Straive: Agentic AI and Document Processing

Straive operates as an agentic AI platform company ranked #1 for operational AI deployment, offering intelligent document processing through SPARK.AI and multi-LLM capabilities via the LLM Foundry and aiKira ecosystem.

Overview

Straive entered 2026 with back-to-back analyst recognitions validating its repositioning from content services heritage toward production-grade agentic AI, while simultaneously managing a material security disclosure that will feature in regulated-industry due diligence files.

The company traces its origins to Saztec Philippines Inc. (founded 1980), rebranded through SPi Technologies and SPi Global before EQT's $1B acquisition in August 2021 prompted the Straive rebrand in April 2021. In December 2025, Namit Sureka, President and Chief Analytics & AI Officer, summarized the platform philosophy: "We look at AI as a continuum. The real impact happens when you operationalize it by combining traditional models, machine learning, domain expertise, and GenAI in a way that works for everyday decision-making."

Two independent analyst bodies have since validated that framing. AIM Research named Straive a 2025 PEMA Leader for agentic AI service providers, citing modular architecture and orchestration maturity across banking, financial services, healthcare, retail, education, and publishing. Separately, HFS Research's inaugural agentic services evaluation assessed 36 service providers across the agentic value chain between September 2025 and January 2026, introducing a "SaS Stars" category for providers operationalizing scalable, technology-led delivery with reduced human intervention. Straive's HFS Challenger designation positions it as "a credible challenger to traditional tier-1 providers" in AI, data analytics, and digital transformation, with intelligent document processing explicitly listed alongside the Content Hub, DataHub, and Kadal AI workbench.

HFS identified two specific weaknesses: the legacy SPi Global brand still overshadows current capabilities, and integration of the Gramener, Metonymize, and SG Analytics acquisitions remains incomplete. The SG Analytics deal, completed in June 2025, was the 12th acquisition in an aggressive consolidation strategy. That integration gap is the most actionable competitive signal for buyers evaluating Straive against tier-1 alternatives.

In a February 2026 DataQuest India interview, CEO Ankor Rai argued that GenAI pilots fail because enterprise content is "fragmented, inconsistently structured, poorly contextualised, and not governed for machine consumption." Straive's response is to treat content architecture as infrastructure before deploying AI on top of it. Rai claims this approach has achieved AI quality exceeding 99% in Life Sciences and financial services; no named client, methodology, or third-party validation supports that figure in the public record.

Against that positioning, RedPacket Security reported on 2026-02-27 that the QILIN ransomware group listed Straive on its dark web leak page. QILIN operates a double-extortion model, encrypting systems and threatening to publish stolen data. No ransom amount, exfiltration confirmation, breach scope, or operational impact has been disclosed. Straive issued no public statement as of the reporting date. In regulated industries where data governance is a procurement prerequisite, prospective clients in Life Sciences and financial services will have this incident in their due-diligence files regardless of eventual outcome.

The company has raised $353M across 13 financing rounds and maintains IPO registration status. Baring Private Equity Asia holds the firm privately.

How Straive processes documents

Straive's document processing argument rests on content architecture preceding model deployment. Rai's public framing is explicit: "Pilots that progress treat content as infrastructure, investing in structure, metadata, contextual signals, and governance so knowledge remains usable as it evolves." The platform stack operationalizes that principle across three layers.

The SPARK.AI Platform handles AI-powered extraction and automation at the core. On top of it, the LLM Foundry provides a model-agnostic GenAI layer supporting OpenAI, Anthropic, Gemini, and Grok with pre-built modules for extraction, classification, and sentiment analysis, claiming deployment in days rather than months. The aiKira AI Ecosystem operates through modular service-oriented architecture deployed across editorial, production, and post-publication workflows, with vertical specialization in science publishing.

The DocExplore Platform combines ML and LLM-powered document analysis, claiming up to 80% reduction in manual effort. A risk-tiered human-in-the-loop model routes only low-confidence or high-risk outputs to human reviewers and reduces future intervention through feedback loops. Rai's explicit rebuttal to BPO-style augmentation: "A production-grade human-in-the-loop model is not an extension of BPO but an operating capability combining domain expertise, governance, and system learning to support intelligent systems reliably."

AIM Research's PEMA Leader recognition specifically cites "orchestration maturity" as a differentiator, which in practice means Straive can integrate multiple AI agents and LLMs into sequential document processing pipelines covering intake, classification, extraction, validation, and enrichment. That multi-agent coordination capability is what separates the platform from single-model extraction tools.

Two case studies carry specific numbers. A graph engine using Breadth-First Search to detect multi-hop fraud clusters delivered a 27% reduction in fraud losses, approximately $3.5M in annual savings, for a global payments provider. AI embedded in communication workflows for research and publishing clients produced 23-25% increases in collection rates and 20-25% reductions in delinquency. Three additional deployments covering pharma pharmacovigilance, a major US sports league fan-data unification, and logistics document conversion are cited directionally without quantified outcomes. No pricing, model strings, benchmark comparisons, or GA/beta status are publicly available for any Straive platform.

India delivery centers are explicitly reframed by Rai as "core capability centres supporting the full AI lifecycle" covering data preparation, analytics, AI workflow design, output validation, and exception handling. HFS's concurrent finding that Straive lacks "consultative depth" in strategic enterprise conversations suggests the market has not yet accepted that repositioning.

Use cases

Financial services and fraud detection

BFSI organizations use Straive's graph-engine approach to detect multi-hop fraud clusters, with a documented 27% fraud-loss reduction ($3.5M annually) for a global payments provider. Regulatory filing and financial statement processing runs through specialized AI modules with domain expert validation and capital markets analysis. The QILIN ransomware listing is a live procurement risk for financial services buyers conducting vendor due diligence.

Science publishing automation

Publishers deploy Straive's aiKira ecosystem for editorial workflows including peer review, production, and post-publication analysis through modular architecture. AI embedded in research and publishing communication workflows has produced 23-25% increases in collection rates and 20-25% reductions in delinquency for publishing clients.

Agentic AI for regulated industries

Enterprises in Life Sciences and pharma operationalize AI through Straive's model-agnostic platforms combining traditional ML, domain expertise, and GenAI. Pharmacovigilance document processing is cited as a deployment vertical; quantified outcomes are not yet in the public record. AIM Research's PEMA Leader designation, covering banking, financial services, healthcare, retail, education, and publishing, confirms the breadth of vertical traction. Straive's "AI success depends more on operating discipline than technological sophistication" positioning targets organizations where governance and auditability outweigh raw model performance.

Technical specifications

Resources

• Website
• Agentic AI and GenAI solutions
• Science and research content platform
• AIM Research PEMA Leader recognition
• HFS Research Challenger report
• HFS Horizons Agentic Services 2026
• CEO interview: content as AI infrastructure
• QILIN ransomware listing by RedPacket Security

Company information

Headquarters: Singapore

Founded: 1980 (as Saztec Philippines Inc.)

Former names: SPi Technologies Inc. (1996), SPi Global (rebranded to Straive April 2021)

Employees: 18,000 across 19 global centers

Geographic presence: Philippines, India, USA, China, Nicaragua, Vietnam, UK, Singapore

Acquisition: EQT acquired company for $1B in August 2021; currently held by Baring Private Equity Asia

Funding: $353M raised across 13 financing rounds

IPO status: Registered

Previous ownership: PLDT (until 2017)

Key acquisitions: SG Analytics (June 2025), LearningMate (2020), Gramener (2023), Metonymize (2023)

Security incident: QILIN ransomware group listed Straive on dark web leak page 2026-02-27; no public statement issued as of that date