Automation & AI

Analysing processes

Most organisations have processes that have grown over time and are rarely written down. We start by talking to the people who carry them out, mapping each step, who is involved, which systems are used and where information gets copied, chased or lost.

Identifying component tasks

A process (e.g. “onboarding a new volunteer”) is really a series of smaller tasks: collecting details, checking eligibility, updating a database, sending a welcome pack, notifying a manager. Breaking processes down this way shows which parts need human judgement, which can be automated and which can be handled by responsible use of AI.

Prioritising

Not everything is worth automating. We weigh up how often a task happens, how long it takes, how error-prone it is and how sensitive the information involved is. That gives you a clear, realistic order of work rather than an unstructured wishlist.

Applying AI

Where a task involves reading, summarising, categorising or drafting, large language models such as Claude and Gemini can help. Getting good results depends on giving them the right context: the background documents, examples and guidelines they need to do the job well. We help you set this up using features such as projects (shared workspaces holding your reference material and instructions), memory (so tools retain relevant knowledge between sessions) and skills (reusable instructions for tasks your team does repeatedly, such as formatting reports in your house style).

Applying non-AI automation

Many tasks don’t need AI at all. Orchestration tools such as n8n and Zapier connect the systems you already use through their APIs, so data moves automatically between your website, CRM, email platform and spreadsheets. Rule-based workflows like these are reliable, low-cost and use a fraction of the energy of AI.

Keeping you in control

We design workflows so a person reviews anything AI produces before it’s used, and we’re careful about what data is shared with which tools. You can read more about how we think about these tradeoffs in our AI policy.