Policy, billing, claims — and the data, cloud and AI under all three.
Y&L Consulting, a YASH Technologies company, builds software and data platforms for insurance carriers and insurtechs: product engineering, legacy modernization, data platforms, cloud, and AI that reaches production. We're at ITC Vegas all three days, running thirty-minute working sessions.
Insurance engagements delivered
Lifecycle domains: policy, billing, claims
Service lines, one accountable team
API integrations on one CRM build
The whole lifecycle — and the layer that runs beneath it.
Carriers are pushing data and AI into every stage of the policy lifecycle at once. The stages are where the business value sits. The layer underneath is where the programmes succeed or stall.
Policy
- Submission intake and triage
- Rating, quoting and binding
- Issuance and document generation
- Endorsements and renewals
Billing
- Invoicing and payment processing
- Collections and dunning
- Commission and producer settlement
- Financial reconciliation
Claims
- FNOL intake across channels
- Triage, segmentation and assignment
- Adjudication and reserving
- Settlement, subrogation, recovery
Servicing
- Policyholder and producer portals
- Contact-centre operations
- Complaints and disputes
- Regulatory and statutory reporting
Data, cloud and AI
One governed data estate feeding every stage, hosted on cloud you can evidence to an examiner, with AI applied where it changes a decision rather than where it demos well. This is the layer that decides whether a promising pilot in claims ever reaches the other three stages.
Two very different problems. The same engineering bench.
A carrier modernizing a twenty-year-old estate and an insurtech shipping v2 need opposite things from a partner. We staff them differently and say so upfront.
You have the data. Getting value out of it is the programme.
Policy, billing, claims and finance each hold part of the answer and none of them agree. We modernize the estate underneath and put data and AI into the lifecycle stage by stage, without a big-bang cutover the board won't approve.
- Legacy modernization and core system integration
- Governed data platforms across policy, billing and claims
- Underwriting and claims analytics, with MLOps after go-live
- Application management and production support
- Regulatory, audit and statutory reporting support
You have the product. Scaling it is the problem.
The prototype won the pilot; now a carrier's security review, data volumes and integration requirements are in the way. We build product alongside your team and bring the insurance domain your engineers shouldn't have to learn from scratch.
- Software product development and platform engineering
- Carrier integrations: policy, billing, claims and rating APIs
- Multi-tenant architecture and cloud cost engineering
- AI features built to survive a carrier security review
- Dedicated squads that scale up and down with the roadmap
Six service lines, shaped around insurance.
Insurance technology work fails on domain, not on engineering. Everything below is built around policy, billing, claims and the regulator — and is running in a carrier or an insurtech today.
Most programmes cut across three or four. One team carries them, so the seams are ours to manage, not yours.
Product engineering with the insurance domain already in the room.
For insurtechs building the platform and for carriers building digital products their policyholders and producers actually use. Dedicated squads, your backlog, our engineers — who already know what a rating call and an FNOL payload look like.
- Discovery, product architecture and technical design
- Full-stack build: web, mobile, APIs and services
- Carrier integrations — policy, billing, claims, rating
- Multi-tenant architecture and cloud cost engineering
- QA automation, CI/CD and release engineering
- Dedicated squads that flex with the roadmap
Insurtechs past the pilot and heading into carrier procurement, where security review, data volume and integration depth suddenly become the roadmap.
Carriers tooPolicyholder and producer portals, quote-and-bind journeys, and the service applications that sit in front of a core system nobody wants to expose.
Moving the workload and the knowledge off legacy together.
Legacy ETL, undocumented logic and batch windows that no longer fit the business. The migration is the easy half; carrying the domain knowledge across is what decides whether you are back here in three years.
- Legacy assessment and fit-gap analysis
- SAS and legacy ETL modernization
- API and integration layer development
- Cloud migration and re-platforming
- Application management and production support
Claims data-universe and SAS modernization, with legacy ETL migrated to Databricks alongside Python, API and NLP pipelines.
Specialty insurerClaims-aligned application management covering data QA, architecture, modernization and analytics.
One governed estate feeding policy, billing and claims.
Every AI ambition in insurance resolves back to this. We design and build the platform underneath, sequenced so the business gets something usable long before the full estate is finished.
- Target-state architecture and platform selection
- Lakehouse and warehouse build-out
- Legacy ETL migration and pipeline engineering
- Data quality instrumentation and observability
- Underwriting and claims analytics, pricing and loss trends
- BI enablement for internal and external consumption
A scalable data foundation with analytics enablement, staged so the first business-facing reports shipped ahead of full platform completion.
Global life & annuity groupContact-centre data integrated into an AWS data lake, opening it up for analytics and downstream AI.
AI aimed at the decisions that actually hold up the business.
Submission triage, document extraction, claims segmentation, servicing copilots. We build on Claude, Azure OpenAI and open models — deployed inside your tenancy, with the evaluation and guardrails a carrier's risk function will ask for before anything goes live.
- Use-case shaping and value-vs-complexity prioritisation
- Document intelligence: OCR, extraction, classification
- Retrieval over policy wordings, guidelines and claim files
- Agentic workflows with deterministic guardrails
- Evaluation harnesses, red-teaming and human-in-the-loop design
- MLOps: deployment, drift monitoring, retraining, audit trail
Underwriting automation using email classification, OCR and entity extraction, feeding automated dashboard updates — with MLOps in place so model performance stayed observable after go-live.
The cloud footprint underneath it, and someone to run it.
Migration is a project; operations is the rest of the decade. We do both, and we would rather hand you a platform your own team can run than a dependency.
- Cloud migration and landing-zone design
- Infrastructure as code and environment automation
- Platform engineering and developer tooling
- FinOps and cloud cost optimisation
- Application management services and production support
- 24x7 monitoring and incident response
Onshore, offshore or blended, with the domain leads staying constant across the engagement — the part that usually breaks when a programme moves from build into run.
Making the trustworthy path the fastest one.
Governance people route around is not governance, and an AI programme without lineage is a compliance conversation waiting to happen. We build the trust layer as infrastructure, not as a committee.
- Metadata management and business glossary
- End-to-end lineage, source to consumption
- Data quality rules, monitoring and stewardship
- Governed self-service enablement
- Cloud-native managed detection and response
- NIST and Zero Trust alignment, audit evidence
A data-wellness capability spanning metadata, business glossary, lineage, quality monitoring, governed self-service and regulatory compliance support.
Regulated healthcare clientAI-augmented managed detection and response on Microsoft Sentinel, aligned to NIST and Zero Trust principles.
Come in at any stage. Most people don't start at the beginning.
You may already have the roadmap and need delivery. You may have the platform and need someone to run it. We pick up at whichever stage is actually open.
Discovery
Current-state architecture, process flows, fit-gap analysis and stakeholder workshops. Four to six weeks, with a written point of view at the end.
Roadmap
Target architecture, sequencing and scope, with the first increment defined tightly enough to fund the rest of the programme.
Build
Product, platform and analytics delivery against the roadmap — onshore, offshore or blended, with the domain leads staying constant.
Run
Application management, production support and enhancement, so the capability keeps working after the programme team has moved on.
Insurance AI is a data problem in costume.
There will be several hundred AI demos on the floor at ITC this year. Almost all of them work. Very few reach production at scale, and the reason is almost never the model — it is what the model is standing on. Three failure modes, over and over:
The data isn't wrong. It's unaccountable.
The number is probably right. Nobody can prove it. Without lineage, a glossary and quality monitoring, every model output becomes a negotiation with risk and compliance — and that negotiation is what kills the rollout, not the accuracy.
The pilot ran on an extract.
Somebody hand-pulled a clean dataset to make the demo work. Production means policy, billing, claims and finance on a live footing, with batch windows and a migration path off the legacy estate. That is a platform problem wearing an AI costume.
Nobody owns it after go-live.
Models drift, volumes move, the vendor's consultant leaves. Without MLOps, observability and a support model your own team can operate, month four looks nothing like week one — and by then the sponsor has moved on.
- AWS
- Azure
- Databricks
- Microsoft Fabric
- Snowflake
- Kubernetes
- Terraform
- Power BI
- Claude
- Azure OpenAI
- Databricks ML
- Open models
- MLOps & evaluation
- Salesforce FSC
- MuleSoft
- Microsoft Sentinel
- Talkdesk
- Calabrio
Thirty minutes with the people who build it.
We're at Mandalay Bay from 29 September to 1 October. Rather than a booth demo, we're running short working sessions with carrier and insurtech technology leaders on whichever service line is live for you right now. We'll confirm the exact meeting point when we lock your time.
Slots are limited and ITC weeks fill up fast — most attendees book their calendar out before they land.
- 01A straight answer on whether this is a six-week discovery or an eighteen-month programme.
- 02The architecture pattern we would propose, and the two we would rule out.
- 03Comparable delivery references, named on request under NDA.
- 04No follow-up sequence unless you ask for one.