Sarah Osman
If I had to describe my 2025 work in one sentence, it would be this: applied social and behavioural science is moving upstream.
This year I’ve worked across vaccine uptake in Cameroon, safety and accountability in nightlife settings in Georgia, and national social protection behavioural design in Rwanda, along with a number of assignments around team capacity building.
On paper, these look like very different assignments. In practice, they all asked the same question: what has to be true in a person’s day-to-day reality for the desired behaviour to become the easiest, safest, most normal thing to do?
So, as we head into 2026, here are the shifts I’m seeing and the ones I’m hoping we will accelerate.
One of the strongest signals for this shift came from supporting a government agency with developing a social and behaviour change strategy. It came from the agreed approach to not focus on campaigning as there was an acknowledgement that this was not providing the desired results. Instead, we chose to focus on system function: embedded into programme design, performance measurement, staff capability, and community feedback loops. That distinction changes the mainstream focus on media campaigns to understanding the bottlenecks caused by the system as a whole.
In 2026, I expect to see more governments and large implementers asking for:
This is good news because when behaviour change is embedded in systems, it’s more likely to survive budget cycles, staff turnover, and shifting political priorities.
The immunisation work I was engaged in reminded me that trust is an experience.
The intervention design was grounded in the WHO BeSD framework and then translated into a human-centred, community co-creation process resulting in two main components, both co-created and led by community members. Importantly, we didn’t just engage community members in the design process, they were active actors in the implementation as well.
Within three weeks, 728 children were vaccinated, and caregivers described shifting from hesitancy to action because the information felt emotionally and socially real.
Two 2026 implications here:
This is the most under-appreciated evolution I see coming.
In many projects I worked on, critical barriers sat at the opportunity level: environmental design, visibility, space structure, and procedures that enable staff to act consistently.
That’s a very different social and behaviour change (SBC) posture. It says: people may already reject a negative behaviour; the question is whether the daily environment makes the alternative possible.
In 2026, I expect the best SBC work to include organisational psychology, address programme design as a whole and look more like:
This is where behavioural science becomes a practical partner to programme design, quality improvement, and implementation.
I’m seeing two parallel trends: more teams experimenting with AI, and more anxiety about doing it badly.
This year, I worked with AI to analyse qualitative data responsibly (stay tuned for a webinar soon where we will explain how!) that provided immense productivity gains, especially when teams are working under time pressure and drowning in data.
At the same time, the field is waking up to the risks: privacy, bias amplification, false confidence in outputs, and the temptation to replace community engagement with AI insight.
That’s why I’m glad to see training offers that explicitly centre responsible use, like the two-day virtual course on responsible AI in SBC programming (run by me and Isabelle Amazon-Brown).
My 2026 take: AI will be most valuable as an assistant, not an authority. Use it to accelerate synthesis, prototype options, translate drafts, and generate testing variants but keep humans accountable for meaning, ethics, and decisions.
I’ve spent a lot of time this year in rooms where people are trying to make a ToC do too much, with too little behavioural specificity.
For one assignment, I developed a guidance note on applying a behavioural science lens to ToC design that makes a simple point that many teams still resist: behaviour is the outcome and intermediate outcomes (increased knowledge, improved self-efficacy, more positive norms) are stepping stones, not the finish line.
In 2026, I expect stronger demand for:
This is where SBC becomes easier to fund, easier to manage, and harder to hand-wave.
In one of my training decks this year, I included a point I think we need to keep repeating: much of behavioural science is still WEIRD (built from Western, educated, industrialised, rich and democratic contexts) and global majority expertise is still too easily sidelined.
What I’m watching for in 2026 is the shift from naming the problem to changing how we work, including:
Equity won’t be a paragraph in a proposal. It will be visible in who shapes the problem definition and what trade-offs are acceptable.
We’re entering a period where SBC teams will be asked to do more with less, and that will expose something many of us already know: a lot of frameworks are good at describing problems, but weak at helping people design solutions.
In a recent deep dive into SBC frameworks, one finding really stuck with me: across 26 widely used frameworks, only six provided practical guidance on how to match behavioural drivers to appropriate techniques. That’s a huge gap, because different barriers need different responses.
In 2026, I think good SBC will increasingly mean: can your team move cleanly from insight → mechanism → technique → intervention → delivery and fidelity? The strongest programmes will have a defensible design logic you can actually implement, track, and adapt.
One of the most useful things we can do for practitioners is make systems thinking feel doable.
A simple behaviour ecosystem mapping approach I’ve been using starts with: who are the key actors around a behaviour? Then you map barriers at different levels (individual, family/community, service/institution, policy). You can explain the socio-ecological model all day but when you put names and roles to it, teams suddenly see where leverage sits.
In 2026, I expect more programme teams to start planning like this because it stops us wasting time on interventions that assume the people we are directing interventions at have full control over the behaviour. It also forces a more honest design question: if the behaviour depends on a teacher, a supervisor, a father, a manager, a policy-maker, or a service queue, why are we designing as if only one person exists?
Right now, many organisations use applied behavioural science, behavioural insights and SBC interchangeably and then get frustrated when a nudge-style tweak doesn’t deliver norm change, or when a multi-level SBC approach doesn’t fit a six-week rapid experiment brief.
These approaches overlap, but they come from different traditions:
In 2026, I think stronger teams will be explicit about the primary lens they’re using and blend intentionally rather than accidentally. That clarity will save time and improve projects.
Applied social and behavioural science is growing up. In 2026, the most impactful work will be the work that builds behavioural thinking into everyday systems.
Best wishes for 2026 and I hope we get to work together this coming year!
👉🏾 How skilled are you at designing for change? Start with the FREE assessment: https://lnkd.in/dK7YPKgR
👉🏾 Join my mailing list for exclusive insights and content! subscribe.osmanadvisoryservices.com