Supporting presenters with an AI based product service system
2022 | PRAISE
PRAISE (Presenter Reflection, Audience Interaction, Speaker Engagement) is a system designed to improve the quality of presentations by facilitating richer communication and feedback between presenters and their audiences. Developed as part of the Artifice squad project at TU Eindhoven, PRAISE utilizes a collection of interconnected devices and interfaces, powered by AI, to gather, process, and deliver actionable feedback, ultimately helping presenters enhance their engagement and tailor content more effectively.
The Challenge: Bridging the Gap Between Presenter Delivery and Audience Experience
Presenters often struggle to accurately gauge audience engagement and understanding in real-time. Simultaneously, audience members may hesitate to provide direct, honest feedback due to factors like social pressure or lack of a suitable channel. Key challenges identified included:
-
Lack of Actionable Feedback: Presenters receive limited specific insight into which parts of their presentation resonate well or poorly with the audience.
-
Audience Hesitation: Concerns about anonymity and potential confrontation can prevent audiences from sharing candid feedback.
-
Information Overload & Distraction: Real-time feedback mechanisms can potentially distract both the presenter and the audience if not carefully designed.
-
Varying Needs & Experience Levels: Both presenters and audiences have diverse needs, preferences, and technological comfort levels that a single solution must accommodate.
-
Meaningful AI Integration: Effectively using AI to provide valuable insights beyond generic tips, tailored to the specific presentation and audience context.
The Approach: Iterative Design Focused on Engagement and Reflection
An iterative, user-centered design process was employed, exploring the potential of AI within the context of presentations and audience engagement. Key design principles and themes emerged and guided the process:
-
Anonymity: Ensuring audience members feel safe providing honest feedback.
-
Confrontationality: Delivering feedback in a non-disruptive, constructive manner.
-
Control: Balancing the input and influence between audience and presenter.
-
Experience: Catering to different presenter skill levels and preferences through modularity.
-
Specificity: Providing feedback relevant to the specific audience and presentation context, enhanced by AI.
The Process: From Exploration to Realisation
The project was conducted in 3 iterations going from exploration of AI based concepts to realising a prototype.
Iteration 1: Exploration, Ideation & Initial Concept
Initial explorations revealed that real-time, complex feedback systems (like dynamic lights/slides) were often perceived as distracting or chaotic. Interviews highlighted the presenter's primary responsibility for engagement, the potential for phone-based input to be distracting, and differing needs based on presenter experience.
The scope was defined: focus on presenters in university lecture settings (30-100 people), aiming to assist skill development and audience connection. An initial concept emerged featuring a physical audience input device (knob), a physical presenter display (showing audience mood), and a post-presentation reflective interface. AI was conceptualized to predict engagement and generate tips. This was prototyped low-fidelity for a midterm demo day.
Iteration 2: User Testing and Concept Refinement
User testing confirmed audience preference for direct, actionable feedback ('requests') over abstract 'emotions'. The physical presenter device was seen as potentially anxiety-inducing and distracting in a live setting, though more acceptable online. Phone-based input remained a concern for distraction. The reflective interface was valued, but needed simplification. Trust in AI was higher for post-hoc reflection than real-time suggestions.
Iteration 3: Insight Synthesis, Design Challenges & Finalisation
We synthesized user study insights into Needs & Requirements. Defined core Design Challenges (e.g., accommodating different feedback types, balancing power, reducing input distraction). Iterated the concept significantly based on these challenges and developed final UI/UX designs and functional prototypes.
Several pivots were made based on user testing. A Discussion Segment was added. The presenter device was made Modular/Optional. Audience Input shifted from 'emotions' to 'engagement scale'. Tips were refined to become more actionable and the role of AI was clarified to generate specific insights an discussion suggestions.
The Outcome: The PRAISE System
The final PRAISE design is a modular ecosystem comprising:
-
Audience Web App: An anonymous, low-effort web-based tool for audience members to input their engagement level and specific, contextually relevant feedback points during a presentation.
-
Presenter Device (Optional): A physical device displaying real-time audience engagement levels (via LED matrix) and key feedback points (on screen), designed for modular use based on presenter preference.
-
Discussion Suggestions: AI-generated prompts displayed post-presentation (e.g., on the main screen or reflective interface) suggesting specific moments (high/low engagement) for focused discussion.
-
Reflective Interface/Dashboard: A post-presentation tool for presenters offering AI-driven insights, tips for improvement linked to specific presentation moents (speech rate, clarity, audience feedback), and visualizations of engagement trends.
PRAISE offers presenters a nuanced, multi-stage approach to understanding and acting upon audience feedback. By separating real-time awareness from deep reflection and discussion, and leveraging AI for specific insights, it aims to foster presenter growth and improve communication effectiveness in a controlled, non-confrontational, and user-centric manner.