Accelerator Use Case
Peek Assist, an AI-powered WhatsApp assistant for frontline eye health screeners.
Type:
Social enterprise
Program:
Turn.io Accelerator
Focus:
Eye health
Location:
United Kingdom

Over one billion people live with vision loss, and most cases are treatable.
Using novel software in schools, communities, and workplaces, Peek Vision has already helped eye health providers screen more than 18 million people and connect nearly 2 million to care across 13 countries. Along with this rapid growth in a short timeframe, pressure grew on human-led onboarding and support of frontline eye health screeners, necessitating a more scalable approach.
Work through the Accelerator focused on creating a scalable, self-serve support layer for frontline workers that extends, rather than replaces, human support teams. Peek Vision launched a WhatsApp assistant that delivers instant answers grounded in curated internal knowledge resources, supporting workers in English, Hindi, Swahili, and Urdu.
The service provides on-demand guidance through structured text and is deploying voice and multimedia formats, including video tutorials, to improve learning and issue resolution.
Work through the Accelerator focused on creating a scalable, self-serve support layer for frontline workers that extends, rather than replaces, human support teams. Peek Vision launched a WhatsApp assistant that delivers instant answers grounded in curated internal knowledge resources, supporting workers in English, Hindi, Swahili, and Urdu.
The service provides on-demand guidance through structured text and is deploying voice and multimedia formats, including video tutorials, to improve learning and issue resolution.
Impact
18M+
People screened
2M+
Connected to care
13
Countries reached
Demo:
Watch Peek Assist in action
Routine queries are handled automatically, while more complex cases are routed to human operators through integrated workflows.
The system was tested with 142 users across five countries over two months, with strongly positive feedback. A core lesson was that high-quality, structured training content is essential for reliable AI performance.
A core lesson was that high-quality, structured training content is essential for reliable AI performance.
Peek Vision is tracking resolution rates, content usage, and response times, with success defined by faster problem resolution and reduced workload for central teams.
The organisation estimates each team member could gain at least 130 hours of productivity per quarter, time that can be redirected to onboarding new partners and expanding screening reach as programs scale across countries.
Peek Vision is tracking resolution rates, content usage, and response times, with success defined by faster problem resolution and reduced workload for central teams.
The organisation estimates each team member could gain at least 130 hours of productivity per quarter, time that can be redirected to onboarding new partners and expanding screening reach as programs scale across countries.

LLM Models
used
Chat GPT 5.0 Nano
Chat GPT 4.1 Mini

AI
Built through the Chat for Impact Accelerator
This project was developed through the Chat for Impact Accelerator, a global programme supporting organisations to design, test, and scale impactful conversational services on WhatsApp.
Through the programme, teams work closely with product and technical experts to prototype, launch, and iterate on AI-powered chat solutions that expand access to health, education, and essential services for communities around the world.
Through the programme, teams work closely with product and technical experts to prototype, launch, and iterate on AI-powered chat solutions that expand access to health, education, and essential services for communities around the world.


