// ── HERO ────────────────────────────────────────────────────────── // Same dark-green motion-blur treatment as Home/Contact heroes. // Headline = approved Short description from SERVICE-DESCRIPTIONS.md. // Sub-paragraph = first paragraph of the approved Expanded description. const SolutionHero = ({ onCTA }) =>
; // ── OVERVIEW ─────────────────────────────────────────────────────── // Mirrors the Homepage "30+ years of experience" two-column rhythm: // eyebrow + hairline rule, then a 1.15fr / 1fr split with H2 on the // left and a body paragraph on the right. // Copy = approved 2nd paragraph of the Expanded description. const Overview = () =>
Overview

Designed around your model objective.

Each program is designed around the model objective, evaluation criteria, benchmark framework, reviewer qualifications, and scoring rubrics. Most evaluation work is delivered through Argos Myriad — the Argos Data Platform — with SmartSuite, our customizable tooling environment, providing custom evaluation workflows, scalable reviewer deployment, and embedded quality controls. Argos Data can also work inside client evaluation platforms or through structured file exchange when that fits the program better.

; // ── USE CASES ────────────────────────────────────────────────────── // 6 bullets, verbatim from the approved "Use cases" list in // SERVICE-DESCRIPTIONS.md under AI Data Collection. Each renders as // a numbered row inside a single bordered column block — light, calm // rhythm consistent with the rest of the site's hairlined dividers. const UseCases = () => { const items = [ "Comprehensive LLM Evaluation for quality, relevance, accuracy, helpfulness, safety, and task performance", "Multilingual Evaluation for the cross-language, comparative view of model behavior across languages, locales, and regional contexts", "In-Language Model Validation for the deep, single-market view of whether a model is ready for deployment in a specific locale", "A/B Testing & Model Comparison for side-by-side evaluation of model versions, prompts, or system configurations", "Domain-Specific Benchmarking for evaluating models against industry, task, or subject-matter standards", "Human-in-the-Loop Validation for expert review, adjudication, and quality assurance across model evaluation workflows", "Agentic AI Evaluation & Workflow Validation for assessing how AI agents plan, reason, use tools, and complete multi-step tasks", ]; return (
Use cases

Where Model Evaluation and Benchmarking is applied.

{items.map((it, i) =>
{String(i + 1).padStart(2, "0")}
{it}
)}
); }; // ── SUB-SERVICES GRID ────────────────────────────────────────────── // Reuses the Homepage Solutions card pattern verbatim: numbered card, // rounded image tile, title, approved short description, pill CTA, // on the same cream/yellow radial wash. 6 cards = the 6 sub-services // documented under AI Data Collection in SERVICE-DESCRIPTIONS.md. // // Image assignments: where a Homepage card uses an approved photo // for the same topic, we reuse it. Where the doc names no thumbnail, // we pick from the approved photography pool (no new imagery // invented). If no approved photo clearly matches, a literal // "Lorem ipsum…" placeholder tile is shown so missing imagery is // visible to the manager — per project rules. // SvcIcon is loaded globally from components/svc-icons.jsx // (28 icons from the supplied Argos iconset, addressable by name "01"–"28"). const SubServices = () => { const services = [ { id: "01", slug: "comprehensive-llm-evaluation", parent: "model-evaluation", title: "Comprehensive LLM Evaluation", desc: "Enterprise-grade evaluation for assessing LLM quality, reliability, safety, and production readiness", icon: "stack", bg: "#0E1F3E", fg: "#EEE36B" }, // navy + yellow { id: "02", slug: "multilingual-evaluation", parent: "model-evaluation", title: "Multilingual Evaluation", desc: "The cross-language, comparative view of how a model performs across many languages, locales, and regions", icon: "arcs", bg: "#30AA74", fg: "#FFFFFF" }, // green + white { id: "03", slug: "in-language-model-validation", parent: "model-evaluation", title: "In-Language Model Validation", desc: "The deep, single-market view of whether a model is ready for deployment in a specific locale", icon: "half-arcs", bg: "#EEE36B", fg: "#0E1F3E" }, // yellow + navy { id: "04", slug: "a-b-testing-and-model-comparison", parent: "model-evaluation", title: "A/B Testing & Model Comparison", desc: "Enterprise-grade model comparison workflows for confident release, upgrade, and deployment decisions", icon: "speech", bg: "#002042", fg: "#30AA74" }, // navy-deep + green { id: "05", slug: "domain-specific-benchmarking", parent: "model-evaluation", title: "Domain-Specific Benchmarking", desc: "Production-readiness benchmarks for evaluating AI performance in defined business workflows and domains", icon: "layers", bg: "#000000", fg: "#EEE36B" }, // black + yellow { id: "06", slug: "human-in-the-loop-validation", parent: "model-evaluation", title: "Human-in-the-Loop Validation", desc: "Governed human validation workflows for consistent, auditable AI quality control", icon: "orbit", bg: "#3CA276", fg: "#0E1F3E" }, { id: "07", slug: "agentic-ai-evaluation-and-workflow-validation", parent: "model-evaluation", title: "Agentic AI Evaluation & Workflow Validation", desc: "Structured human-in-the-loop evaluation for reliable, safe, and production-ready AI agent workflows", icon: "orbit", bg: "#3CA276", fg: "#0E1F3E" }]; ; // green-soft + navy // Shared "Learn more" text link — sits at the bottom of every card. const LearnMoreLink = ({ href = "#" }) => Learn more ; // Shared icon plate — full-height navy column on the LEFT of every // card. Icon sits near the top, number anchored top-left in cream. // Width scales between full and half cards. const IconPlate = ({ s, width }) =>
{s.id}
; // ── FULL CARD — horizontal, wide navy column + content ────────── const FullCard = ({ s }) =>
{s.title}

{s.desc}

; // ── HALF CARD — horizontal, navy column left, content right ───── // Two halves stacked vertically = same overall height as one full card. const HalfCard = ({ s }) =>
{s.title.replace(/\n/g, " ")}

{s.desc}

; return (
); }; // ── SUB-SERVICES — variant B ─────────────────────────────────────── // Alternate card design for the same six services. No icon plate // background; the icon sits top-left of the card in the service's // brand color. The number moves to the bottom-left and pairs with // a "Learn more" affordance on the right that alternates between // the text-link and the older pill button, every other card. const SubServicesAlt = () => { const services = [ { id: "01", slug: "comprehensive-llm-evaluation", parent: "model-evaluation", title: "Comprehensive LLM Evaluation", desc: "Enterprise-grade evaluation for assessing LLM quality, reliability, safety, and production readiness", icon: "18", color: "#0E1F3E" }, { id: "02", slug: "multilingual-evaluation", parent: "model-evaluation", title: "Multilingual Evaluation", desc: "The cross-language, comparative view of how a model performs across many languages, locales, and regions", icon: "24", color: "#0E1F3E" }, { id: "03", slug: "in-language-model-validation", parent: "model-evaluation", title: "In-Language Model Validation", desc: "The deep, single-market view of whether a model is ready for deployment in a specific locale", icon: "05", color: "#0E1F3E" }, { id: "04", slug: "a-b-testing-and-model-comparison", parent: "model-evaluation", title: "A/B Testing & Model Comparison", desc: "Enterprise-grade model comparison workflows for confident release, upgrade, and deployment decisions", icon: "20", color: "#0E1F3E" }, { id: "05", slug: "domain-specific-benchmarking", parent: "model-evaluation", title: "Domain-Specific Benchmarking", desc: "Production-readiness benchmarks for evaluating AI performance in defined business workflows and domains", icon: "17", color: "#0E1F3E" }, { id: "06", slug: "human-in-the-loop-validation", parent: "model-evaluation", title: "Human-in-the-Loop Validation", desc: "Governed human validation workflows for consistent, auditable AI quality control", icon: "09", color: "#0E1F3E" }, { id: "07", slug: "agentic-ai-evaluation-and-workflow-validation", parent: "model-evaluation", title: "Agentic AI Evaluation & Workflow Validation", desc: "Structured human-in-the-loop evaluation for reliable, safe, and production-ready AI agent workflows", icon: "27", color: "#0E1F3E" }]; ; const Card = ({ s, idx }) => { return (
{s.title}

{s.desc}

Learn more
{s.id}
); }; return (
); }; // ── WHY ARGOS / DIFFERENTIATOR ───────────────────────────────────── // Dark navy section, mirrors the Homepage "What we stand for" rhythm. // Three column blocks — content = the three approved paragraphs of // the "Argos differentiator" section, lightly framed with derived // 2-word headings so the layout has structure. (Paragraph bodies // remain verbatim from SERVICE-DESCRIPTIONS.md.) const WhyArgos = () => { const blocks = [ { id: "The risk", body: "As AI systems move into production, evaluation becomes a business-critical control point. Enterprise teams need to understand whether models are accurate, consistent, safe, relevant, and fit for production across languages, domains, and scenarios. Surface-level testing is not enough when model behavior affects customer experience, operational trust, compliance exposure, and brand risk." }, { id: "Our approach", body: "Argos Data combines vetted human evaluators, multilingual and domain-specific expertise, and structured operational governance to support benchmarking, model comparison, regression testing, and ongoing performance monitoring. We define evaluation criteria, scoring rubrics, reviewer calibration, and adjudication workflows before evaluation begins. Programs are designed for repeatability — so evaluation evidence holds up across model versions, release cycles, and downstream review." }, { id: "Why it matters", body: "For enterprise AI teams, this connects evaluation directly to release readiness and production reliability — providing the human-reviewed evidence needed to move models from development to deployment with confidence." } ]; return (
Why Argos

Evaluation, treated as an engineered AI data operation.

{blocks.map((v, i) =>
{v.id}

{v.body}

)}
); }; // ── OUTCOME ──────────────────────────────────────────────────────── // Light section, eyebrow + headline + verbatim Outcome paragraph // from SERVICE-DESCRIPTIONS.md. const Outcome = () =>
Outcome

Reliable evaluation signals for measuring AI quality, safety, and release readiness.

Model Evaluation & Benchmarking provides enterprise AI teams with reliable evidence of model performance across quality, safety, language, domain, and user experience criteria. The result is clearer release decisions, stronger model reliability, reduced deployment risk, and a more disciplined path from model development to production AI. ---

; // Publish the six section components for the shared svc-template.jsx App. window.SVC_SECTIONS = { SolutionHero, Overview, UseCases, SubServicesAlt, WhyArgos, Outcome };