01
Start with people
What this means
Design decisions should flow from a deep understanding of real needs and experiences. This isn’t a one-time research phase at the start – it’s ongoing. From initial exploration through design, build, launch, and evaluation, you’re listening to and learning from the people who will use your tool, rely on it, fund it, or support others using it.
“People” means a broad ecosystem of actors:
- Communities (end-users) seeking legal information or support
- Lawyers and advocates who refer clients or use the tool
- Service providers and legal aid organisations
- Support workers, community organisations, and trusted intermediaries
- Funders and policy makers
- Back-office teams who maintain and improve the tool
Genuine co-design means involving these people in shaping solutions from the very start, not just testing them at the end. This takes real investment in time, in resources, and in creating the conditions for meaningful participation.
Many people come to access to justice tools in crisis, distress, or with low digital literacy. A trauma-informed approach recognises this. Design for diverse needs, not just the “average user.” Understand what success actually looks like for the people you’re trying to help – not just engagement metrics or completion rates.
Why it matter
When you start with people, you design tools that actually work. You catch assumptions early. You build trust. You’re more likely to create something that closes gaps rather than deepens them.
What this isn’t
This isn’t about designing for one user archetype. It’s not just listening at the beginning. It’s not assuming all users have the same needs, confidence, or access to technology.
02
Build trust through transparency and accountabilityStart with people
What this means
Trust is the foundation everything else is built on. Without it, the most well-designed tool will fail the people it’s meant to serve.
At its core, this principle is about a fair bargain: the tool is useful and accurate; it’s transparent about what it does, what it can’t do, and what happens to data; and it honestly acknowledges the risks involved. End-users, and the professionals who support them, should know what they’re getting into.
Accountability for that bargain is distributed across everyone involved:
Tool builders are accountable for how the tool functions, what data it collects and protects, and its broader system impacts
Service providers are accountable for implementation, support and ensuring the tool is used appropriately
Legal expertise should be embedded in design and governance and not consulted after the fact
Funders and policy makers are accountable for the incentives they create and the standards they require
Build accountability mechanisms into your tool from the start: ways of recording how decisions are made, evaluating real impact, and creating genuine feedback loops with users. Don’t add these as afterthoughts.
Why it matter
When trust is the north star, everything else aligns. And when the sector holds itself to this standard collectively, it builds shared credibility.
What this isn’t
This isn’t about perfect tools. It’s about honest, useful tools built on fair terms and being transparent when things aren’t working.
03
Design for equity, fairness, and inclusion
What this means
Design to close gaps, not widen them and actively test whether your tool is doing so. Understand who your tool serves and, just as importantly, who it excludes. Design for all use cases and circumstances, not just the average user or the easiest scenario.
Efficiency and equity are not always in tension. More timely responses, better triage, and reduced administrative burden are equitable outcomes. The goal is not to avoid efficiency; it’s to ensure that efficiency never comes at the expense of dignity, clarity, and agency. Where trade-offs between equity, scale, and cost exist, name them honestly rather than designing around them.
Many people come to A2J tools in crisis or distress, or with low digital or legal literacy. A trauma-informed approach shapes not just what your tool does, but how it communicates: what it asks of people, in what order, and with what tone.
Bias, particularly in AI and automated decision making, requires active, ongoing attention. Identifying and mitigating bias is not a one-time check. The sector needs clear standards for how to test, evaluate, and communicate about bias in practice.
Consider the systemic context your tool sits within. A well-designed tool can reduce complexity and open pathways, but it can also inadvertently reinforce existing inequities or lock people into systems that weren’t working for them to begin with. Be honest about what your tool can and cannot change.
Why it matter
Equity isn’t about treating everyone the same, it’s about closing real gaps. When you design with this lens, you’re more likely to create tools that actually serve the people who need them most.
What this isn’t
This isn’t about designing a tool that works for everyone equally. And it’s not about solving systemic inequality with a tool alone, but it is about being honest about what your tool can and can’t do in that context.
04
Provide actionable guidance
What this means
People come to A2J tools needing more than information. They need to know what to do next, or to feel less alone in a difficult situation. Guidance isn’t always a concrete action. Sometimes what people need most is clarity, confidence, context, or reassurance that they’re on the right track.
The tension between being helpful and being cautious is real, but over-restriction is itself harmful. Withholding genuinely useful information in the name of legal caution leaves people stranded. You can support informed decision making without crossing into unauthorised legal advice and being transparent about where that line sits for your tool is part of good design, not a limitation of it.
At the same time, be honest about what your tool cannot do. Tools cannot substitute for systemic change or for human expertise when a situation genuinely requires it. Design with that honesty built in, not as a disclaimer bolted on at the end, but as a genuine part of how the tool guides people forward
Why it matter
When guidance is clear and actionable, people move forward. They feel less stuck, more confident. Vague warnings and over-caution leave people stranded.
What this isn’t
This isn’t about giving legal advice. It’s not about being reckless with disclaimers. It’s about finding the useful middle ground between “we can’t tell you anything” and “we’re your lawyer” and being transparent about where your tool sits on that spectrum.
05
Use technology purposefully
What this means
Start by asking: is technology the right answer here? And if so, is building something new better than strengthening what already exists?
In an emerging field, multiple approaches to the same problem create value. Don’t suppress experimentation in the name of avoiding duplication, but build on strong foundations rather than creating isolated one-off tools that can’t connect to the wider ecosystem.
If you decide to build, commit to constant testing, evaluation, and iteration. Define what “good enough” looks like. Work toward shared evaluation frameworks across the sector so that tools can be assessed against common, meaningful outcomes.
Where AI is used, it must create demonstrably better social impact than not using it. That case needs to be made honestly and evidenced over time, not assumed at the point of build.
Why it matter
When technology is used purposefully, it solves real problems without creating new ones. The sector learns and improves together.
What this isn’t
This isn’t about building perfect tools. It’s not about avoiding all experimentation. It’s about being honest about whether technology is the right answer, and committing to evaluate whether it actually is.
01
Start with people
What this means
Design decisions should flow from a deep understanding of real needs and experiences. This isn’t a one-time research phase at the start – it’s ongoing. From initial exploration through design, build, launch, and evaluation, you’re listening to and learning from the people who will use your tool, rely on it, fund it, or support others using it.
“People” means a broad ecosystem of actors:
- Communities (end-users) seeking legal information or support
- Lawyers and advocates who refer clients or use the tool
- Service providers and legal aid organisations
- Support workers, community organisations, and trusted intermediaries
- Funders and policy makers
- Back-office teams who maintain and improve the tool
Genuine co-design means involving these people in shaping solutions from the very start, not just testing them at the end. This takes real investment in time, in resources, and in creating the conditions for meaningful participation.
Many people come to access to justice tools in crisis, distress, or with low digital literacy. A trauma-informed approach recognises this. Design for diverse needs, not just the “average user.” Understand what success actually looks like for the people you’re trying to help – not just engagement metrics or completion rates.
Why it matter
When you start with people, you design tools that actually work. You catch assumptions early. You build trust. You’re more likely to create something that closes gaps rather than deepens them.
What this isn’t
This isn’t about designing for one user archetype. It’s not just listening at the beginning. It’s not assuming all users have the same needs, confidence, or access to technology.
02
Build trust through transparency and accountabilityStart with people
What this means
Trust is the foundation everything else is built on. Without it, the most well-designed tool will fail the people it’s meant to serve.
At its core, this principle is about a fair bargain: the tool is useful and accurate; it’s transparent about what it does, what it can’t do, and what happens to data; and it honestly acknowledges the risks involved. End-users, and the professionals who support them, should know what they’re getting into.
Accountability for that bargain is distributed across everyone involved:
Tool builders are accountable for how the tool functions, what data it collects and protects, and its broader system impacts
Service providers are accountable for implementation, support and ensuring the tool is used appropriately
Legal expertise should be embedded in design and governance and not consulted after the fact
Funders and policy makers are accountable for the incentives they create and the standards they require
Build accountability mechanisms into your tool from the start: ways of recording how decisions are made, evaluating real impact, and creating genuine feedback loops with users. Don’t add these as afterthoughts.
Why it matter
When trust is the north star, everything else aligns. And when the sector holds itself to this standard collectively, it builds shared credibility.
What this isn’t
This isn’t about perfect tools. It’s about honest, useful tools built on fair terms and being transparent when things aren’t working.
03
Design for equity, fairness, and inclusion
What this means
Design to close gaps, not widen them and actively test whether your tool is doing so. Understand who your tool serves and, just as importantly, who it excludes. Design for all use cases and circumstances, not just the average user or the easiest scenario.
Efficiency and equity are not always in tension. More timely responses, better triage, and reduced administrative burden are equitable outcomes. The goal is not to avoid efficiency; it’s to ensure that efficiency never comes at the expense of dignity, clarity, and agency. Where trade-offs between equity, scale, and cost exist, name them honestly rather than designing around them.
Many people come to A2J tools in crisis or distress, or with low digital or legal literacy. A trauma-informed approach shapes not just what your tool does, but how it communicates: what it asks of people, in what order, and with what tone.
Bias, particularly in AI and automated decision making, requires active, ongoing attention. Identifying and mitigating bias is not a one-time check. The sector needs clear standards for how to test, evaluate, and communicate about bias in practice.
Consider the systemic context your tool sits within. A well-designed tool can reduce complexity and open pathways, but it can also inadvertently reinforce existing inequities or lock people into systems that weren’t working for them to begin with. Be honest about what your tool can and cannot change.
Why it matter
Equity isn’t about treating everyone the same, it’s about closing real gaps. When you design with this lens, you’re more likely to create tools that actually serve the people who need them most.
What this isn’t
This isn’t about designing a tool that works for everyone equally. And it’s not about solving systemic inequality with a tool alone, but it is about being honest about what your tool can and can’t do in that context.
04
Provide actionable guidance
What this means
People come to A2J tools needing more than information. They need to know what to do next, or to feel less alone in a difficult situation. Guidance isn’t always a concrete action. Sometimes what people need most is clarity, confidence, context, or reassurance that they’re on the right track.
The tension between being helpful and being cautious is real, but over-restriction is itself harmful. Withholding genuinely useful information in the name of legal caution leaves people stranded. You can support informed decision making without crossing into unauthorised legal advice and being transparent about where that line sits for your tool is part of good design, not a limitation of it.
At the same time, be honest about what your tool cannot do. Tools cannot substitute for systemic change or for human expertise when a situation genuinely requires it. Design with that honesty built in, not as a disclaimer bolted on at the end, but as a genuine part of how the tool guides people forward
Why it matter
When guidance is clear and actionable, people move forward. They feel less stuck, more confident. Vague warnings and over-caution leave people stranded.
What this isn’t
This isn’t about giving legal advice. It’s not about being reckless with disclaimers. It’s about finding the useful middle ground between “we can’t tell you anything” and “we’re your lawyer” and being transparent about where your tool sits on that spectrum.
05
Use technology purposefully
What this means
Start by asking: is technology the right answer here? And if so, is building something new better than strengthening what already exists?
In an emerging field, multiple approaches to the same problem create value. Don’t suppress experimentation in the name of avoiding duplication, but build on strong foundations rather than creating isolated one-off tools that can’t connect to the wider ecosystem.
If you decide to build, commit to constant testing, evaluation, and iteration. Define what “good enough” looks like. Work toward shared evaluation frameworks across the sector so that tools can be assessed against common, meaningful outcomes.
Where AI is used, it must create demonstrably better social impact than not using it. That case needs to be made honestly and evidenced over time, not assumed at the point of build.
Why it matter
When technology is used purposefully, it solves real problems without creating new ones. The sector learns and improves together.
What this isn’t
This isn’t about building perfect tools. It’s not about avoiding all experimentation. It’s about being honest about whether technology is the right answer, and committing to evaluate whether it actually is.
01
Start with people
What this means
Design decisions should flow from a deep understanding of real needs and experiences. This isn’t a one-time research phase at the start – it’s ongoing. From initial exploration through design, build, launch, and evaluation, you’re listening to and learning from the people who will use your tool, rely on it, fund it, or support others using it.
“People” means a broad ecosystem of actors:
- Communities (end-users) seeking legal information or support
- Lawyers and advocates who refer clients or use the tool
- Service providers and legal aid organisations
- Support workers, community organisations, and trusted intermediaries
- Funders and policy makers
- Back-office teams who maintain and improve the tool
Genuine co-design means involving these people in shaping solutions from the very start, not just testing them at the end. This takes real investment in time, in resources, and in creating the conditions for meaningful participation.
Many people come to access to justice tools in crisis, distress, or with low digital literacy. A trauma-informed approach recognises this. Design for diverse needs, not just the “average user.” Understand what success actually looks like for the people you’re trying to help – not just engagement metrics or completion rates.
Why it matter
When you start with people, you design tools that actually work. You catch assumptions early. You build trust. You’re more likely to create something that closes gaps rather than deepens them.
What this isn’t
This isn’t about designing for one user archetype. It’s not just listening at the beginning. It’s not assuming all users have the same needs, confidence, or access to technology.
02
Build trust through transparency and accountabilityStart with people
What this means
Trust is the foundation everything else is built on. Without it, the most well-designed tool will fail the people it’s meant to serve.
At its core, this principle is about a fair bargain: the tool is useful and accurate; it’s transparent about what it does, what it can’t do, and what happens to data; and it honestly acknowledges the risks involved. End-users, and the professionals who support them, should know what they’re getting into.
Accountability for that bargain is distributed across everyone involved:
Tool builders are accountable for how the tool functions, what data it collects and protects, and its broader system impacts
Service providers are accountable for implementation, support and ensuring the tool is used appropriately
Legal expertise should be embedded in design and governance and not consulted after the fact
Funders and policy makers are accountable for the incentives they create and the standards they require
Build accountability mechanisms into your tool from the start: ways of recording how decisions are made, evaluating real impact, and creating genuine feedback loops with users. Don’t add these as afterthoughts.
Why it matter
When trust is the north star, everything else aligns. And when the sector holds itself to this standard collectively, it builds shared credibility.
What this isn’t
This isn’t about perfect tools. It’s about honest, useful tools built on fair terms and being transparent when things aren’t working.
03
Design for equity, fairness, and inclusion
What this means
Design to close gaps, not widen them and actively test whether your tool is doing so. Understand who your tool serves and, just as importantly, who it excludes. Design for all use cases and circumstances, not just the average user or the easiest scenario.
Efficiency and equity are not always in tension. More timely responses, better triage, and reduced administrative burden are equitable outcomes. The goal is not to avoid efficiency; it’s to ensure that efficiency never comes at the expense of dignity, clarity, and agency. Where trade-offs between equity, scale, and cost exist, name them honestly rather than designing around them.
Many people come to A2J tools in crisis or distress, or with low digital or legal literacy. A trauma-informed approach shapes not just what your tool does, but how it communicates: what it asks of people, in what order, and with what tone.
Bias, particularly in AI and automated decision making, requires active, ongoing attention. Identifying and mitigating bias is not a one-time check. The sector needs clear standards for how to test, evaluate, and communicate about bias in practice.
Consider the systemic context your tool sits within. A well-designed tool can reduce complexity and open pathways, but it can also inadvertently reinforce existing inequities or lock people into systems that weren’t working for them to begin with. Be honest about what your tool can and cannot change.
Why it matter
Equity isn’t about treating everyone the same, it’s about closing real gaps. When you design with this lens, you’re more likely to create tools that actually serve the people who need them most.
What this isn’t
This isn’t about designing a tool that works for everyone equally. And it’s not about solving systemic inequality with a tool alone, but it is about being honest about what your tool can and can’t do in that context.
04
Provide actionable guidance
What this means
People come to A2J tools needing more than information. They need to know what to do next, or to feel less alone in a difficult situation. Guidance isn’t always a concrete action. Sometimes what people need most is clarity, confidence, context, or reassurance that they’re on the right track.
The tension between being helpful and being cautious is real, but over-restriction is itself harmful. Withholding genuinely useful information in the name of legal caution leaves people stranded. You can support informed decision making without crossing into unauthorised legal advice and being transparent about where that line sits for your tool is part of good design, not a limitation of it.
At the same time, be honest about what your tool cannot do. Tools cannot substitute for systemic change or for human expertise when a situation genuinely requires it. Design with that honesty built in, not as a disclaimer bolted on at the end, but as a genuine part of how the tool guides people forward
Why it matter
When guidance is clear and actionable, people move forward. They feel less stuck, more confident. Vague warnings and over-caution leave people stranded.
What this isn’t
This isn’t about giving legal advice. It’s not about being reckless with disclaimers. It’s about finding the useful middle ground between “we can’t tell you anything” and “we’re your lawyer” and being transparent about where your tool sits on that spectrum.
05
Use technology purposefully
What this means
Start by asking: is technology the right answer here? And if so, is building something new better than strengthening what already exists?
In an emerging field, multiple approaches to the same problem create value. Don’t suppress experimentation in the name of avoiding duplication, but build on strong foundations rather than creating isolated one-off tools that can’t connect to the wider ecosystem.
If you decide to build, commit to constant testing, evaluation, and iteration. Define what “good enough” looks like. Work toward shared evaluation frameworks across the sector so that tools can be assessed against common, meaningful outcomes.
Where AI is used, it must create demonstrably better social impact than not using it. That case needs to be made honestly and evidenced over time, not assumed at the point of build.
Why it matter
When technology is used purposefully, it solves real problems without creating new ones. The sector learns and improves together.
What this isn’t
This isn’t about building perfect tools. It’s not about avoiding all experimentation. It’s about being honest about whether technology is the right answer, and committing to evaluate whether it actually is.