The first AI management assistant for NGOs

Meet your NGO’s AI teammate.

Everyone talks about using AI. Impact Lighthouse makes it work for NGOs. It’s a tool built around the way your team already works — field data, follow-ups, funder reports — with an AI teammate that takes on the legwork at every step and checks with you before it changes anything. We set it up and run it for you.

Built for NGO workflowsWorks in Impact Lighthouse or the AI you already use
Faster. Better. More focused.
Your AI teammatein Impact Lighthouse or your own AI
Get the follow-ups ready for Monday’s caseworker meeting.
Found 12 students below 60% attendance this term
Grouped them by school and caseworker — 4 are at School A
Added each student’s last home visit and its outcome
4
3
3
2
School ASchool BSchool CSchool D
Shall I save this as the list Follow-up this term and share it with the Caseworkers group?
Save and shareChange something
The problem

Running an NGO shouldn’t mean fighting your tools.

Your data is scattered across paper forms, spreadsheets, WhatsApp threads and people’s memories. Reporting to funders swallows weeks. Field staff battle software that was never built for them. And while everyone says “you should use AI,” nobody shows you how.

You didn’t start an NGO to wrestle with admin — Impact Lighthouse takes it off your plate so you can get back to the work that matters.

Paper everywhere
Registers and forms that never make it into a system.
Scattered data
Spreadsheets across laptops, WhatsApp threads, memories.
Reporting season
Weeks of manual work to answer funder questions.
“Just use AI”
Everyone says it. Nobody shows you how.
Lighty, Impact Lighthouse’s AI assistant: a smiling lighthouse with a star for a lamp
Meet Lighty

Lighty is the AI teammate that just helps. No PhD required.

Most NGOs hear “AI” and picture something expensive, technical and out of reach. Impact Lighthouse flips that. Lighty, our AI assistant, isn’t a feature bolted onto the side — he’s how you work with your data, inside Impact Lighthouse or through the AI you already use. He reads only what you’re allowed to see. Anything he can’t undo, he can’t do alone: deleting and changing a field’s type wait for your approval in Impact Lighthouse, and automation rules he sets up start switched off. You never have to “adopt AI”: you ask, in plain words, and he does the heavy lifting with you.

Built in

Ask AI, inside Impact Lighthouse

Ask Lighty a question on any dashboard and get a clear answer and a chart back. He reads only what you’re allowed to see. Your admins set a monthly AI budget for each group, and there’s a daily limit for each person.

Fully connected

The AI you already use, connected to your data

Connect ChatGPT, Claude or Gemini through a secure connection (MCP), and it can do what you can do on screen: find, add and update records and, for admins, build dashboards, lists, calculations and automation rules. It reads only what you’re allowed to see. Deleting and changing a field’s type wait for your approval in Impact Lighthouse, and automation rules it sets up start switched off.

Spot what’s changing.
“Compare this term’s attendance to last term.”
Ask for a trend and get the comparison and a chart back.
Ask your data anything.
“How many girls attended in December?”
A clear answer and a chart, instantly.
Build it with you.
“Build a dashboard of attendance by school for the board.”
It proposes the charts, checks each one against your data, and builds it once you agree.
Your donor report, drafted for you.
“Draft our Q4 report for the education funder.”
It writes the narrative around your verified figures; you edit, redraft and approve.
Harmonise your data.
“Find families entered twice and fix the village spellings.”
It finds the duplicates and inconsistent entries, shows you each fix, and makes them once you agree.
Speaks your team’s language.
“បង្ហាញជាភាសាខ្មែរ” → shown in Khmer
Bridges languages for multilingual field teams.
Service as software

Not another tool to learn. A service that does the work.

Software hands you a licence and a manual, and leaves the work to you. Impact Lighthouse works the other way round. We set it up around your programmes and run it for you, and your AI teammate takes on the legwork — building the dashboards, harmonising the data, linking your framework, writing the report — while your team stays in charge of every decision.

You get the outcome, not another system to look after.

01

Faster

Nothing to install or configure, because we set it up. Ask for a dashboard or a report and the assistant drafts it in minutes.

02

Better

Every figure comes from the NGO’s own records, defined once and checked before it reaches a funder. The framework shows what’s really working.

03

More focused

No servers, no IT hire, no “reporting season”. The hours go back to the people the NGO serves.

Clean data

Data you can trust, from the first entry.

Forms keep mistakes out. Required fields, number ranges, rules that make one answer depend on another, dropdowns and linked records instead of free typing, and values the form fills in for you. Every entry is checked before it’s sent.

Your AI teammate keeps what’s already there in order. Bring years of spreadsheets as they are: it helps shape them into a proper database, finds the duplicates, the five spellings of one village and the values that don’t add up, shows you each fix, and makes it once you agree. Field staff work in their own language — write each form in whatever language your team speaks, and switch the app itself to Khmer.

No-codeYour team’s languageChecked before it’s sentRequired fields & rangesDependent rulesLinked records, not typosPre-filled recordsRegister entryPublic forms
Home Visit — form builderPublished
LNKFamily · គ្រួសារLinked record
DTEVisit dateDate
SELVisit outcomeDropdown
IMGPhoto evidencePhoto upload
NUMDays attendedNumber · 0–31
IFFollow-up needed?Conditional
Days attended must be at most 31 — fix it before sending
Your AI teammate
I found 3 spellings of one village in Family (“Kien Svay”, “Kean Svay”, “kien svay”) across 41 records, and 2 families entered twice. Harmonise them?
Show me the fixes
Automations

The right people know the moment it happens.

Tell your AI teammate what should happen — “when a home visit finds a child out of school, tell the caseworkers” — and it sets the rule up for you to check and switch on. From then on, Impact Lighthouse acts the moment data is saved: it emails the caseworker, alerts the programme lead, or updates the record itself, so nobody has to remember to pass it on. Every run is logged and kept for 90 days, and if a rule ever fails, the person who made it hears about it.

Runs the moment data is savedEmail alerts to groups or peopleUpdates records for youActivity logFailure alerts
Child not in school — alertEnabled
WhenA Home Visit form is submitted
Only ifVisit outcome is “Child not in school”
DoEmail the Caseworkers group, and set “Follow-up needed” to Yes
Ran 2 minutes ago · emailed 3 caseworkers · 1 value set
Business intelligence

Dashboards your AI teammate builds for you.

Tell your AI teammate what the board needs to see — “attendance by school, against our 90% target” — and it proposes the charts, checks each one against your data, and builds the dashboard once you agree. Trends, pivot tables, heatmaps and progress-to-target gauges, always current as data arrives. Publish to the teams that need it, each seeing only its own slice, and download a branded PDF for the board or a funder.

Business intelligence without the analyst.

Live dashboardsPivot tablesProgress to targetHeatmapsReusable calculationsBranded PDF exportPer-group data scope
Attendance — board dashboardDownload PDF
Attendance vs target
87%
Target 90%
Pivot · attendance by school and term
SchoolT1T2T3
North84%86%89%
River79%83%88%
Market76%80%84%
Total80%83%87%
Heatmap · attendance by weekday, last 10 weeks
Mon
Tue
Wed
Thu
Fri
01
Built for you, in minutes
Describe what you need; your AI teammate proposes the charts and builds them once you agree.
02
Figures worked out with you
Your AI teammate works out a figure like “attendance rate”, shows you the number, and saves it for every chart, list and report.
03
Lists with totals
Group records, see subtotals, then filter, search and download.
04
Share it safely
Publish to the groups that need it, and each school sees only its own figures.
05
Ready for the board
Download any dashboard as a PDF with your logo, filters and date on it.
Q3 report — Education funderDraft
Students reached
1,284
Attendance
87%
Home visits
342
Narrative · written by your AI teammate

This quarter, 1,284 students attended classes across 4 schools. Average attendance rose to 87%, up from 81% in Q2, and caseworkers completed 342 home visits to families of students at risk of dropping out.

Every number traced to your data · ready to approve
Reporting

Funder reports your AI teammate writes for you.

Set up a template once for each funder: the reporting period, the key figures, the charts and tables, and what each section should cover. Press Generate, and your AI teammate pulls the period’s figures and writes the report around them, in the length and tone you chose. You edit, ask for a redraft, and approve. A number that doesn’t come from your data can’t be approved — so what reaches the funder is what your records say.

Templates per funderWritten by your AI teammateFigures from your dataReview & approvePrint or save as PDF
Fundraising

Better funder communication. Less work to get there.

Funders want clear numbers, on time, that agree with what you told them last quarter. Impact Lighthouse gives you that without a reporting season. Every report, dashboard and answer comes from the same live records, and your AI teammate does the drafting — so your team spends its time on the relationship, not the spreadsheet.

Drafted for you

One template per funder. Press Generate and the report is drafted from the period’s figures, in the structure and tone each funder expects. You edit and approve.

Same-day answers

When a funder asks a question between reports, the answer is already in your data. Ask your AI teammate and reply the same day.

Numbers that agree

Every number traces back to your records, so your figures agree from one funder to the next, and a branded dashboard shows progress against the targets you promised.

M&E framework manager

Your M&E framework, live against your data.

Most Theories of Change and logframes live in a Word document, far from the data that proves them, and the actuals get filled in by hand before every donor report. Impact Lighthouse links the two. Every indicator points at a real figure and carries its baseline and target, and each reporting period shows at a glance what’s on track, at risk or off track. Switch to the logframe view when a donor asks for it.

Already have a framework? Upload it. Your AI teammate reads it in, links each indicator to your data, and flags the ones nothing measures yet.

Starting from scratch? Build one with your AI teammate — from a conversation, a sector template, or the data you already collect — then let it set up what the missing indicators need.

Every period, checked. When a period closes, your AI teammate flags assumptions your figures no longer support.

Education programmeTerm 1, 2027
9 on track3 at risk2 off track
Goal
Every child in the programme completes primary school
Outcome
Children attend school regularly
Indicator
Average attendance rate
87% / target 90%At risk
Indicator
Students attending 80%+ of days
1,061 / target 1,000On track
Outcome
Families stay engaged with their child’s schooling
Indicator
Home visits completed
342 / target 300On track
Indicator
Students re-enrolled after dropping out
14 / target 30Off track
Where it breaks down

Collecting the data was never the hard part.

Most teams solve collection eventually — a form app, a tablet, a spreadsheet that mostly works. Then the data lands, and the second half of the job starts: exporting, cleaning, and turning it all into something a funder will read. Impact Lighthouse is built for that half as much as the first.

How the second half of the job compares
StageHow it usually goesWith Impact Lighthouse
After collection
How it usually goes: The data lands as an export, and lives in whichever spreadsheet was opened last.
With Impact Lighthouse: One organised place, and an AI teammate that finds duplicates and inconsistencies and fixes them with you.
Making sense of it
How it usually goes: Someone has to build the analysis by hand, and it’s the same person every time.
With Impact Lighthouse: Your AI teammate builds the dashboards, and they update as data arrives.
Proving impact
How it usually goes: The logframe lives in a Word file. Someone fills in the actuals by hand before each donor report.
With Impact Lighthouse: Your framework is linked to your data. Every indicator shows its latest figure against its target.
Funder questions
How it usually goes: A funder asks for a figure between reports, and someone spends a day rebuilding it.
With Impact Lighthouse: The answer is already in your data. Ask your AI teammate and reply the same day.
Reporting season
How it usually goes: Weeks of pulling numbers together, separately, for each funder.
With Impact Lighthouse: Your AI teammate writes the report from the period’s figures, ready for you to edit and approve.
Getting started
How it usually goes: Setup falls to whoever on the team is most technical.
With Impact Lighthouse: We set it up with you and run it. No servers, no IT team, nothing to install.

Ready to hand the legwork to your AI teammate?

How it works

Collect. Understand. Improve.

From the first form to the final donor report — with your AI teammate at every step.

1

Start from what you have.

Bring your spreadsheets and paper forms. We set up the forms your programme needs with you — no developers, no consultants.

2

Collect anywhere, on any device.

Phones or tablets, classroom or field. Every entry lands the moment it’s submitted.

3

Understand what’s happening.

Your AI teammate builds the dashboards, and they update themselves as data arrives.

4

Improve your impact — and prove it.

Your framework shows which results are on track and which need attention, while the programme is still running. When a funder asks, your AI teammate writes the report from the same figures. Reporting becomes a by-product, not a project.

See the whole journey in a demo
Getting started

Your data is a mess? That’s where we start.

Most NGOs come to us with years of spreadsheets, paper registers and forms that don’t match. That’s normal, and it’s no problem. Your AI teammate helps us structure it into a proper database, harmonise and clean it, and get your team going. There’s nothing to install and no IT team needed.

Start with a 30-minute call
1

Show us how you work today

A 30-minute call about your programmes, your forms and your spreadsheets — messy parts included.

2

We structure and clean it, with your AI teammate

Your AI teammate proposes the tables and links your spreadsheets belong in, harmonises spellings and formats, and finds duplicates and gaps. We check every step with you, then build your first forms and dashboards.

3

Your team just starts

Staff open a link on the phone or laptop they already have, in their own language. When they’re unsure how to do something, they ask the AI teammate, and our support is there too.

Case study photo — ISF Cambodia programme team
ISF programme team, Phnom Penh — photo to follow.
Case study

Indochina Starfish Foundation, Cambodia

ISF’s teams replaced paper attendance registers and scattered spreadsheets with live dashboards and reports drafted from the same figures — freeing time that now goes to the children they serve.

[X]
students tracked
[Y]
programmes live
Read the full case study
Who it’s for

Whatever you need to track, you can build it.

Impact Lighthouse isn’t built for one sector — it’s a flexible system you shape to your own programmes.

AttendanceHome visitsDistributionsSurveys & assessmentsCase notesScreeningsBeneficiary registersMonitoring & evaluation…or anything else your programme runs on.

Your data stays yours.

Your data is always yours — never locked away or sold on.

When the AI is involved

What the AI sees, what you can keep from it and what waits for your approval. How we handle your data

Grant-friendly pricing.

Priced for NGO budgets, not enterprise ones — no per-seat surprises, no IT department required. We set it up, we run it, and the fee scales with the size of your organisation.

One flat fee — no per-seat pricing that punishes growth
Setup, hosting and support included
Tell us your staff count and we’ll quote in a day
Tell us your size and we’ll quote it
Why we exist

We believe the organisations changing the world shouldn’t be held back by the tools they use to do it.

Impact Lighthouse started on the frontline, not in a boardroom. We watched dedicated teams pour hours into paper registers and tangled spreadsheets — time that should have gone to the people they serve — while the data they needed to prove and improve their work sat locked away. So we set out to free it.

A lighthouse doesn’t move the ships — it gives them the clarity to navigate safely. That’s our role: not to run your mission, but to give you the clear view you need to steer it well.

What we believe

Mission over admin. Every hour spent fighting tools is an hour taken from the work that matters.

Data should serve the people doing the work — not the other way around.

Powerful technology should reach everyone — not just those who can afford consultants and IT teams.

Your data is yours. Always. Protected, and never sold on.

Team photo
Team photo to follow.

Who we are

[Placeholder: a short paragraph on the team. The Indochina Starfish Foundation in Cambodia was our first deployment and remains a close partner. Add names and photos.]

AI & data security

Safe to use AI with beneficiary data.

Before you let AI near the people you serve, ask what it sees, what you control, what it can do alone, what it keeps and where your data is hosted. Here are our answers.

What does the AI see?

Lighty, and any AI you connect to Impact Lighthouse, reads only what the person asking is allowed to see. It can’t run its own database queries or open attached files.

What can we keep from it?

Anything you choose, with Keep from AI. An admin marks a field or attachment column, and that field, and anything worked out from it, such as lookups, calculations and linked names, never reaches the AI. The check runs on every request: if it can’t run, the AI refuses rather than answers. Admin → Kept from AI lists every kept field, with its change history.

What can’t it do without a person?

Deleting and changing a field’s type wait for your approval in Impact Lighthouse: that covers deleting records, fields, tables, lists, dashboards, calculations and automation rules. An approval expires after 24 hours. Automation rules the AI sets up start switched off, and it never approves a funder report. Your admins switch AI on and set each group’s monthly AI budget, and each person has a daily message limit.

How long are our questions kept?

The text of each question asked of the AI is deleted after 90 days, and the rest of its record after 24 months. A record you delete is erased for good 30 days later. AI requests pass through our central system, which keeps no question or answer.

Where is our workspace hosted?

Your workspace is hosted in the EU, and your organisation’s application and database run separately from every other customer’s. Every third party that handles your data for us, including for AI, is in our sub-processor list.

Do you train AI on our data?

No. We never use your data to train AI or to improve AI models.

Our data processing agreement and sub-processor list are available on request. Questions about data protection? privacy@impactlighthouse.io

See what your data could tell you.

Book a 30-minute demo and we’ll show you Impact Lighthouse with your kind of programme in mind.

Your programmes, and where data gets stuck today
A live walkthrough — collect, dashboard, framework, report
What setup and pricing would look like for your team

“[Director quote — one line on what changed for the team.]”

ISF
[Name, Role]
Indochina Starfish Foundation
Book a demo

No obligation, 30 minutes. We use your details only to reply — see how we handle data.