Contractor: IPPF Harm Mitigation Task Force AI/LLM Pilot Project

Background

IPPF is a global service provider and a leading advocate of sexual and reproductive health and rights (SRHR) for all. IPPF provides SRHR services in 151 countries and runs approximately 28,000 health service delivery points worldwide. It seeks to influence governments and other key decision-making bodies to make policy and legislative changes that support or defend SRHR. IPPF also conducts a range of education, awareness and empowerment programmes that supports its key mandate of SRHR for all.

Overview of the engagement

IPPF, in recognition of the urgent and significant changes in the funding landscape for sexual and reproductive health and rights (SRHR) programming, convened a global, multi-disciplinary Harm Mitigation Task Force (HMTF) to coordinate an effective response, to Member Associations (MAs). This led to channeling support through Harm Mitigation Emergency Grants, with the purpose of sustaining critical existing services, with a focus on harm reduction for the most vulnerable and excluded populations.

It is within this context that IPPF is seeking to engage a qualified consultant or agency to develop and customize an AI/LLM-based solution to strengthen the proposal review and grant-reporting-review process. The solution will be designed to assess the quality of MA proposals and reports, draw actionable insights from programme documentation, and support the HMTF in efficiently processing and analyzing HMTF proposals and reports generated from the delivery of the implemented programmes. There are an estimated 20 projects per funding cycle, with proposals and reports of no more than 10 pages each, plus historical review and reference documents of the same page length (e.g. Request for Proposals and Reporting guidance and templates).

Purpose

Reason the project is taking place

There is an urgent need for a scalable, accurate, and efficient solution to review and quality-assure these proposals and reports at pace, and to draw meaningful programme insights to inform decision-making. To address this, IPPF seeks to develop a Reports Review and Quality Assurance (QA) solution, built within the Microsoft Co-pilot Enterprise environment, to enable the HMTF to review multiple reports and proposals at scale, with speed and accuracy.

The solution would achieve the following:

• Operate within the IPPF SharePoint and Co-pilot environment, ensuring seamless integration with existing organisational systems and workflows;Function as a reliable reference tool, supporting structured and consistent review processes across the Secretariat;
• Handle English language reports and documents, with the potential to scale across IPPF's other official languages - French, Spanish, and Arabic - with accuracy and consistency;
• Integrate Explainable AI (XAI) mechanisms, ensuring that hallucinations and semantic collapse/inconsistencies can be identified, audited, and traced as document volumes scale; and
• Incorporate quality assurance dynamics through a design that enables ongoing integration with IPPF's QA frameworks and review standards.

Roles and Responsibilities

Contractor’s specific roles and responsibilities:

• Develop a progressive training pipeline, iteratively refining the model using HMTF report and proposal data to improve performance, contextual accuracy, and cross-domain reasoning;
• Test, identify, and mitigate semantic collapse, and strengthen the model's multi-modality and multi-data types cross-reasoning capabilities;
• Work closely with the HMTF team to ensure the report and proposal review environment attains high levels of accuracy and is developed in full alignment with the requirements set out in these Terms of Reference;
• Identify and implement robust solutions for monitoring, auditing, and addressing model decay or degradation in output quality over time, ensuring the long-term reliability of the solution;
• Lead the deployment of the solution within the IPPF SharePoint and Co-pilot Enterprise environment, ensuring seamless integration with existing organisational systems and QA workflows;
• Develop a comprehensive user manual and prompting guidelines, providing staff with clear, practical guidance on effective interaction with the AI/LLM solution;
• Design and deliver targeted training for staff who will be using the model (4-6 staff members) and system administrators
• Deploy a proof-of-concept prototype in a sandbox environment before a full roll out and functioning model / system;

IPPF’s specific roles and responsibilities: 

• Support the contractor in achieving agreed outcomes by providing appropriate guidance and direction
• Support testing, review, and improvement of the solution

Timeline

Commencement date:

Contractor services to commence July 2026

Proof of concept: Mid to late August

Minimum viable product: Late September to early October

• The solution includes, but is not limited to connectors (API’s) integration, prompt guidelines, user manual, maintenance documentation and monitoring processes
• Warranty period of 60 days following completion should be included to address model and integration errors.
• Ongoing support and maintenance should be addressed in additional quote (see section 7).

Milestones and deliverables

Key milestones are as follows:

1. Proof of concept (PoC) – Estimated level of effort: 15 consultancy days. 
 

• Submission and approval of a prototype system requirements specification (SRS) and system design document
• Completion of iterative model training, validation against test datasets, reports, spreadsheets, and formal evaluation of model accuracy, performance, and cross-reasoning capabilities across the HMTF’s target data repository.
• Functional integration of the prototype within the IPPF Secretariat Microsoft SharePoint environment, with connected and functional data pipelines capable of processing HMTF reports and documents in Word, PDF, and Excel formats.
   

2. Minimum Viable Product (MVP): Estimated level of effort- 20consultancy days.

• Fully working system integrated within Microsoft Copilot Enterprise developed in line with the approved SRS document
• Fully deployed Roll-out, and Staff Training both IT system management and staff users of the HMTF-Decision-Support lite system
• User guide and maintenance guide documentation.

Contractor requirements

Experience required:

• Demonstrated expertise in the design, development, and optimisation of deep learning (neural networks) systems
• Deep knowledge of multimodal LLM architectures, including the ability to process and reason
• Knowledge of Microsoft Co-Foundry for Multiple LLM deployments
• Knowledge of Co-pilot and OpenAI integration with office 365
•  Implementation of agentic AI frameworks for structuring quality assurance
• Agentic AI Expertise
• Knowledge in UI & UX designPrior experience developing applications with multilingual capabilities is desirable.

Skills:

• Project/Programme Management
• Excellent communication skills, oral, written and presentation
• User-centred training across diverse regions and languages 

Application information 

To apply for this consultancy, please submit a proposal responding to the requirements of the consultancy opportunity. Proposal(s) are to be submitted at  [email protected] with the subject line ‘Proposal for AI / LLM Solutions Architect’ no later than COB on 13th July 17:30 BST 2026. 

Proposals must include both a technical and financial proposal and must follow the format detailed in the attached ToR.