TOR for Onboarded Agency for Quality Control Protocols of AI Platform for SELCO Foundation

 

Title	Title Quality Control Protocols of AI Platform for SELCO Foundation
Timeline	Timeline 06 Months
Expected area of expertise	Expected area of expertise • Experience in developing products based on the Gen AI Platform • Experience in AI quality assurance, content curation • Expertise in AI Testing Professional
Apply Googe Link	Apply Googe Link https://forms.gle/n5YEt6zYwZbCBU8Z8 (Contact Procurement for Form Link)
Last Date of Apply	17th November, 2025

About SELCO Foundation:

SELCO Foundation is a Not for Profit, based out Bangalore, India which works at the intersection of development, clean energy and climate. On one hand it innovates new solutions which can augment core development sectors like healthcare delivery and improved livelihoods, and on the other hand it builds systems which help the solutions scale successfully. Being a practitioner driven knowledge and technical expert, SELCO Foundation firmly believes that an end-user centric, demand driven approach that focuses on ecosystem building, leads to interventions that are replicable and long term (Read more about SELCO here: http://www.selcofoundation.org/ )

Program Overview:

SELCO Foundation for the past two years, has been pilot-testing a GenAI platform in collaboration with Apurva.ai. It is aimed to explore the power of leveraging technology to incubate transformative systems change at scale i.e. building systems of change that provide timely information for complex decision-making for organizations, by reducing the time taken to arrive at those decisions. This initiative aims to prove out the potential of scaling a learning methodology in the development sector, which can balance diverse views and agile decision making.

The Gen AI platform powered by Apurva.ai is an open innovation platform built specifically to help enable social change leaders to amplify the collective wisdom of their networks. This means enabling the organizations to tap into the rich theoretical, and subject-matter knowledge and expertise, and practical, on-ground experiences of the other organizations in their respective networks and learning from their perspectives and narratives regarding addressing complex problems in the development sector. This also serves as a unique platform that triangulates knowledge real-time. In partnership with Bhashini (BHASha Interface for India, an initiative by the Indian government under the Digital India program with the primary goal to break down language barriers in India by leveraging artificial intelligence (AI) and natural language processing (NLP) technologies), it offers processing of 22 languages of India with the aim to include more as the platform progresses.

This Gen AI platform has integrated systems thinking framework and design thinking in its inference engine, enabling a holistic learning mechanism even if the queries come from certain biases. Currently the following products fall under the purview of the Gen AI platform that is customisable based on needs, data sizing etc.

Cocreate/ Conversations:

Participatory AI designed to be part of meetings, workshops, roundtable conferences to offer systems thinking informed perspectives, novel ideas and customisable insights.

Wisdom/ Insights:

• Query-able platform based on curated knowledge inputs (audio files, videos, reports, research materials). This also has a WhatsApp integration as an extension.

Impact:

• Customizable WhatsApp-based survey tool

Goal and Objective

 All AIs require significant effort in QC and training. Through this project and selected vendor SELCO will leverage on learnings so far to customise methodologies in the sector for its purpose and create appropriate SOPs.

• To collectivise knowledge from the sector that enables interdisciplinary and intersectional developmental action
• To equip problem solvers and program designers with actionable knowledge that will reduce costs and time taken to research
• To develop a physical center that is committed to learning and knowledge by constantly engaging with relevant ecosystem players to generate and share knowledge

Scope of Work:

The agency sought will specialize in comprehensive quality assurance for SELCO Foundation’s initiative Platform for Collective Wisdom – a collective knowledge platform for the development sector powered by Apurva.ai --a development sector focused curation based GPT ecosystem. The agency will rigorously audit and curate the knowledge ingested, continuously monitor GPT outputs for accuracy and against misuse, analyze user interactions to detect risks, and enforce best practices for content and system quality. The agency will provide detailed reports and actionable insights enabling ongoing refinement and safeguarding the GPT’s content reliability, security, and alignment with organizational needs.

Knowledge Curation QC

• Conduct rigorous quality audits on the organization’s ingested knowledge to ensure accuracy, consistency, completeness, and relevance as per Apurva.ai platform curation compatibility and guidelines.
• Validate that content is aligned with platform policies like terms of use, and compliance requirements
• Develop and implement standardized quality check protocols for content ingestion.
• Identify knowledge gaps, redundancies, and potential sources of misinformation in the curated datasets.

Platform Output Quality Monitoring

• Continuously monitor the GPT’s generated outputs for accuracy, coherence, context relevance, and alignment with organizational standards and platform guidelines.
• Detect and report instances of hallucination, misinformation, or inappropriate response patterns.
• Provide recommendations to refine output filters, guardrails, and response tuning to minimize errors and misleading information.
• Establish testing frameworks that simulate real-world user interactions for ongoing model validation.

User Interaction Analysis and Misuse Prevention

• Analyze how users interact with the GPT-generated information to identify usage patterns, potential misunderstandings, or misuse scenarios.
• Anticipate and simulate scenarios where platform misuse or jailbreak attempts could compromise system integrity or expose sensitive data
• Propose user access controls, content delivery restrictions, and operational guardrails tailored to mitigate organizational risk

Best Practices Development and Enforcement

• Work closely with the organization to co-create and continuously refine best practices for knowledge curation, model usage, and output quality assurance.
• Develop guidelines for identifying sensitive or critical information and methods for its secure handling within the GPT environment.
• Assist the organization in enforcing adherence to these best practices across relevant teams.

Collaborative Reporting and Insights Generation

• Generate reports based on aggregation of findings from quality audits, platform output monitoring, and user behavior analysis.
• Generate detailed, actionable reports summarizing:
• Quality assurance outcomes and identified risks o Effectiveness of current guardrails and control measures
• Recommendations for content improvement
• Potential vulnerabilities and misuse trends

Cross-Entity Coordination, Documentation and Training Support

• Liaise regularly with Platform for Collective Wisdom to align objectives, clarify requirements, and escalate critical issues.
• Keep abreast of industry developments, compliance regulations, and AI ethics standards to inform quality assurance processes.
• Document quality testing protocols, curation standards, user guidelines
• Support the organization in capacity-building efforts by training relevant personnel on quality assurance frameworks and risk mitigation strategies.

Selection Criteria:

A qualified agency should have demonstrated expertise in AI quality assurance, content curation, risk assessment, and AI ethics, with experience auditing large language models or enterprise-grade AI systems. Certifications such as "AI Testing Professional," "AI Quality Assurance," or relevant data security credentials add credibility, and a portfolio evidencing prior work in knowledge platform quality control or AI/ML safety will be essential.

Proposal

 For proposal submission, the agency should provide:

• A step-by-step methodology outlining their approach to knowledge curation audits, GPT behavior validation, user risk analysis, and mis/abuse scenario testing.
• A detailed plan for continuous monitoring, feedback, and development of quality check protocols tailored for each organization's needs.
• A Human Resource allocation plan to fulfil the entire scope of work for a period of 6 months along with profiles of team members
• Clear formats for reporting findings and actionable recommendations
• Please provide your proposal and quotation as per deliverables based and mentioned program requirements. Capture Timelines and Split cost wherever possible.
• SELCO Foundation reserves the right to engage agencies at its discretion. Engagement of multiple agencies as per core expertise may also be considered

 For pricing, the proposal should break down the unit cost for performing a full quality assurance cycle for one organization, and specify incremental pricing for extending services to each additional organization (often a lower per-unit rate due to process standardization and system familiarity)

Timeline:

Duration of the Program: The program is planned up to 6-7 months

• The service providing agency is expected to commence the detailed scope (outlined above) of for 10 instances starting November 2025,
• The contract for fulfilling the scope of work in Phase 1 will be for a period of 6 months (November 2025-June 2026) following which the terms of engagement and scope of work for a Phase 2 can be mutually explored.

Payment Terms:

Payments will be made based on the completion of agreed deliverables. A payment schedule will be included in the contract and based on considered each party agrees.

To Apply:

Interested consultants / organisations, meet the eligibility criteria with relevant experience and is interested in collaborating with the SELCO Foundation, (please include samples and/or references of the previous similar work as proof of experience) and based out of India are requested to reach out with a detailed proposal giving a brief on the methodology and the process they will uptake for this project, including budgets (with break-ups and explanation), timelines and milestones and submit the same to google form https://forms.gle/cDUh1CHZuTSPijtC9 on before 17th November, 2025.

Any further queries please write to procurement@selcofoundation.org with a subject line: “Quality Control Protocols of AI Platform for SELCO Foundation” (Name of Project)

Note: The application will be evaluated on a rolling basis.

Tags Innovations and Technology Research Others