico Data’s production deployments demonstrate quantifiable improvements across multiple carrier implementations. A global specialty insurer with over $4 billion in gross written premium reduced submission processing time from two hours to 40 seconds while achieving 91% accuracy with zero human intervention. A $50 billion carrier cut underwriting and renewal processing time from 7 days to 15 minutes. Convex Insurance reported reducing time spent on Statements of Value by over 90%.
Beyond operational efficiency, carriers have realized business growth: Indico Data’s homepage highlights a $30 million increase in quarterly net premiums attributed to platform use. The company reports a 97% implementation success rate in case study documentation. Market adoption metrics show momentum: Indico Data achieved record ARR with 60% year-over-year growth in new carrier customers during H1 2025. Strategic investors including Guidewire and Aviva Ventures have validated the platform’s direction.
