For decades, focus groups have been the primary method for innovation teams to understand users. Although they offer insightful information and authentic input, they also require time-consuming recruitment, incur high costs, and are prone to inherent biases that may affect outcomes. These disadvantages were tolerable in slower innovation cycles. These days, they create a gap since important decisions are frequently made before insights become available. Synthetic user research addresses this challenge by extending, rather than replacing, conventional methods.
From static personas to interactive counterparts
Synthetic users are AI-powered personas based on extensive language models. Unlike in traditional workshops, these personas are interactive, responding and reacting in real time. They create a digital version of the target audience based on multiple inputs, including demographics, behaviors, motivations, and ethnographic insights.
This approach changes how ideas are evaluated. Concepts can be tested right before validation, and features can be assessed prior to development. Unexpected resistance, such as hesitation, contradiction, or previously unconsidered concerns, frequently yields the most insightful discoveries.
By leading conversations, delving further, and analyzing responses, researchers maintain close involvement through a human-in-the-loop approach. The value is not found in automation alone, but rather in the combination of speed and human judgment.
A rigorous study logic
Identification and customization of personas are the first steps. Personas can be derived from a larger, structured database or constructed precisely around an idea, depending on the goal. Upon modification of behavioral characteristics and expectations, they are adjusted to reflect a specific setting or industry.
Interaction and simulation come next. A researcher introduces a value proposition, concept, or design trade-off. The synthetic user reacts, questions presumptions, and draws attention to areas of conflict, often bringing to light responses that would be challenging to predict internally.
A ‘human-in-the-loop’ approach comes into the stage to increase the level of depth and enhance the understanding. A human moderator interacts with the AI personas more in detail to examine answers and delve into underlying motives. This is crucial for determining the ‘why’ and not only what succeeds or fails.
From assumptions to structured exploration
This approach is especially powerful in managing uncertainty.
Instead of relying on assumptions, teams can run structured simulations. Multiple concepts can be explored in parallel across different personas, contexts, or markets.
This enables A/B testing much earlier in the process. Variations can be compared before developing a prototype:
- Does a faster feature justify increased noise?
- Is a lower price more compelling than improved repairability?
- How does the same value proposition resonate across different markets?
By running these comparisons repeatedly, even at scale, patterns emerge. This shift helps teams move from instinct to evidence and reduce early-stage risk before real-world validation. It ensures teams build the right solution before refining how it is built.
An approach that travels across contexts
While this method is rooted in product development, its relevance extends far beyond it.
- In healthcare, synthetic personas can help anticipate how less tech-savvy users interact with new devices.
- In retail, they can reveal how messaging shifts between cultures or consumer mindsets.
- In sustainability, they can surface how people navigate trade-offs between cost, durability, and environmental impact.
- In UX and service design, they can expose friction in journeys that don’t yet exist.
These applications are unified by one logic: testing decisions against realistic, behavior-driven perspectives before finalization.