Array ( [0] => Array ( [_id] => 682491af2ad6dcc6e68a94b2 [name] => RFQ Response Evaluation Agent [description] =>

ZBrain RFQ Response Evaluation Agent automates the evaluation of vendor submissions across implementation, pricing, technical and qualification categories. Leveraging structured inputs from upstream screening agents and LLM-driven analysis, it delivers standardized evaluations and cross-vendor insights. This ensures transparent, audit-ready outputs that accelerate vendor selection while reducing manual effort and compliance risks.

Challenges the ZBrain RFQ Response Evaluation Agent addresses

Manual evaluation of RFQ responses is resource-intensive, fragmented and often prone to bias. Procurement teams struggle to consolidate evaluator remarks, interpret scores consistently and compare vendors objectively across categories. These challenges delay procurement cycles, increase the risk of subjective or inconsistent decisions and create compliance gaps. As RFQ response volumes grow, the lack of structured comparative analysis further erodes transparency, stakeholder confidence and timely vendor selection.

ZBrain RFQ Response Evaluation Agent uses an LLM to transform structured screening outputs into clear, standardized evaluation reports. The LLM consolidates evaluator remarks, generates document-wise assessments and synthesizes vendor-level narratives alongside cross-vendor insights in neutral, factual language. It also frames precise and unbiased recommendations, ensuring fair and audit-compliant evaluations. By automating this analysis, the agent reduces manual effort, accelerates procurement cycles and enables consistent, data-driven decisions at scale.

How the Agent Works

ZBrain RFQ response evaluation agent automates comparison of vendor RFQ submissions. Leveraging structured inputs from upstream agents and a large language model (LLM), the agent automates systematic evaluations and delivers comprehensive evaluation reports. Below are the detailed steps that define the agent’s workflow:

Step 1: Structured Input Data Ingestion

This step initiates the workflow. The agent receives structured evaluation data from the RFQ response screening compiler agent and prepares it for analysis.

Key Tasks:

• Structured data capture: The agent ingests vendor name, evaluation criteria, pass/partial/fail results, contextual remarks and overall scores.
• Input integration: Data is received through structured Google Sheets populated by the upstream screening agent, which are updated via webhook integrations.
• Category alignment: Ensures all inputs are mapped to the correct categories – implementation, pricing, technical and qualification – for consistent downstream evaluation.

Outcome:

• Evaluation data readiness: All vendor submissions are standardized and structured, ensuring they are ready for systematic comparative analysis.

Step 2: Comprehensive Analysis and Evaluation

The agent performs a detailed evaluation of structured inputs to produce factual, category-level and vendor-level insights.

Key Tasks:

• Document-wise evaluation: Reviews implementation, pricing, technical and qualification submissions and generates structured findings for each.
• Remark consolidation: Builds three-column evaluation tables (vendor | evaluation summary | score), consolidating evaluator remarks with pass/partial/fail indicators.
• Score interpretation: Interprets provided scores in context, highlighting risks where thresholds are not met.
• Vendor-level narratives: Synthesizes insights across categories to highlight each vendor’s strengths, weaknesses and consistency patterns.
• Cross-vendor insights: Compares vendor performance side by side, identifying relative advantages or gaps in neutral, factual language.

Outcome:

• Structured analysis outputs: Comprehensive evaluations at both the document and vendor level, supported by comparative insights that form the foundation for report generation in the next step.

Step 3: Detailed Report Generation

The agent compiles evaluation outputs into clear, structured reports designed for procurement teams.

Key Tasks:

• Report compilation: Compiles implementation, pricing, technical, and qualification analysis tables, along with vendor-level narratives and cross-vendor insights, into a unified evaluation report.
• Formatting and sectioning: Applies plain-text formatting and aligned three-column tables to ensure readability, auditability and dashboard compatibility.
• Cross-vendor summary generation: Groups insights vendor by vendor, presenting strengths, concerns and comparisons in clear, balanced language.
• Procurement-ready recommendations: Frames structured recommendations for each vendor, highlighting next-step considerations while maintaining clarity and factual accuracy.

Outcome:

• Comprehensive evaluation reports: Transparent, standardized and unified reports that present evaluation results in a user-friendly format, enabling informed and timely procurement decisions.

Step 4: Continuous Improvement Through Human Feedback

The agent incorporates user feedback to refine evaluation quality, improve report clarity and enhance overall learning.

Key Tasks:

• Feedback collection: Enables users to review generated reports, analyze gaps and provide feedback on accuracy, clarity and completeness.
• Feedback analysis and learning: The agent analyzes this feedback to identify recurring issues, formatting inconsistencies and areas needing improvement.

Outcome:

• Agent Improvement: The agent continuously improves by incorporating user feedback, ensuring its evaluation process remains accurate, consistent and aligned with evolving procurement requirements.

Why use RFQ Response Evaluation Agent?

• Faster procurement cycles: Accelerates vendor evaluation, enabling organizations to finalize procurement decisions with speed.
• Consistent and unbiased assessment: Delivers objective, fact-based vendor evaluations free from subjective bias, ensuring fairness and consistency.
• Cost efficiency: Reduces operational overhead by minimizing manual evaluation time, freeing procurement experts for higher-value strategic tasks.
• Process standardization: Establishes a standardized, enterprise-wide framework for vendor evaluation, reducing variability across teams and projects.
• Scalable vendor analysis: Processes large volumes of RFQ responses efficiently, ensuring accuracy and consistency even in high-volume, multivendor scenarios.
• Risk mitigation: Identifies gaps, compliance issues and performance concerns early, reducing the likelihood of vendor misselection.

[image] => https://d3tfuasmf2hsy5.cloudfront.net/assets/worker-templates/response-time-alert-agent.svg [icon] => https://d3tfuasmf2hsy5.cloudfront.net/assets/worker-templates/response-time-alert-agent.svg [sourceType] => FILE [status] => REQUEST [department] => Procurement [subDepartment] => Sourcing Management [process] => RFQ Response Handling [subtitle] => Automates evaluation of RFQ responses across key criteria, delivering structured, comparative reports to support procurement decisions. [route] => rfq-response-evaluation-agent [addedOn] => 1747227055592 [modifiedOn] => 1747227055592 ) [1] => Array ( [_id] => 68248c2a2ad6dcc6e68a5a31 [name] => RFQ Response Documents Retrieval Agent [description] =>

ZBrain RFQ Response Document Retrieval Agent automates vendor RFQ intake by filtering relevant emails, extracting and standardizing multi-format attachments, and converting them into metadata-rich documents, ready for seamless downstream evaluation without manual effort.

Challenges the RFQ Response Document Retrieval Agent Addresses

Manually processing RFQ emails is time-consuming and error-prone; teams must sift through messages, download attachments in various formats and manually extract critical details before evaluation can begin. Incomplete or malformed files create validation bottlenecks, while manual forwarding to screening systems introduces delays and inconsistencies. As RFQ volumes grow, these inefficiencies compound, risking missed deadlines and strained vendor relationships.

ZBrain RFQ Response Agent eliminates these pain points by auto-ingesting emails, using an LLM to confirm RFQ relevance, and validating, classifying, and extracting text from attachments using the best method. Extracted data is enriched with key metadata (RFQ number, project title, vendor name, contact details) and output as structured Markdown, then routed directly to the RFQ screening agent. This end-to-end automation removes manual bottlenecks, ensures data completeness, and accelerates procurement decisions with confidence and clarity.

How the Agent Works

ZBrain RFQ response document retrieval agent follows a structured, step-by-step process to automatically identify, extract, and prepare vendor-submitted RFQ response documents for downstream evaluation. Below is a detailed breakdown of how the agent streamlines the intake and pre-screening stages of the RFQ process.

Step 1: Email Ingestion and Relevance Checking

The agent begins by capturing incoming emails and validating whether the message is relevant to an RFQ submission.

Key Tasks:

• Email Trigger: A Gmail webhook activates the agent upon receipt of an incoming email.
• Email Field Extraction: A code component extracts essential details such as the subject, body text, and list of attachments.
• Relevance Check: An LLM analyzes the email content to determine whether the email pertains to an RFQ. Only relevant emails are passed forward.

Outcome:

• Automated RFQ Email Filtering: Non-relevant emails are filtered out, ensuring the workflow only processes valid RFQ submissions, reducing manual review efforts.

Step 2: Attachment Handling and Text Extraction

The agent examines each attachment in the email and extracts the necessary textual content for further processing.

Key Tasks:

• Attachment Processing: The agent processes each attached file individually in a loop.
• File Type Validation: The agent checks if the file is a supported format, PDF, Word (.doc/.docx), or Text (.txt). Unsupported types are flagged with an appropriate message.
• PDF Classification: If the attachment is a PDF, the agent determines whether it is a native (digitally readable) or scanned (image-based) PDF.
• Content Extraction:
  • Native PDFs: Text is extracted directly using a PDF-to-text utility.
  • Scanned PDFs: Converted into images and processed using a multimodal LLM to extract text.
  • Word/Text Files: Text is directly extracted.

Outcome:

• Accurate Multi-format Text Extraction: Each attachment is accurately interpreted and converted into usable plain text, regardless of input format.

Step 3: Key Metadata Extraction and Formatting

The extracted text is analyzed to retrieve key details and then structured into a standardized format for downstream processing.

Key Tasks:

• RFQ Detail Extraction: An LLM identifies and extracts key RFQ details from the text, such as:
  • RFQ Number
  • Project Title
  • Vendor Name
  • Contact Details
• Markdown Structuring: A dedicated LLM reformats the extracted text into well-structured Markdown, adding only formatting syntax without rewriting, summarizing, or omitting any content. This approach preserves the original structure and ensures clarity for subsequent processing stages.

Outcome:

• Metadata Enriched Structured Document: The extracted document is enriched with structured metadata and formatted in a consistent layout for efficient downstream consumption.

Step 4: Document Routing to Screening Agent

Once formatted, each document is routed to the downstream agent responsible for evaluation.

Key Tasks:

• HTTP POST Call: The agent sends each attachment individually via a POST request to the ZBrain RFQ response screening agent
• Input Transfer: The formatted content serves as the input for screening, allowing evaluation workflows to proceed without delay.
• Sequential Handling: Documents are processed one at a time to ensure precise alignment with the downstream agent’s input requirements.

Outcome:

• Efficient Evaluation Transfer: Processed documents are seamlessly transferred to the evaluation workflow, allowing the screening agent to begin scoring and validation.

Step 5: Submission Summary Compilation

Once all documents have been processed and routed, the agent compiles a consolidated summary for dashboard visibility.

Key Tasks:

• Summary Generation: A final LLM aggregates key metadata, document names and submission context from the processed attachments.
• Dashboard Output: The summary is displayed in the agent’s dashboard for review.
• Human Feedback Integration: Users review each submission summary, and their feedback iteratively fine‑tunes the agent, continuously increasing accuracy.

Outcome:

• Consolidated Submission Summary: A comprehensive submission summary is created, offering clarity on the number of attachments processed and the vendor-specific metadata, supporting visibility and downstream decision-making.

Why use RFQ Response Document Retrieval Agent?

• Time Efficiency: Automates the retrieval and processing of RFQ documents, reducing manual effort and accelerating response cycles.
• Accuracy: Extracts and preserves complete document content while accurately identifying key RFQ metadata.
• Scalability: Handles multiple attachments and high submission volumes, supporting enterprise-scale operations.
• Workflow Automation: Automatically routes processed documents to downstream agents, enabling end-to-end workflow automation.
• Error Reduction: Minimizes manual errors through automated classification, extraction, and validation steps.
• Transparency: Provides real-time visibility into processed submissions through dashboard summaries.

[image] => https://d3tfuasmf2hsy5.cloudfront.net/assets/worker-templates/contract-expiry-alert-agent.svg [icon] => https://d3tfuasmf2hsy5.cloudfront.net/assets/worker-templates/contract-expiry-alert-agent.svg [sourceType] => FILE [status] => REQUEST [department] => Procurement [subDepartment] => Sourcing Management [process] => RFQ Response Handling [subtitle] => Automatically filters Gmail for RFQ emails, extracts document content, and shares it with the RFQ Screening Agent for streamlined processing. [route] => rfq-response-documents-retrieval-agent [addedOn] => 1747225642628 [modifiedOn] => 1747225642628 ) [2] => Array ( [_id] => 6824886a2ad6dcc6e68a41a6 [name] => RFQ Response Screening Compiler Agent [description] =>

ZBrain's RFQ Response Screening Compiler Agent automates the classification and evaluation of RFQ response documents across key categories like pricing plan, implementation plan, technical plan, and qualification plan. By leveraging a Large Language Model (LLM), it ensures faster, rules-based scoring and audit-ready outputs, streamlining vendor shortlisting while improving compliance and consistency.

Challenges the ZBrain RFQ Response Screening Compiler Agent Addresses

Manual RFQ screening is slow and error-prone, often causing inconsistent classifications, missed evaluation criteria, and delays in vendor selection. These issues create procurement bottlenecks, heighten compliance risks, and reduce transparency, especially as response volumes increase. Such inefficiencies extend procurement cycles, hinder data-driven decisions, and ultimately impact project timelines and vendor relationships.

RFQ Response Screening Compiler Agent delivers fast, objective, and auditable assessments by automatically categorizing and consistently scoring RFQ responses. Results are output directly into the appropriate Google sheet, minimizing errors and freeing procurement teams to focus on supplier relationships and strategic initiatives. By reducing manual intervention, the agent ensures every vendor is evaluated fairly and efficiently, boosting procurement agility, strengthening compliance, and enabling teams to focus on higher-value work.

How the Agent Works?

RFQ response screening compiler agent automates the classification and evaluation of RFQ responses across key categories. Leveraging an LLM, the agent classifies RFQ response document type, applies standardized scoring logic to vendor submissions, and compiles all evaluation results into structured, audit-ready reports. Below, we outline the detailed steps that define the agent's workflow:

Step 1: RFQ Response Details Intake and Classification

This step initiates the workflow. The agent receives input for each vendor RFQ response from upstream agents and ensures each response is routed to the correct evaluation category within the integrated Google Sheets.

Key Tasks:

• Structured Response Intake: The agent receives input for each vendor response—including document type (Implementation Plan, Pricing Plan, Technical Plan, or Qualification Plan), vendor name, and screening status—from the RFQ response screening agent, which analyzes all incoming submissions. It also receives the evaluation criteria from the RFQ response screening rules creation agent.
• Response Category Mapping: Leveraging an LLM, the agent reverifies the response type, ensures it aligns with one of the four response categories (Implementation Plan, Pricing Plan, Technical Plan, Qualification), and routes it to the appropriate Google Sheet tab. This step ensures accurate categorization and prevents misclassification from any upstream errors.
• Validation: Ensures that each type matches an allowed category; if an unrecognized or irrelevant type is received, the agent displays an appropriate message.

Outcome:

• Category Assignment: Each document type is accurately mapped to its designated Google sheet tab category, ensuring all subsequent evaluations apply the correct criteria.

Step 2: Response Evaluation

Once classified, the agent conducts a detailed, rules-driven evaluation using criteria created upstream by the RFQ response screening rules creation agent.

Key Tasks:

• Evaluation Criteria Retrieval: The agent references the ordered evaluation criteria from column names in Row 1 of the evaluation sheet, provided by the RFQ response screening rules creation agent for the specific category.
• Score Assignment: The agent uses an LLM to evaluate each vendor response strictly according to the screening status: Pass (1 point), Partial (0.5 points), Fail (0 points). If a criterion is present in headers but not in the screening status, its value is left blank and excluded from scoring.
• Blank/Missing Handling: Blank or missing responses in screening status are treated as Fail (0 points). If the criterion is not in screening status, the cell remains blank and does not count toward the score calculation.
• Overall Score Calculation: The agent computes the overall score as a percentage (Total Points Earned / Total Criteria Evaluated) × 100, rounding to the nearest integer and returning as a percent string (e.g., "94%").

Outcome:

• RFQ Response Scoring: Vendor responses are objectively scored against standardized, rules-based criteria, producing transparent results for downstream compilation.

Step 3: Output Generation

The agent compiles and structures all evaluation results for downstream review and reporting.

Key Tasks:

• Structured Output Creation: Consolidates each evaluated response into a clean JSON object, precisely matching Google Sheet columns.
• Comprehensive Reporting: Generates a report for each RFQ response that includes the document type, vendor name, screening criteria, and overall evaluation score (as a percentage).
• Automated Sheet Entry & Link Sharing: Populates scoring outputs directly into the appropriate Google Sheet tab (e.g., Implementation Plan, Technical Plan) and provides a direct link to the updated sheet for traceability.

Outcome:

• Streamlined Vendor Shortlisting: Procurement teams receive real-time reports containing evaluation scores, document type, vendor name, and direct access to the compiled results in Google Sheets, enabling rapid, transparent, and informed vendor selection.

Step 4: Continuous Improvement Through Human Feedback

The agent incorporates user feedback to refine evaluation accuracy and align with evolving procurement requirements.

Key Tasks:

• Feedback Collection: Allows users to review and annotate evaluation results for clarity, relevance, or alignment with procurement standards, helping flag unclear scoring, missing logic, or areas needing improvement.
• Feedback Analysis and Learning: The agent reviews submitted feedback to identify and address recurring issues, such as inconsistent scoring or overlooked evaluation criteria.

Outcome:

• Agent Enhancement: The agent continuously improves by incorporating human feedback, ensuring its evaluation process remains accurate, consistent, and aligned with changing business requirements.

Why use ZBrain's RFQ Response Screening Compiler Agent?

• Accelerated Vendor Scoring: Automatically classifies and evaluates RFQ responses, significantly reducing turnaround time for vendor shortlisting.
• Enhanced Evaluation Consistency: Applies LLM-driven scoring logic to ensure all vendor responses are assessed objectively and in line with procurement standards.
• Audit-ready Results: Delivers structured, machine-readable outputs with transparent scoring, supporting compliance and simplifying downstream audits.
• Reduced Manual Intervention: Minimizes the need for procurement teams to interpret responses or manage complex scoring logic manually.
• Scalable Processing: Efficiently handles large volumes of RFQ responses across multiple categories without compromising accuracy or speed.
• Enhanced Transparency for Stakeholders: Provides clear scoring and documentation, giving all stakeholders visibility into vendor decisions.

[image] => https://d3tfuasmf2hsy5.cloudfront.net/assets/worker-templates/response-suggestion-agent.svg [icon] => https://d3tfuasmf2hsy5.cloudfront.net/assets/worker-templates/response-suggestion-agent.svg [sourceType] => FILE [status] => REQUEST [department] => Procurement [subDepartment] => Sourcing Management [process] => RFQ Response Handling [subtitle] => Automates scoring of RFQ responses, classifying vendor documents and updating evaluation results in a structured Google Sheet for seamless vendor selection. [route] => rfq-response-screening-compiler-agent [addedOn] => 1747224682852 [modifiedOn] => 1747224682852 ) )

Procurement

RFQ Response Evaluation Agent

Automates evaluation of RFQ responses across key criteria, delivering structured, comparative reports to support procurement decisions.

Procurement

RFQ Response Documents Retrieval Agent

Automatically filters Gmail for RFQ emails, extracts document content, and shares it with the RFQ Screening Agent for streamlined processing.

Procurement

RFQ Response Screening Compiler Agent

Automates scoring of RFQ responses, classifying vendor documents and updating evaluation results in a structured Google Sheet for seamless vendor selection.