Seeds of Promise| An AstraJax initiativeFund the pilot

Turn access into agency.

A practical pilot for Seeds of Promise: connect the computer centre, capture the community context, then give local leaders narrow AI agents that turn questions into usable coaching, plans, and proposals.

Partner with us Read the plan

Matthew Hopkinson coaching community leaders at a chalkboard session in Malawi — AI coaching with community leaders, Malawi

Community leaders seated for an AI workshop in Malawi — Community workshop, Malawi

The need

Short of access, not ambition.

Seeds of Promise already has the rare ingredients: trusted leadership, a ready community, and a computer-centre vision. The missing layer is access that actually works.

Young people have limited routes into digital learning and guided practice
Expert advice often means long journeys, scarce visitors, or word of mouth
Farming, greenhouses, tailoring, and fundraising need specialist support
Generic tech training rarely fits the language, culture, or local constraints

The problem is not lack of vision. It is that knowledge, coaching, and specialist support are too far away from the people who could use them.

Seeds of Promise computer room with desktop workstations — Existing computer centre - ready for the next layer

Layer 1 - Infrastructure

Make the computer centre usable.

This is the part funders can see immediately: a specific community in Malawi has the will and the space. The pilot helps make that space reliable enough for learning and work.

Connectivity, likely Starlink or the best local option
Power resilience, including solar where needed
Usable laptops, tablets, routers, and supporting equipment
A maintainable computer-room setup with clear local ownership

Facilitators leading a community planning session at a flipchart — Understanding context before building agents

Layer 2 - Context

Codify the local reality.

Once the room works, AstraJax does the work that makes AI useful: collecting the context that a generic tool will never arrive with.

Train local champions who can own the day-to-day use
Capture real goals, available resources, local language, costs, and constraints
Translate outside expertise into examples the community can actually use
Keep people in control, with AI supporting rather than replacing judgement

This is the AstraJax connection: the foundation changes, but the principle does not. Structure the work. Structure the context. Then let agents use both.

Matthew Hopkinson and a community leader during an outdoor coaching session — Coaching shaped around local context

Layer 3 - Agentic support

A small specialist layer, shaped around Seeds.

The unique layer is not a chatbot and not a one-off training day. It is a set of narrow, context-aware agents that local champions can use for specific jobs.

A learning coach that helps young people practise questions, digital skills, and explanations at the right level
A greenhouse and farming coach grounded in local crops, climate, resources, and project goals
A fundraising agent that can turn a Chichewa or English idea into donor updates, proposals, and pitch material
A small-enterprise planner for tailoring, pricing, stock thinking, and simple business plans
A community-leadership planner that helps elders compare options, prepare meetings, and structure decisions

The agents do not replace local leaders. They bring the specialist closer, ask better questions, and package guidance in a form the community can act on.

Why this is different

The agentic layer is the bridge between outside expertise and local action.

Not generic AI training

Training teaches people what AI is. The agentic layer gives them practical helpers for real jobs: learning, farming, fundraising, enterprise, and leadership.

Not loose chatbot access

Each agent is narrow, grounded in approved local context, and designed around human judgement. It proposes, explains, and coaches - people decide.

Not Matthew as the bottleneck

The goal is local champions using a toolkit that carries the method forward without waiting for the next visitor, Zoom call, or outside expert.

The AstraJax proof in another setting

In commercial teams, AstraJax turns messy work and scattered data into AI-ready operating systems. Here, the same method turns access, context, and trust into community-ready AI support.

Children playing together at sunset in Malawi — Access to learning and tools for the next generation

What it unlocks

Democratise context-aware education.

This is the practical version of the bigger promise: specialist support at the point of need, adapted to the person and the place.

Education - learning support and digital skills adapted to level and language
Farming and greenhouses - guidance at the point of need, not a two-hour journey
Fundraising - local vision shaped into proposals and donor-ready narratives
Small enterprise - business planning and grant support for tailoring and local trade
Community leadership - plans and options without waiting for external experts

Even when experts are willing to help, their advice often has to be translated across culture, language, resources, and lived reality. The agentic layer does that packaging work again and again.

Community event at Seeds of Promise with local leaders and youth — Seeds of Promise community gathering

Malawi

A known community with existing vision.

Seeds of Promise is not an abstract beneficiary. Matthew reached the community through Links, a UK charity with long-standing partner relationships in Malawi and across Africa.

Matthew visited in October 2025 and delivered an introductory AI coaching session
Strong local appetite, with feedback that content needs deeper local adaptation
Introduced and supported through Sam at Links, who knows the community and the wider partner network
The next step is validating the delivery model, not whether the community cares

Large group photo of Seeds of Promise community members and visitors — Seeds of Promise - October 2025

Proof on the ground

Already more than an idea.

This work has started. What comes next is a focused pilot: connect, equip, capture context, build the first agents, train champions, and document the playbook.

Coaching sessions rooted in local setting, language, and community needs
Training materials shared for continued use
Computer-centre vision and leadership ready for the infrastructure layer
A model that could travel through Links-connected partners if the pilot works

Trusted partner route

Routed through Links, not invented from scratch.

This pilot sits inside an existing charity relationship. That matters for funders: Seeds of Promise is a known community, reached through a partner organisation that already works for lasting change.

Who Links are

Links

Links is a UK charity that connects people, confronts poverty, and creates potential by working with partner communities around the world - including in Malawi - on sustainable solutions in health, education, justice, and enterprise.

Visit links.charity

Why Sam matters

Sam at Links

Partner contact and route into Seeds of Promise

Matthew's relationship with Sam at Links is how this work started. Sam connected the October 2025 visit, shared training materials with community leaders, and confirmed both the local appetite and the need for deeper local adaptation. Sam also sees potential to extend the model to other Links partners across Malawi, Uganda, Kenya, and Zambia if the pilot works.

Seeds of Promise is a Links-connected community, not a cold outreach project
Local leadership and charity partner oversight are already in place
The pilot can be documented as a repeatable model for other Links partners
Funders are backing a specific community with a known route in, not a vague geography

The ask

Fund the three layers: access, context, agents.

This is not "fund AI in Africa." It is a concrete, relationship-rooted pilot at Seeds of Promise:

Help connect and equip a specific community in Malawi, then build the context-aware agent toolkit that supports education, farming, fundraising, local enterprise, and community leadership.

Starlink or local connectivity

Solar or power resilience

Laptops, tablets, routers, and supporting equipment

Local context capture and champion training

First agent toolkit design, testing, and handover

Documentation, filming, and a repeatable playbook

Seeds of Promise x AstraJax is a living proof point for the bigger belief: AI becomes useful when it is grounded in real context, narrow enough to trust, and placed in the hands of people who already have the vision.

Fund the pilot Back to AstraJax

Seeds of Promise team with laptop and camera equipment

Community members gathered under trees for a coaching session

Same method, different foundation

Three layers in parallel.

The foundation changes. The principle does not. AstraJax and Seeds of Promise both move from access to context to bounded agents.

AstraJax

Commercial operating systems

Layer 1
The boring layer
Clean data, clear workflows, trusted numbers, and role-scoped interfaces that turn messy operations into a system people can rely on.
Layer 2
Context governance
Clive keeps the operational context current: what the business does, how decisions are made, who approves what, and what agents are allowed to touch.
Layer 3
Bounded agents
Narrow agents propose, humans approve, and work executes with an audit trail. Useful AI on top of a foundation the team already trusts.

Seeds of Promise

Community pilot, Malawi

Layer 1
Infrastructure
Connectivity, power, devices, and a usable computer centre with clear local ownership - the access layer that makes everything else possible.
Layer 2
Local context
Language, constraints, resources, community goals, and trained local champions - the reality a generic tool will never arrive with.
Layer 3
Agentic support
Narrow, context-aware agents for learning, farming, fundraising, enterprise, and leadership - specialist support packaged for the people who live there.

Structure the work. Structure the context. Then let agents use both.