Our engagement practice works closely with our in-house insights and digital team to scrutinise and assess developments in AI engagement and bring the best ones to our clients. Our in-house reputation practice also plays a role in assessing potential risks, specifically stakeholder or community sentiment towards AI-assisted engagement.
We use artificial intelligence (AI) to reduce the time and cost it takes to get from engagement to meaningful insights for our clients. We appreciate, of course, that there are varying levels of comfort in the use of AI to capture, analyse and present information and we work with our clients to design approaches that work for them. We also ensure human mediation thresholds – the extent of our (human) intervention to ensure accuracy – are agreed on with clients and strictly upheld and that the technologies we use are secure – protecting data privacy through encryption and access restrictions.
Our AI-assisted approach can incorporate AI at each point in the engagement process:
- Design: prompting large language models (LLM) e.g. ChatGPT-4 Omni to assist with the creation of questions and communication
- Synthetic audiences: using AI to build detailed personas representing audience segments of interest to our clients, which can then be interrogated for market intelligence
- Data collection and collation: using platforms like Qualtrics and Q Research to collect and structure data
- Digitisation: using automatic speech recognition (ASR) and optical character recognition (OCR) technologies to capture and digitise information gathered during engagement sessions/workshops
- Analysis: using walled garden models to analyse and interrogate large datasets
- Visualisation: using SenateSHJ’s proprietary dashboard to present data in real time
- Reporting: using AI to assist with writing and data visualisation.
Practical application of AI-assisted engagement includes:
- Simulated focus groups: We use synthetic audiences to complement more traditional face-to-face focus groups, usually as a first pass to test ideas, concepts or messages, and draw out worries and concerns.
- Simulated community reference groups: Similarly, we can use our synthetic audiences tool to create a replica community reference group to identify priority areas of support or concern which warrant face-to-face engagement.
- Rapid reporting: We use AI to capture conversations, Post-it note content, whiteboard ideation and other intelligence from face-to-face engagement and turn it into a digital format which can be quickly analysed for themes and insights. We can also ask the model specific questions of interest which emerge during the process.