Artificial Intelligence for charities:
start small, stay safe, prove it works.
Most charities already use Artificial Intelligence. Far fewer have a plan, a policy or a way to keep supporter data safe. This page answers the questions we are asked most and shows how to make progress without a big commitment.
Who can help a charity implement Artificial Intelligence?
Cirrico helps charities and nonprofits in the United Kingdom adopt Artificial Intelligence safely and usefully. We are a Salesforce Summit Partner working only with the charity sector, we hold ISO 27001 and the Data Security and Protection Toolkit, and we use Artificial Intelligence across our own business under a published policy. We help you decide whether Agentforce, Claude, Microsoft Copilot or none of them yet is the right first step, we put the data protections in place before anything connects, and we run a short pilot on your own data so you can see the result before you commit. It starts with Explore, a two-hour session we run at no cost to you.
Most charities have already started.
79%
of charities use Artificial Intelligence routinely
56%
say limited skills are the biggest barrier
28%
are using, testing or considering agents
45%
have taken no action on workforce planning
51%
have no Artificial Intelligence policy
6%
have started preparing their data
Charity Digital Skills Report 2026; Corndel, Zoe Amar Digital and Microsoft white paper on the charity workforce, 2026.
Adoption has grown from the ground up, through staff drafting emails and summarising meetings, often on personal accounts. Planning has not kept pace: half of charities have no policy, and most have not worked out what Artificial Intelligence means for their people. The organisations you compete with for funding and attention are not waiting, and trustees and funders are starting to ask what your plan is. The risk now is less in trying Artificial Intelligence than in using it unmanaged, or in standing still while others get the hours back. Skills and trust are the two things holding charities back, and both are fixable. Our job is to turn curiosity into a safe, measured first step.
How we think about it
Start from the process and the cost, not the tool
We look first at where the manual effort and the money go. The tool is the last decision, not the first.
Supporter data comes first
Nothing connects to your data until the access is scoped, written down and tested. If a use case touches sensitive data, it waits until the controls are in place.
Small steps, measured
A pilot you can switch off, with hours saved and adoption tracked, beats a strategy document nobody reads.
The questions we are asked most
Is our supporter data safe if we use Artificial Intelligence?
Do we need Agentforce, or will the tools we already have do?
What should a charity start with?
Do we need an Artificial Intelligence policy before we start?
What does it cost?
What if the answer is not yet?
How we help, in three steps
Artificial Intelligence runs through the way we already work, so there is nothing new to buy into.
Step one
Explore (two hours, we cover the cost)
A structured conversation about your goals, where the effort goes, your data and your red lines. You leave with a written note and a next step, including “not yet” if that is the right answer.
Step two
Discovery and Planning, with Artificial Intelligence readiness
Data foundations, use cases scored for value and risk, a version one policy, the tool recommendation and a roadmap. From £4,000.
Step three
Pilot, then adoption
A focused Programme block puts two or three use cases live on your own data with access scoped and tested, then your Managed Service keeps it governed and measured.
What it costs to start
Starting is cheaper, and less binding, than most charities expect.
You may already have it
Agentforce Coworker is included in many Salesforce editions and available in Salesforce Foundations, with usage drawing on the credits in your edition. If you are on Salesforce, switching it on for a pilot usually needs no new licence. We check your edition first.
Pilots without a long contract
A pilot is a short, fixed-price block of work, usually four to six weeks, with licences at the charity rate where a licence is needed at all. There is no multi-year commitment. If it does not earn its keep, you switch it off and keep what you learned.
Explore, at no cost to you
Two hours with our founder and a written note, with the cost covered by us. We run a limited number each month, so we ask you to apply.
Which Artificial Intelligence tools do we work with?
We are not tied to one tool, and the right answer is sometimes none of them yet.
Agentforce
Salesforce’s Artificial Intelligence. Agentforce Coworker answers questions and hands tasks to agents inside Salesforce; agents such as the Prospect Research Agent and Volunteer Capacity Agent take on defined jobs; Agentforce Nonprofit adds a library of ready-made use cases.
Claude
The assistant from Anthropic that reaches your Salesforce, your documents, your inbox and your chat tools, with charity pricing. We use it ourselves across Cirrico.
Microsoft Copilot
For charities that live in Microsoft 365. It works inside your existing permissions and now lets you choose Claude and other models. We help you decide where it fits alongside Salesforce.
Not yet
Sometimes the data is not ready, the policy does not exist or the use case touches people at risk. Then the right advice is to fix the foundations first, and we will say so.
We use it ourselves, every day
The best evidence we can offer is our own set-up, which is the one we help customers put in place.
Connected to our own systems
Claude is our approved tool, connected to our own Salesforce, Google Workspace, Slack, Zoom and Confluence under a written approval rule, so it only sees what the signed-in person already can.
Governed by a published policy
Our policy says which tools we use, which data never goes in, who reviews what goes out and who stays accountable. Our founder is our Data Protection Officer and Information Security Officer, and the policy is public.
Used in the work we do for charities
Drafting and reviewing documents, summarising meetings, checking Salesforce configuration and preparing reports, with a named person accountable for everything that reaches a customer.
Several of our customers already use Claude in their technical teams, and Artificial Intelligence assessments are under way with others. Their stories will appear here as they go live. Read our success stories →
Our commitments
We hold ISO 27001 and the Data Security and Protection Toolkit, our founder is our Data Protection Officer and Information Security Officer, and we use Artificial Intelligence ourselves under a published policy that says which tools we use, which data never goes in, and who stays accountable. We adopt Artificial Intelligence on clean data, with governance and the security the sector demands, and we advise against uses that carry risk to the people you serve, such as beneficiary-facing crisis agents.