{"id":15444,"date":"2026-05-25T16:46:32","date_gmt":"2026-05-25T11:16:32","guid":{"rendered":"http:\/\/10.10.164.133\/?p=15444"},"modified":"2026-05-25T17:35:21","modified_gmt":"2026-05-25T12:05:21","slug":"agentic-ai-at-payment-scale-how-payu-modernized-verify-without-losing-legacy-behavior","status":"publish","type":"post","link":"https:\/\/payu.in\/blog\/agentic-ai-at-payment-scale-how-payu-modernized-verify-without-losing-legacy-behavior\/","title":{"rendered":"Agentic AI at Payment Scale: How PayU Modernized Verify Without Losing Legacy Behavior"},"content":{"rendered":"\n

A single verify-payment call sounds simple \u2014 until it sits on the critical path of billions<\/strong> <\/strong>of payment checkouts. Behind that one call lived six different encryption schemes, hundreds of merchant and payment-gateway overrides, and a decade of implicit business logic buried inside a PHP monolith. Verify v1.0 powers one of the most payment-critical workflows at PayU<\/strong>: determining whether money actually moved<\/em>.<\/p>\n\n\n\n

The challenge was not replacing a failing system. The monolith worked. The challenge was modernizing <\/strong>a business-critical platform into a scalable, observable, and extensible Java-based UPI service \u2014 without breaking years of undocumented runtime behavior. It was a payment-grade behavior migration not just a language migration.<\/p>\n\n\n\n

We paired Agentic<\/strong> <\/strong>AI <\/strong>tools like Cursor and Claude with engineering discipline <\/strong>and human-in-the-loop validation. This post is how PayU<\/strong> <\/strong>used<\/strong> <\/strong>AI<\/strong> <\/strong>to<\/strong> <\/strong>accelerate <\/strong>the modernization of a real production payment system \u2014 not a sandbox experiment.<\/p>\n\n\n\n

The Problem Nobody Puts on a Sl<\/strong>i<\/strong>de Deck<\/strong><\/h2>\n\n\n\n

Most migration stories sound clean in presentations: \u201cWe migrated from a monolith to microservices.\u201d What rarely gets discussed is the invisible complexity hidden inside legacy systems.<\/p>\n\n\n\n

Our legacy repository had years of deeply coupled runtime-driven behavior. At first glance, the execution paths looked straightforward \u2014 but in reality, most of the logic was dynamically controlled through injected runtime objects, configurations, feature flags, and gateway \/ merchant-specific implementations.<\/p>\n\n\n\n

As developers new to this ecosystem, we weren\u02bct just migrating code. We first had to discover how the system actually behaved in production.<\/p>\n\n\n\n

The Breakthrough – Understanding the Unknown & Runtime Discovery Before Migration:<\/strong><\/p>\n\n\n\n

Before writing a single line of migration code, we spent time reverse-engineering the execution flow.<\/p>\n\n\n\n

We started by analyzing production logs from a few unique merchants to understand how requests traversed through the system. Those logs revealed only part of the picture.<\/p>\n\n\n\n

To uncover the h<\/strong>i<\/strong>dden runt<\/strong>i<\/strong>me paths<\/strong>, <\/strong>we<\/strong>:<\/strong><\/p>\n\n\n\n

            \u2022 Added deeper observability and targeted instrumentation into the legacy repository<\/p>\n\n\n\n

            \u2022 Deployed the enhanced version into pre-production<\/p>\n\n\n\n

            \u2022 Ran controlled test suites across multiple merchant scenarios<\/p>\n\n\n\n

            \u2022 Traced runtime-driven flows that previously had no visibility<\/p>\n\n\n\n

This process helped us identify execution branches that were entirely configuration-driven and impossible to infer statically from the codebase alone.<\/p>\n\n\n\n

Only after building this runtime map did we feel confident enough to begin the migration.<\/p>\n\n\n\n

Architecture First, AI Second:<\/strong><\/p>\n\n\n\n

Once the legacy flow was understood, the next challenge was designing a modern architecture that could scale with future onboarding and evolving PG integrations or to any evolutions.<\/p>\n\n\n\n

We collaborated closely with senior engineers and engineering managers to define a clean and extensible Spring Boot\u2013based UPI service architecture.<\/p>\n\n\n\n

The<\/strong> <\/strong>key<\/strong> <\/strong>des<\/strong>i<\/strong>gn<\/strong> <\/strong>pr<\/strong>i<\/strong>nc<\/strong>i<\/strong>ples<\/strong> <\/strong>were<\/strong>:<\/strong><\/p>\n\n\n\n

            \u2022 Clear orchestration boundaries<\/p>\n\n\n\n

            \u2022 Gateway-specific abstractions<\/p>\n\n\n\n

            \u2022 Reusable utilities and encryption layers<\/p>\n\n\n\n

            \u2022 Strict ownership separation between orchestration and core transaction persistence paths<\/p>\n\n\n\n

            \u2022 Extensibility for future PG integrations without service rewrites<\/p>\n\n\n\n

What looked like \u201cone payment flow\u201d from the outside was actually dozens of unique behaviors hidden behind gateway-specific implementations.<\/p>\n\n\n\n

            \u2022 Every payment gateway spoke its own dialect<\/p>\n\n\n\n

            \u2022 Hundreds of merchant parameters and feature flags<\/p>\n\n\n\n

            \u2022 Different request\/response contracts<\/p>\n\n\n\n

            \u2022 Different checksum and signature validation rules<\/p>\n\n\n\n

            \u2022 AES, RSA, and custom encryption strategies<\/p>\n\n\n\n

            \u2022 Nested success\/failure semantics<\/p>\n\n\n\n

            \u2022 Unique timeout and retry behaviors<\/p>\n\n\n\n

The complexity wasn\u02bct in the API calls. It was in preserving years of edge-case behavior without breaking reconciliation and settlements.<\/p>\n\n\n\n

The Agent<\/strong>i<\/strong>c M<\/strong>i<\/strong>grat<\/strong>i<\/strong>on Loop – Where AI Became a Force Mult<\/strong>i<\/strong>pl<\/strong>i<\/strong>er<\/strong>:<\/strong><\/h2>\n\n\n\n

After finalizing the architecture skeleton, we onboarded both the legacy monolith and the new microservice into a shared workspace and fed the latest runtime logs into our development workflow.<\/p>\n\n\n\n

This is where AI dramatically accelerated execution.<\/p>\n\n\n\n

Instead of manually port<\/strong>i<\/strong>ng thousands of l<\/strong>i<\/strong>nes of log<\/strong>i<\/strong>c gateway by gateway<\/strong>, <\/strong>we leveraged AI-ass<\/strong>i<\/strong>sted development to<\/strong>:<\/strong><\/p>\n\n\n\n

            \u2022 Understand legacy method-level intent<\/p>\n\n\n\n

            \u2022 Convert language and framework patterns<\/p>\n\n\n\n

            \u2022 Map implementations into the correct architectural layers<\/p>\n\n\n\n

            \u2022 Identify abstractions and orchestration boundaries<\/p>\n\n\n\n

            \u2022 Generate migration scaffolding aligned with the new design<\/p>\n\n\n\n

Because the architecture skeleton was already well-defined, the AI systems were intelligent enough to infer where code belonged:<\/p>\n\n\n\n

            \u2022 orchestration layers<\/p>\n\n\n\n

            \u2022 utility modules<\/p>\n\n\n\n

            \u2022 gateway implementations<\/p>\n\n\n\n

            \u2022 abstractions and interfaces<\/p>\n\n\n\n

AI was highly effective for structured scaffolding and repetitive transformation, while engineers retained ownership of correctness, edge cases, and production readiness.<\/p>\n\n\n\n

What would traditionally take weeks of repetitive engineering effort was reduced to days of focused validation and refinement.<\/p>\n\n\n\n

We completed the core migration activity in under a week.<\/p>\n\n\n\n

\n
\"\"<\/figure>\n<\/figure>\n\n\n\n

Val<\/strong>i<\/strong>dat<\/strong>i<\/strong>on for Payment-Grade Trust and Product<\/strong>i<\/strong>on Read<\/strong>i<\/strong>ness<\/strong>:<\/strong><\/h2>\n\n\n\n

The migration itself was only half the challenge.<\/p>\n\n\n\n

In payment systems, even a minor behavioral deviation can impact reconciliation, settlements, and downstream financial operations. So our biggest focus became validation.<\/p>\n\n\n\n

We designed a bidirectional validation strategy:<\/p>\n\n\n\n

            \u2022 Process the first half of a transaction flow through the legacy system<\/p>\n\n\n\n

            \u2022 Feed the intermediate processed state into the modernized service<\/p>\n\n\n\n

            \u2022 Validate whether the final outcome matched expected production behavior <\/p>\n\n\n\n

Then we reversed the flow:<\/strong><\/p>\n\n\n\n

            \u2022 Process the initial stages in the new system<\/p>\n\n\n\n

            \u2022 Hand over execution into the legacy platform<\/p>\n\n\n\n

            \u2022 Compare outputs and downstream effects<\/p>\n\n\n\n

This cross-system execution model gave us confidence that the migrated service preserved legacy behavior while remaining compatible with existing operational flows.<\/p>\n\n\n\n

Six PG integrations reached production readiness in about four weeks, with the core migration scaffold completed much earlier and the majority of effort spent on validation.<\/p>\n\n\n\n

Outcomes<\/strong> <\/strong>and<\/strong> <\/strong>Lessons<\/strong>:<\/strong><\/h2>\n\n\n\n

The hardest part of modernization is rarely rewriting code.<\/p>\n\n\n\n

It\u02bcs discovering the undocumented assumptions accumulated over years of production evolution.<\/p>\n\n\n\n

What made this migration successful was not just AI-assisted code generation. It was the combination of:<\/p>\n\n\n\n

            \u2022 Runtime observability<\/p>\n\n\n\n

            \u2022 Architectural clarity<\/p>\n\n\n\n

            \u2022 Controlled validation strategies<\/p>\n\n\n\n

            \u2022 Deep understanding of payment edge cases<\/p>\n\n\n\n

AI didn\u02bct magically solve the migration. But it significantly compressed the time required to understand, transform, and modernize a highly complex legacy ecosystem.<\/p>\n\n\n\n

Our discovery: Agentic AI shines when the work is <\/strong>structured<\/em><\/strong>, not when it\u02bcs <\/strong>j<\/em><\/strong>udgment<\/em><\/strong><\/p>\n\n\n\n

We stopped treating AI as a magic translator and started treating it as a loop<\/strong>. That consistency didn\u02bct come from the model. We<\/strong> <\/strong>designed<\/strong> <\/strong>the<\/strong> <\/strong>template;<\/strong> <\/strong>AI<\/strong> <\/strong>filled<\/strong> <\/strong>the<\/strong> <\/strong>slots.<\/strong><\/p>\n\n\n\n

Our team still had to:<\/p>\n\n\n\n

            \u2022 Human-in-the-loop <\/strong>reviews for every AI-generated code path, ensuring protocol-level accuracy before merge.<\/p>\n\n\n\n

            \u2022 Run real transactions <\/strong>and diff logs: Legacy repo vs new microservice, same transaction identifiers.<\/p>\n\n\n\n

            \u2022 Paste production-like failures back into Cursor <\/strong>\u2014 \u201c gateway returned a 400 with an empty parameter set; what did the legacy system do differently?”<\/em>\u201d<\/p>\n\n\n\n

            \u2022 Catch implicit legacy behavior <\/strong>AI invented or skipped (status pre-checks, transaction cancel counts, auth-leg field merges \u2014 the kind of logic that never appears in a README)<\/p>\n\n\n\n

> Our<\/strong> <\/strong>rule: <\/strong>No generated verify path merged without a human who could explain it to an on-call at 2 a.m.<\/p>\n\n\n\n

Prompt quality mattered more than model choice. Vague prompts gave us code snippet that failed parity. Specific prompts \u2014 \u201cmirror<\/em> <\/em>the<\/em> <\/em>legacy<\/em> <\/em>verification<\/em> <\/em>response<\/em> <\/em>validator<\/em> <\/em>for<\/em> <\/em>status<\/em> <\/em>U.\u201d<\/em><\/p>\n\n\n\n

\u2014 gave us mergeable code.<\/p>\n\n\n\n

Before vs after: What Improved \u2014 Without the Hype :<\/strong> <\/h2>\n\n\n\n
Aspect<\/strong>  <\/td>Without Agentic AI (estimate)<\/strong>  <\/td>With Agentic AI + strict review<\/strong> <\/strong> <\/td><\/tr>
Time per PG integration  <\/td>High \u2014 full manual rewrite  <\/td>Significantly lower<\/strong> <\/strong> <\/td><\/tr>
Understanding legacy PHP <\/td>Months of archaeology <\/td>Days with AI-assisted walkthroughs <\/td><\/tr>
Boilerplate (DTOs, clients, tests) <\/td>Manual <\/td>Mostly generated, human-edited (Only wherever needed)<\/strong> <\/td><\/tr>
Risk of missing implicit branches <\/td>High <\/td>Lower \u2014 but only with log-driven iteration <\/td><\/tr>
Cross-PG consistency <\/td>Variable <\/td>Standardized gateway template<\/strong> <\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Standardized<\/strong> <\/strong>gateway <\/strong>template<\/strong><\/p>\n\n\n\n

We\u02bcre not claiming \u201c90% less code\u201d or \u201czero bugs.\u201d We\u02bcre claiming something more honest: we <\/strong>shipped six PG integrations with a small team, a repeatable pattern, and Verify v1.0 now in Production <\/strong>\u2014 without recreating monolithic complexity separately for every gateway.<\/p>\n\n\n\n

The work wasn\u02bct done by engineers who had maintained legacy repo for years. Our squad was a Lead<\/strong>, an ASE<\/strong>, and an intern<\/strong> <\/strong>\u2014 none of us had prior exposure to the monolith repo when we started. Agentic AI compressed the legacy ramp; the Lead anchored design, parity, and what we were allowed to ship to production.<\/p>\n\n\n\n

Clos<\/strong>i<\/strong>ng thought<\/strong><\/h2>\n\n\n\n

The hard part isn\u02bct syntax, it\u02bcs trust<\/strong>. Trust that a captured UPI transaction in Bengaluru behaves exactly like it did when legacy system owned the thread.<\/p>\n\n\n\n

AI proposes; engineers dispose. We used Agentic AI to reduce friction \u2014 not to outsource judgment. Our team discovered that the winning combo is a boring<\/strong> <\/strong>modern<\/strong> <\/strong>architecture<\/strong>, exciting<\/strong> <\/strong>iteration<\/strong> <\/strong>speed<\/strong>, and engineers who still read hex dumps when the model says \u201clooks good to me.\u201d<\/p>\n\n\n\n

PayU <\/strong>did not simply use Cursor\/Claude to port code. PayU <\/strong>designed a payment-grade, agentic migration operating model: runtime discovery, architecture-first modernization, AI-assisted implementation, human review, and parity validation before production.<\/p>\n\n\n\n

If you\u02bcre mid-flight on a legacy migration, start with one provider, one stage, one log diff. Let AI draft the RequestProcessor.x<\/strong>. You keep the keys to production.<\/p>\n","protected":false},"excerpt":{"rendered":"

A single verify-payment call sounds simple \u2014 until it sits on the critical path of billions of payment checkouts. Behind that one call lived six different encryption schemes, hundreds of…<\/p>\n","protected":false},"author":84,"featured_media":0,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1232],"tags":[],"yoast_head":"\nAgentic AI at Payment Scale: How PayU Modernized Verify Without Losing Legacy Behavior - PayU Blog<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/payu.in\/blog\/agentic-ai-at-payment-scale-how-payu-modernized-verify-without-losing-legacy-behavior\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Agentic AI at Payment Scale: How PayU Modernized Verify Without Losing Legacy Behavior - PayU Blog\" \/>\n<meta property=\"og:description\" content=\"A single verify-payment call sounds simple \u2014 until it sits on the critical path of billions of payment checkouts. Behind that one call lived six different encryption schemes, hundreds of...\" \/>\n<meta property=\"og:url\" content=\"https:\/\/payu.in\/blog\/agentic-ai-at-payment-scale-how-payu-modernized-verify-without-losing-legacy-behavior\/\" \/>\n<meta property=\"og:site_name\" content=\"PayU Blog\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/PayUind\/\" \/>\n<meta property=\"article:published_time\" content=\"2026-05-25T11:16:32+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2026-05-25T12:05:21+00:00\" \/>\n<meta property=\"og:image\" content=\"http:\/\/10.10.164.133\/wp-content\/uploads\/2026\/05\/Traditional-vs-Agentic-Ai-Assisted-Approach-1-1.png\" \/>\n<meta name=\"author\" content=\"Nikita Ekka\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:creator\" content=\"@payuind\" \/>\n<meta name=\"twitter:site\" content=\"@payuind\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Nikita Ekka\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"10 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"http:\/\/10.10.164.133\/agentic-ai-at-payment-scale-how-payu-modernized-verify-without-losing-legacy-behavior\/\",\"url\":\"http:\/\/10.10.164.133\/agentic-ai-at-payment-scale-how-payu-modernized-verify-without-losing-legacy-behavior\/\",\"name\":\"Agentic AI at Payment Scale: How PayU Modernized Verify Without Losing Legacy Behavior - PayU Blog\",\"isPartOf\":{\"@id\":\"https:\/\/payu.in\/blog\/#website\"},\"datePublished\":\"2026-05-25T11:16:32+00:00\",\"dateModified\":\"2026-05-25T12:05:21+00:00\",\"author\":{\"@id\":\"https:\/\/payu.in\/blog\/#\/schema\/person\/a20ce79fd4ff1edcb9e22b0b8b64d1d8\"},\"breadcrumb\":{\"@id\":\"http:\/\/10.10.164.133\/agentic-ai-at-payment-scale-how-payu-modernized-verify-without-losing-legacy-behavior\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"http:\/\/10.10.164.133\/agentic-ai-at-payment-scale-how-payu-modernized-verify-without-losing-legacy-behavior\/\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"http:\/\/10.10.164.133\/agentic-ai-at-payment-scale-how-payu-modernized-verify-without-losing-legacy-behavior\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/payu.in\/blog\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Engineering Blogs\",\"item\":\"http:\/\/10.10.164.133\/category\/engineering-blogs\/\"},{\"@type\":\"ListItem\",\"position\":3,\"name\":\"Agentic AI at Payment Scale: How PayU Modernized Verify Without Losing Legacy Behavior\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/payu.in\/blog\/#website\",\"url\":\"https:\/\/payu.in\/blog\/\",\"name\":\"PayU Blog\",\"description\":\"\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/payu.in\/blog\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"en-US\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/payu.in\/blog\/#\/schema\/person\/a20ce79fd4ff1edcb9e22b0b8b64d1d8\",\"name\":\"Nikita Ekka\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/payu.in\/blog\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/secure.gravatar.com\/avatar\/4448daa49886616547819223e7d11196?s=96&d=mm&r=g\",\"contentUrl\":\"https:\/\/secure.gravatar.com\/avatar\/4448daa49886616547819223e7d11196?s=96&d=mm&r=g\",\"caption\":\"Nikita Ekka\"},\"sameAs\":[\"http:\/\/payu.in\/blog\"],\"url\":\"https:\/\/payu.in\/blog\/author\/nikita-ekka\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Agentic AI at Payment Scale: How PayU Modernized Verify Without Losing Legacy Behavior - PayU Blog","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/payu.in\/blog\/agentic-ai-at-payment-scale-how-payu-modernized-verify-without-losing-legacy-behavior\/","og_locale":"en_US","og_type":"article","og_title":"Agentic AI at Payment Scale: How PayU Modernized Verify Without Losing Legacy Behavior - PayU Blog","og_description":"A single verify-payment call sounds simple \u2014 until it sits on the critical path of billions of payment checkouts. Behind that one call lived six different encryption schemes, hundreds of...","og_url":"https:\/\/payu.in\/blog\/agentic-ai-at-payment-scale-how-payu-modernized-verify-without-losing-legacy-behavior\/","og_site_name":"PayU Blog","article_publisher":"https:\/\/www.facebook.com\/PayUind\/","article_published_time":"2026-05-25T11:16:32+00:00","article_modified_time":"2026-05-25T12:05:21+00:00","og_image":[{"url":"http:\/\/10.10.164.133\/wp-content\/uploads\/2026\/05\/Traditional-vs-Agentic-Ai-Assisted-Approach-1-1.png"}],"author":"Nikita Ekka","twitter_card":"summary_large_image","twitter_creator":"@payuind","twitter_site":"@payuind","twitter_misc":{"Written by":"Nikita Ekka","Est. reading time":"10 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"http:\/\/10.10.164.133\/agentic-ai-at-payment-scale-how-payu-modernized-verify-without-losing-legacy-behavior\/","url":"http:\/\/10.10.164.133\/agentic-ai-at-payment-scale-how-payu-modernized-verify-without-losing-legacy-behavior\/","name":"Agentic AI at Payment Scale: How PayU Modernized Verify Without Losing Legacy Behavior - PayU Blog","isPartOf":{"@id":"https:\/\/payu.in\/blog\/#website"},"datePublished":"2026-05-25T11:16:32+00:00","dateModified":"2026-05-25T12:05:21+00:00","author":{"@id":"https:\/\/payu.in\/blog\/#\/schema\/person\/a20ce79fd4ff1edcb9e22b0b8b64d1d8"},"breadcrumb":{"@id":"http:\/\/10.10.164.133\/agentic-ai-at-payment-scale-how-payu-modernized-verify-without-losing-legacy-behavior\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["http:\/\/10.10.164.133\/agentic-ai-at-payment-scale-how-payu-modernized-verify-without-losing-legacy-behavior\/"]}]},{"@type":"BreadcrumbList","@id":"http:\/\/10.10.164.133\/agentic-ai-at-payment-scale-how-payu-modernized-verify-without-losing-legacy-behavior\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/payu.in\/blog\/"},{"@type":"ListItem","position":2,"name":"Engineering Blogs","item":"http:\/\/10.10.164.133\/category\/engineering-blogs\/"},{"@type":"ListItem","position":3,"name":"Agentic AI at Payment Scale: How PayU Modernized Verify Without Losing Legacy Behavior"}]},{"@type":"WebSite","@id":"https:\/\/payu.in\/blog\/#website","url":"https:\/\/payu.in\/blog\/","name":"PayU Blog","description":"","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/payu.in\/blog\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"en-US"},{"@type":"Person","@id":"https:\/\/payu.in\/blog\/#\/schema\/person\/a20ce79fd4ff1edcb9e22b0b8b64d1d8","name":"Nikita Ekka","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/payu.in\/blog\/#\/schema\/person\/image\/","url":"https:\/\/secure.gravatar.com\/avatar\/4448daa49886616547819223e7d11196?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/4448daa49886616547819223e7d11196?s=96&d=mm&r=g","caption":"Nikita Ekka"},"sameAs":["http:\/\/payu.in\/blog"],"url":"https:\/\/payu.in\/blog\/author\/nikita-ekka\/"}]}},"_links":{"self":[{"href":"https:\/\/payu.in\/blog\/wp-json\/wp\/v2\/posts\/15444"}],"collection":[{"href":"https:\/\/payu.in\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/payu.in\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/payu.in\/blog\/wp-json\/wp\/v2\/users\/84"}],"replies":[{"embeddable":true,"href":"https:\/\/payu.in\/blog\/wp-json\/wp\/v2\/comments?post=15444"}],"version-history":[{"count":7,"href":"https:\/\/payu.in\/blog\/wp-json\/wp\/v2\/posts\/15444\/revisions"}],"predecessor-version":[{"id":15487,"href":"https:\/\/payu.in\/blog\/wp-json\/wp\/v2\/posts\/15444\/revisions\/15487"}],"wp:attachment":[{"href":"https:\/\/payu.in\/blog\/wp-json\/wp\/v2\/media?parent=15444"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/payu.in\/blog\/wp-json\/wp\/v2\/categories?post=15444"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/payu.in\/blog\/wp-json\/wp\/v2\/tags?post=15444"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}