When the AI asks if you have any drafts or notes to share, you can upload documents or paste in information to help it understand your work more deeply. The quality of the ideas you get back depends on the quality of context you provide. Don't worry about having everything neatly organised - bullet points, rough notes, and partial data are all useful. Share what you've got.

Important context (add as much of this as you can)

• What the service currently involves:
  
  •  What activities you deliver, how often, and in what format
  • Who uses it and roughly how many people
  • What outcomes you're working towards

• What's working and what isn't
  
  • Feedback from beneficiaries or service users — even informal comments
  • Things your team regularly notices or raises
  • Any data on take-up, drop-off, or engagement

• Your goals and constraints
  
  • What you'd most like to change or improve
  • Any limits on time, budget, or staff capacity

Additional context that adds depth and detail

• Survey or evaluation data: satisfaction scores, outcome measurements, or themes from beneficiary feedback
• Examples of what isn't reaching people: who your service is missing or underserving
• Relevant external context: changes in funding, policy, or in the needs of your community
• What similar organisations do: if you have any knowledge of comparable services elsewhere
• Previous service reviews or improvement plans: to avoid retreading old ground
• Tone of voice guidelines if you have them

Before you upload files: check for sensitive data

Review your documents for personal or sensitive information about beneficiaries. Anonymise names and identifying details before uploading. Even though Worthwhile Chat is privacy-first, good data practice starts with you. If you're using other AI tools, check their data settings too — many train on your data by default. How much you share depends on your existing privacy and data policies, and how much control you have over the tools you use.

If you’re unsure, ask yourself: “If this document were seen by someone outside our organisation, could anyone be identified or harmed?” If yes, anonymise first. 

The AI will produce a set of improvement ideas, organised by timeframe or theme. You may be asked follow-up questions - for example, about what a particular challenge means in practice, or which outcomes matter most to you. This is normal. Answering them will help improve the output.

Read through what it has produced and ask yourself:

• Do these ideas reflect the real picture of our service — or are some based on a misunderstanding of what we do?
• Are the suggestions genuinely practical for an organisation of our size and capacity?
• Have equity and inclusion considerations been taken into account? Would these changes work for everyone we support?
• Are there ideas here we've already tried or ruled out? If so, why?
• Is anything missing, such as a gap or challenge we mentioned that hasn't been addressed?
• Do any of the suggestions feel too ambitious, or not ambitious enough?

If something doesn't feel right, that's useful information. Note what's off and carry it into the next step.

Getting a useful output often takes more than one go. Think of this as a working conversation; you can keep refining until the ideas feel genuinely relevant to your situation.

Ask for revisions

If the first output isn't quite right, try instructions like:

Tone:

• "Make this sound less like generic advice and more like something our own team would actually say."
• "This feels quite formal, can you make the language more practical and down to earth?"

Length:

• "Remove suggestions that require significant new funding, as we need ideas we can try with existing resources."
• "Take the top three ideas and expand on them: what would each one look like in practice?"

Structure:

• "Focus specifically on the [name of aspect], as that's the area we most need to develop."
• "For each suggestion, can you add a note on what we'd need to do first to make it happen?"

Representation:

• "Can you focus on what would make the biggest difference to people who are hardest to reach?"
• "Are there any ideas here that could unintentionally make things harder for some of the people we support?"

If the conversation goes off track, you can always start a fresh one with slightly different framing. Sometimes a new prompt gets you closer to what you need.

Add what only you know

AI works with what you give it. Once you have a solid draft, fill in the gaps that only your team can provide:

• The history behind the service: what's been tried before and why it did or didn't work
• The specific relationships and partnerships that shape what's possible
• The informal knowledge your frontline staff carry about what beneficiaries actually experience
• Any upcoming changes to funding, staffing, or the needs of your community — that should shape priorities

Involve the people closest to the work

AI can help you generate and organise ideas, but the people closest to the work will spot what's missing, what's unrealistic, and what matters most. Share ideas with the frontline staff who deliver the service, and with beneficiaries or service users where you can, before making final decisions. Their day-to-day knowledge and experience is something no AI can replicate.

Get a colleague to read it through

Before sharing improvement ideas with senior leadership or using them to inform decisions, ask a colleague to read them through. Do they recognise the service in what's been described? Is anything missing, unrealistic, or overstated? Never share AI-assisted content externally without a human review.

Be sceptical of suggestions that sound good in theory

AI generates ideas based on patterns across many organisations. Not every suggestion will fit your context, culture, or community. Treat the output as a prompt for your own thinking — not a blueprint. Ask yourself: "Would this actually work for us and for the people we support?"

Always check facts and figures

If the AI refers to sector benchmarks, research findings, or external examples, verify them before using them to inform decisions. The AI is there to support your thinking - not to replace your own knowledge and judgement.

Be transparent about AI use

If the ideas generated by this process feed into a board paper, a service review, or a funding application, consider noting that AI tools were used as part of the development process. Being open about this is in line with Worthwhile AI values.