Fields / City
First stop
There is a whole economy that sits between the state and the market — organisations that exist not to make a profit and not because a government department was told to, but because a group of people decided that some need was going unmet and set out to meet it.
This is the third sector: charities, community groups, campaigning organisations, foundations, social enterprises, and the vast web of voluntary activity around them. In the UK it is a serious part of national life. The voluntary sector's total annual income was estimated at around £69 billion in the most recent NCVO Civil Society Almanac figures [from salary and job-posting data, NCVO Almanac 2024]. The structural pull at the field level is Spread: the sector exists to take a cause, a need, or a source of goodwill that is contained — a local problem, an unheard group, a pool of money and willingness that isn't yet doing anything — and spread the response until help reaches the people who need it and the idea reaches the people who can act on it. Almost every charity job, underneath the specifics, is a job about widening the reach of something good.
Justice runs directly beneath the surface. A large part of the sector does not just deliver services; it exists to change the conditions that made the services necessary. Campaigning charities, human-rights organisations, environmental groups, and policy bodies spend their days trying to move law, public opinion, and institutional behaviour from unjust toward just. Even service-delivery charities usually carry an advocacy function — the food bank that also lobbies about the causes of hunger, the disability charity that also challenges discriminatory policy. The tension between "relieve the symptom now" and "fix the system that produces it" is one of the sector's oldest and most defining arguments, and it plays out inside almost every organisation.
Care and Connection are the daily texture for the front-line half of the sector. Beneath the fundraising totals and the campaign wins are people running refuges, staffing helplines, coordinating volunteers, supporting families, and building community. A great deal of charity work is relational — it is done in the presence of the people it serves, and it depends on trust, patience, and the ability to hold difficult situations without flinching. The sector attracts people who want their working life to be legibly connected to someone else's wellbeing, and it can deliver that connection more directly than most commercial work. It can also burn people out, because the need is effectively infinite and the resources never are.
What nobody tells you is that the sector runs on a permanent tension between mission and money. The work is values-driven, but the organisations still have to pay rent, salaries, and pension contributions, and most of the sector's income is precarious — grants that end, donations that fluctuate, government contracts that get cut. The distribution of resources is extreme: charities with income above £10 million are under 1% of organisations but take roughly 57% of all sector income, while around 80% of charities have income below £100,000 and collectively account for a tiny share [from salary and job-posting data, NCVO Almanac 2024]. Pay is real but modest — the sector has historically paid several percent less per hour than the wider economy, and charity CEO pay (median around £57,000–£64,000 depending on the survey and gender split) sits far below the commercial equivalent, with a CEO-to-average-worker ratio of roughly 4:1 against the FTSE 100's 113:1 [from salary and job-posting data, ACEVO Pay and Equalities Survey 2025; Harris Hill 2025]. People who thrive here have made a conscious trade: meaning and autonomy in exchange for financial ceiling and structural insecurity.
Entry is unusually open, which is both the sector's gift and its trap. There is no mandatory licence or single qualifying exam to work in a charity. Umbrella and professional bodies exist — the NCVO (National Council for Voluntary Organisations) is the sector's membership body, ACEVO represents chief executives, and the Chartered Institute of Fundraising (CIOF) offers qualifications and a code of practice for fundraisers — but none of them is a statutory gate in the way the SRA or GMC is for law or medicine. That openness means people arrive from every possible background, often through volunteering, internships, or a first paid coordinator role. It also means the sector can be exploitative at entry level: unpaid internships and low starting salaries act as an informal filter that favours people who can afford to work for little, which is one of the sector's live equity problems. In Portugal the same social space is organised around the terceiro setor and formalised through Instituições Particulares de Solidariedade Social (IPSS), non-profit bodies registered with Segurança Social and covered by their own collective labour agreements, sitting within a constitutionally recognised social and cooperative sector alongside the public and private ones [Segurança Social / gov.pt 2025].
The pull
Contained → Widespread
The pull to get what works to more people
Justice runs directly beneath the surface.
The line forks
Service update
What is newly possible in Charity & Third Sector, what stays yours, and how AI is changing the way in. In that order, and only if you want it.
2 barriers have dropped
Producing professional appeals, grant applications, campaign content and impact reports required staff (comms officer, fundraiser, designer) that 80% of the sector (charities under £100k) could not afford. A tiny volunteer-run charity with a genuine cause was structurally out-competed for funding and attention by larger organisations with professional teams. The capacity gap gated which causes could be heard.
Language models let a one-person or volunteer-run charity draft a competent grant application, appeal or campaign message; generative design tools (NCVO's Microsoft Designer recommendation) let them produce volunteer ads and announcements — capacity that used to require hiring.
Tiny grassroots organisations making their case professionally without a communications budget.
A founder launching a cause producing funder-ready materials solo.
Very likely to keep going this way. Small charities really can now produce professional materials. Whether that brings more funding or just more noise for funders is less certain, and funders' responses so far suggest the noise is a real possibility.
Every charity gets the same help and large charities use it better, so the gap in what they can produce narrows while their relative advantage may not move. As AI-drafted applications flood in, funders tighten their filters and 23% refuse AI content. AI improves what a small charity can produce and also makes it harder to get noticed.
Mode 1: traditional grant-making 'asks under-resourced organisations to spend enormous unpaid effort writing applications, most of which fail.' The effort cost of applying was a regressive tax on exactly the small organisations least able to afford it.
AI drafts applications, repurposes impact reports into letters of inquiry, and generates funder-specific case statements — cutting the unpaid effort dramatically.
Small-org fundraisers applying to more funders with less effort per application.
Could go either way. How much it helps is also less certain, because funders are responding fast and could cancel out the benefit within a couple of grant rounds.
Easier applications flood funders, who screen with AI, and 23% refuse AI-written submissions, so time saved on writing goes into disclosure, personalising and standing out. An application that looks AI-written also suggests less care to funders who care about values. The benefit is real but is being competed away as it happens.
These are real jobs that exist now and did not exist before the current wave of AI. Tap one to see what it is and how well we know it.
Owns AI governance, tool selection and workflow-building — the role the '47% with no governance policy' figure implies is missing. Real in large charities, absent in the 80% under £100k that cannot afford it.
Implied by the governance-gap data (47% no policy) and NCVO's guidance; no postings data located. · early signal
Donor analytics and prospect research growing as a distinct function in large charities, powered by predictive ML. Pre-existing but expanding.
Predictive-ML donor-analytics adoption in large charities (reported by people in the field). · early signal
Largely a consultancy/sector-body role (NCVO's programme, philanthropy advisers) rather than in-house — the sector is buying this capability externally because it cannot staff it.
NCVO small-charity AI programme; philanthropy-sector AI commentary (happening now on the programme; inference on it constituting a role). · early signal
Thin, and the thinness is informative — a gfx-shaped absence by a different route. No named exemplars, no postings data, no statutory role creation. Unlike finance (mandated CAIOs) or HR (a statutory bias auditor), the charity sector has NO regulator requiring anyone to act on AI and NO budget to hire specialists, so new roles concentrate in the handful of large charities and the consultancy/sector-body layer, not across the field. A field with no regulation and no money gets few new specialist roles; most charity AI work is being absorbed into existing jobs (the fundraiser who now also prompts) rather than spun into new ones.
Bars above the line are the parts of this work that still need a person. Bars below it are what AI can already do. Tap any column to see the actual work behind it.
The honest read. nfp passes the quality flag comfortably — relational judgement strong, accountability and ambiguity moderate. What makes it structurally different from gfx and fsh is WHERE the protection sits: it coincides with the field's PURPOSE rather than sitting orthogonal to it. The volunteer manager, community organiser, front-line worker and major-donor cultivator entered for the human connection, and the human connection is exactly what AI cannot touch. The exposed tasks (grant-writing, reporting, appeals) are the administrative overhead — Mode 1's 'necessary but slightly grubby business end'. In this field, AI threatens the paperwork and protects the point.
Language and reasoning is core alone — the sector's daily texture is writing (grants, appeals, case statements, briefings), and grant writing is the highest-interest AI application (60% interest, 24.6% already using). Predictive ML significant but concentrated in large charities. The distinctive feature is a genuine dual use dynamic in grant-making: applicants generate applications with AI while funders screen against them and 23% refuse AI-written submissions — an arms race that re-spends the saved effort on gatekeeping. Meta-field rule keeps the sector's CAUSES out of the matrix; the WORK of running a charity is scored.
5 ways into this field, and AI is not doing the same thing to each of them.
First-job tasks — drafting appeals, writing grant applications, donor research — are the exposed ones (grant writing the highest-interest AI task, 24.6% already using). The most exposed entry point, cushioned by the sector's capacity shortage.
Research synthesis and briefing-drafting are exposed entry tasks; coalition and political work is senior and protected.
First-pass application processing exposed on the funder side and caught in the dual use arms race (23% of foundations refuse AI-written submissions).
Relational and often physical from day one; the least exposed entry point. Mode 1: the currency is 'meaning, recognition, and relationship'.
Care and presence; not exposed.
That is everything we currently know about AI in Charity & Third Sector. It shows where things are moving so you can choose which way in suits you.
The job market
People and jobs. A genuine workforce shortage of an unusual kind. Recruitment 'becoming increasingly difficult and vacancies on the rise', turnover ~19%, nearly 7 in 10 nonprofit employees considering leaving in 2025 (reported by people in the field; NCVO Road Ahead and sector surveys). NOT a talent-scarcity shortage like cyber or vet science — a RETENTION-AND-PAY-DRIVEN shortage: the sector loses people to burnout and low pay faster than it can replace them. Mode 1: pay 'several percent per hour below the wider economy', 'the need is effectively infinite and the resources never are', burnout endemic. The sector does not lack willing entrants (entry 'unusually open', people 'arrive from every possible background') — it lacks the money to pay them.
The main limit on getting in. funding budget unambiguously. The sector has willing entrants and open entry; it lacks the money to hire and retain them, which is why unpaid internships act as an informal filter (an equity problem Mode 1 flags) and why the constraint on hiring is 'the grant that ended', not the candidate who did not exist. AI displacement is a minor secondary at the entry writing tier and is cushioned by the shortage.
Entry-level jobs. Entry demand is stable-to-constrained but constrained by FUNDING rather than by AI displacement — the sector wants more staff than it can pay for. AI absorbs some entry-level writing tasks (appeals, applications, reports) but the capacity shortage means freed time flows to unmet need rather than to reduced hiring. Vacancies are rising, not falling; the binding limit is budget, not candidate supply.
Cost of training. Moderate. No mandatory credential (entry is open, degrees common but not required, CIOF qualifications valued not required), so students are not forced into debt-funded specialist training. But pay is modest (Mode 1: CEO median ~£57-64k, far below commercial equivalents; entry roles low), and the unpaid-internship filter imposes an affordability barrier that functions like a hidden cost of entry. The financial ceiling is real and AI does not change it — Mode 1's trade of 'meaning and autonomy in exchange for financial ceiling' is unchanged.
The single most important LMC fact for this field's AI story. Because the sector is short of capacity, AI lands as a MULTIPLIER for overstretched teams rather than a SUBSTITUTE for expensive ones — the nursing/K-12 pattern, not the finance/graphic-design pattern. Freed time flows to unmet need, which is why entry-level impact is moderate-at-worst and never severe: there is no surplus of underused labour for AI to displace, only a deficit of capacity for it to help fill. The field's exposed tasks are its administrative overhead (grant-writing, reporting), which the sector is glad to automate; its protected core (the human relationship) is both what AI cannot do and what people enter the sector for. AI threatens the paperwork and protects the purpose, and slightly improves the sector's standing trade of meaning-for-money by removing some grind.
Interchange
People drawn to Charity & Third Sector are often drawn to these.
End of the line
That is a conversation about this field, not a test. Or keep going on your own. Both doors stay open.