FOR ORGANISATIONS

You've introduced the technology. Now you need to understand what it is doing to your people.

AI adoption changes how people work. It changes how they make decisions, what they check, what they disclose, what they remember and what they trust.

I work with leadership teams, HR and OD, and digital and IT leaders to understand those changes and build AI adoption around human judgement. Much of my work is in the NHS and public sector, and the same questions come up wherever professional judgement matters.

Start a conversation

Your AI adoption figures don’t tell you whether judgement is being protected.

Your people may be using AI confidently and delivering more work. Whether they can still recognise when the system is wrong is a different question. So is whether they can explain how a decision was reached, whether they know when AI should not be used, and whether your policies reflect what people actually do.

The challenge

Most guidance relies on a person checking the output.

That check is the part most likely to weaken quietly over time.

Ways to work with me

FOR SENIOR TEAMS

Leadership workshop
For senior teams shaping their organisation’s approach to AI. Usually a facilitated day, designed with you beforehand around the decisions you’re facing.

You leave with a shared view of where judgement is at risk, what has to stay human, and principles they can take to their Board.

FOR TEAMS: 2 HRS OR HALF A DAY

Judgement by Design
For teams ready to work on their own processes. In two hours we take one real process, or in a half day two, and map it as it actually runs. Together we agree what stays human, where checks belong, and what the time saved will be used for. Sometimes the answer is AI. Sometimes it is simpler automation, or a better process.

Either way, the team leaves with a workflow they have designed themselves and can defend.

TO UNDERSTAND WHAT IS HAPPENING

Behavioural AI Risk Audit
For organisations that want to know what is really happening before they scale further. The audit looks at where AI is entering decisions, where judgement may be being displaced, and where governance does not match behaviour. It tests for automation drift, judgement erosion and misplaced trust.

You leave with a clear picture of current practice and the priorities that matter most.

FOR LASTING CHANGE

12-week programme
For teams ready to change how they work across the board. It combines a behavioural risk audit, governance design and a development programme, From Confidence to Judgement, with in-person workshops and accountability sessions throughout.

You leave with agreed ways of working, governance that matches behaviour, and a team that can keep judgement in the lead.

NHS communications teams can book Judgement by Design and the 12-week programme through HealthCommsPlus.

Reach out for a conversation

What people say…

“Having an expert bring professional credibility to the messages allowed them to land more directly with the audience. I’ve had a lot of direct feedback on how successful the session was and how much people enjoyed the presentation and facilitated session.”


Chief Information Officer, NHS Trust

“We talk about keeping a human in the loop, but Sonya has really made me question whether that person has the confidence, capability and organisational permission to challenge what AI produces.”


HR and OD leader, international HR and OD event

Why me

My work sits where three areas meet.

Twenty-five years in NHS communications and digital transformation. Ten of those years were spent leading communications on major digital transformation programmes, including electronic patient record implementations. I hold an MA in Politics and Public Policy and have worked in government policy, including at the Number 10 Policy Unit and with the European Parliament.

Psychology. I hold an MSc in Cyberpsychology and am a certified business psychologist. My work looks at how technology changes behaviour, trust, disclosure, decision-making and human judgement inside organisations.

AI adoption and governance. I was a founding architect and board member of the NHS Communications AI Taskforce, where I designed the governance architecture and developed the first NHS operating framework for AI use in communications. I advise as an AI Ambassador at the Department of Health and Social Care.

The result is a perspective built around the people using the technology.

What this can help you understand

  • Where AI is becoming embedded in decision-making

  • Where automation bias or over-reliance may be developing

  • Whether people understand the limits of the systems they use

  • Where policy and actual behaviour diverge

  • Whether people feel able to question AI outputs

  • Where professional expertise is being strengthened or displaced

  • What leaders need to know before scaling AI further

Ready to look at what is happening in your organisation?

Let’s talk about where you are, what you’re facing, and which way in makes sense.

Get in touch

Frequently Asked Questions

  • If your leadership team is still shaping its approach, start with the leadership workshop. If AI is already in wide use and you want to know what it is doing, start with the audit. If a team has a specific process it wants to get right, start with Judgement by Design.

  • AI training teaches people how to use a tool. AI adoption is the harder, longer work of integrating AI into the way an organisation makes decisions, without eroding the professional judgement those decisions depend on. Most organisations invest heavily in training and underfund adoption. The result is confident AI use that is not necessarily competent AI use.

  • Judgement erosion is the gradual, often invisible decline in a professional’s capacity to make independent decisions when AI tools are regularly used in place of thinking. It accumulates through small, repeated acts of accepting AI output without interrogating it, until the ability to work without the tool, or to catch its errors, is significantly reduced.

  • Automation drift is the slow shift in what a team accepts as good enough. Over time, people approve AI outputs they would once have questioned, and the standard itself moves towards whatever the tool produces. Nobody decides to lower the bar. It moves a little at a time.

  • The difference shows most clearly under pressure or in edge cases. Teams using AI well can explain why they accepted or rejected an output, can do the task without the tool if they need to, and apply consistent criteria when AI and human judgement conflict. Confident but less competent use tends to show up as uniform outputs, less questioning, and difficulty explaining decisions afterwards.

  • Leadership teams, HR and OD, and digital and IT leaders in organisations where professional judgement has real consequences, whether financial, reputational or human. Much of my work is in the NHS and public sector, and I also work with organisations in other sectors facing the same questions.

  • Sessions are in-house, with your team, built around your context and your work. The work draws on cyberpsychology, the science of how technology affects human cognition, trust and decision-making, applied to how your organisation is using AI now. Participants leave with shared criteria for consistent, defensible AI decisions.

  • No. The aim is calibrated trust. Teams that are appropriately critical of AI output make better use of it. They know when to rely on it, when to interrogate it and when to override it. Blanket scepticism is as unhelpful as blanket acceptance.