← All lessons / AI in practice / Lesson 10 of 12
Interactive lesson
AI in your processes: improve the work you already do.
AI adds most value inside processes you already run, not as a separate experiment. Learn to map a process, choose where AI helps, and keep people at the points that matter.
- About 10 minutes
- Beginner friendly
- No account required
- Original diagrams
- Map a workflow and the people responsible for its steps.
- Decide what to keep human, assist or automate.
Map first, then decide step by step
Write the process down as it really happens today, step by step. Then, for each step, ask two questions: how much judgement does it need, and what happens if it goes wrong?
Keep human
Judgement about people, safeguarding, or anything hard to undo.
AI drafts, person checks
Reading, summarising, sorting and drafting, where a quick check catches errors.
Automate
Routine, low-risk, identical every time, and easy to spot and fix if wrong.
Interactive
Redesign a volunteer onboarding process
A fictional charity handles new volunteer enquiries in six steps. Decide how each step should be handled, and watch the redesigned process build below.
Your redesigned process
Key ideas to take away
- Map the real process first. Write down what actually happens today, including the workarounds.
- Match the role to the risk. Automate the routine, assist the judgement-heavy, keep decisions about people human.
- Design the checkpoint, not just the AI step. Say who checks, what they check, and what they do when something's wrong.
- Measure a baseline. Time and errors before AI, then the same measures after.
- Tell people what changed. Staff, volunteers and the people you serve should know where AI is used.
Check your understanding
3 quick questions. Try as often as you like. Your progress stays on this device; answers clear when you reload.
Want to go deeper?Methods, checklists and standards. Optional.
A simple risk test for each step
| Question | If yes |
|---|---|
| Does it decide something about a person? | Keep a person in charge of the decision. |
| Would an error be hard to spot or undo? | Add a check before the output is used. |
| Does it use personal or sensitive data? | Approved tools only; consider an impact assessment. |
| Is it identical every time and low impact? | A good candidate for automation. |
Watch for automation bias
People tend to accept a machine's suggestion, especially when busy. Make checks meaningful: give reviewers time, show them the source, and occasionally include a known error to confirm they're still catching them.
Workflows before agents
A fixed sequence of steps you design is easier to test than an agent that chooses its own route. Lesson 06 covers when an agent is worth it.
Continue exploring
We've picked resources that explain these ideas from different angles. They are independent of Code With Impact.
- Guidance
AI Playbook for the UK Government (opens in a new tab)
Government Digital Service
Ten principles for using AI safely and effectively, written for public bodies but useful for any organisation.
- Article
Charities and artificial intelligence (opens in a new tab)
Charity Commission
What trustees should consider, including an internal AI policy and keeping decisions with people.
- Framework
AI Risk Management Framework (opens in a new tab)
NIST
A widely used voluntary framework for governing, mapping, measuring and managing AI risk.
Sources for this lesson
The volunteer process, steps and recommendations are fictional and original to Code With Impact. The lesson draws on the frameworks listed under Continue exploring; no text or diagrams from them are reproduced.
How we reference: we link and credit, we never copy other educators' text, diagrams or code. See our full Sources & credits register.