Insurance · APAC · Dedicated Team
The AI Ideas Were There. The Engineers to Build Them Were Not.
Innovation managers at a leading insurance group across Asia-Pacific were drowning in manual review - hundreds of ideas, no consistent framework, no visibility. The right two engineers changed that.
Specialists deployed
2
Duration
6 mo
Engagement model
Dedicated Team
"Technical capability was there from day one. What we didn't expect was how well they navigated the people side"
— Innovation Lead, Insurance Group
Every cohort, the submissions piled up. Every cohort, someone had to read them all by hand.
A leading insurance group was running structured AI innovation programmes across multiple business units in Asia-Pacific. Employees submitted ideas each cohort - claims automation, underwriting support, agent tools, customer experience. But the infrastructure behind the programme had not kept pace. Submissions landed in emails and spreadsheets. Innovation managers reviewed every entry manually, with no consistent scoring framework, no way to detect duplicate proposals across markets, no structured path from raw idea to funded project. Leadership had no pipeline visibility at all.
The organisation needed engineers who understood both modern full-stack architecture and production AI integration - a combination that is nearly impossible to hire for quickly.
Two senior AI engineers. Embedded in the team. Platform delivered in two months.
Talex identified and deployed two senior full-stack developers with production AI experience directly into the client's AI Lab - embedded alongside innovation managers and programme managers as the core engineering function. Part of a six-month engagement with the client, this platform was scoped and delivered within two months.
The engineers built the full system end-to-end: automated idea intake via Microsoft Forms webhooks, AI-powered scoring using Google Vertex AI Gemini 2.5 Pro across five dimensions (strategic fit, benefits, feasibility, innovation, scalability) with weighted scoring and reasoning, intelligent clustering to surface duplicate proposals across business units, a Discovery Canvas scoping workflow with AI-generated insights, DVF validation before any resource commitment, a leadership dashboard with cohort tracking and prioritisation scatter charts, and integration with the organisation's existing AI solutions register to prevent redundant work before it starts.
The project operated within a complex stakeholder environment - innovation managers, programme leads, and external consulting stakeholders with different expectations of the engineering team. Talex's account management stayed close throughout the engagement, ensuring the engineers had clarity on priorities and that stakeholder complexity did not become delivery risk. The team stayed focused. The timeline held.
Team Architecture
Request similar team →Full-Stack Developer
Senior
Stack & Compliance
The innovation programme finally had the infrastructure to match its ambition.
Evaluation dropped from days of manual review to near-instant AI-generated scoring upon submission. Every idea now goes through the same five-dimension framework — outcomes no longer depend on who reviewed it that week. Duplicate proposals are flagged before redundant work begins. Leadership gained live visibility across all business units, strategic pillars, and cohorts for the first time. Innovation managers shifted from triage to strategy. The platform runs multiple cohorts in parallel without scaling the team behind it.
The client noted that the quality and consistency of idea evaluation had improved significantly since the platform went live, with innovation managers reporting more bandwidth for strategic decision-making rather than manual processing.
200% increase (Innovation Throughput)
Doubled the rate of idea processing and evaluation.
Improved (Strategic Alignment)
Ensured alignment of AI projects with business goals.
Enhanced (Scalability)
Platform designed to handle increased submission volume efficiently.
Significant (Duplication Reduction)
Identified and reduced duplicate proposals across units.
Timeline
Discovery & Planning · 2 months
Analyzing existing processes and defining the project scope.
Development · 5 months
Building and integrating the innovation management platform.
Deployment & Training · 2 months
Implementing the platform and training the staff.
Business Outcomes
- →Consistent (Standardised Scoring) : Every submission is evaluated across the same five dimensions by the same AI model - removing the variability that came from reviewer subjectivity.
- →Days to instant (Evaluation Speed) : Manual review that previously took innovation managers days to clear now happens the moment a submission lands.
Engineering Excellence
- →Near-instant scoring : AI Evaluation Speed : Ideas submitted via Microsoft Forms are scored automatically upon intake — no manual review queue, no waiting. Google Vertex AI Gemini 2.5 Pro evaluates each submission across five weighted dimensions and returns reasoning instantly.
- →Zero duplication : Intelligent Clustering : Related ideas from different business units across Asia are automatically grouped and matched against the organisation's existing AI solutions register — preventing redundant work before it starts.
- →Idea to scoped project : Structured Conversion Pipeline : Approved ideas move through Discovery Canvas scoping and DVF validation (Desirability, Viability, Feasibility) before any resource is committed - ensuring only well-defined projects reach implementation.
Why Talex
Expertise in AI Integration 7 days
Talex's deep understanding of AI and innovation processes made them the ideal partner.
Strategic Approach
Focused on transforming the innovation process, not just implementing technology.
Services
Related Projects
The hardest part of building at the edge of what is technically possible is finding the people who have already been there.
1.5 million users. Zero tolerance for downtime. One team to build it right.
