We love your enthusiasm and interest! We are currently working on our final round of consumer testing, heading rapidly to launch. To be the first to know when we're live and ready to accept project requests, please join our waitlist here. 

Happily, we have already trained our Sensory Bot across several diverse product categories – from cola to crisps, from chocolate to chewing gum.  With food & beverage at the heart of MMR’s expertise, this has been our primary focus leading up to the launch in September.  We are though already working in the background to expand our category coverage further, with personal care and home care products next in line.  This is where our Innovation Panel play a role in helping us understand the sensory touchpoints which our Sensory Bot needs training on and then helping us test and refine the bot. 

Yes, it does! As a result of leveraging Generative AI to generate the Sensory Bot specific chatbot dialogue, we can draw on existing approaches suited to Large Language Models, to successfully deploy this solution across markets, supporting 20 languages worldwide.  There are of course some languages that are less well-resourced than others, but we can review these on a case-by-case basis. In more good news, it would also be possible to set up one Sensory Bot project, that can be deployed across different markets and languages to support multi-country research. 

Very good and very relevant question!  This has been on our roadmap since the outset of this development as we recognize the vast potential to capture the consumer’s product experience, so close to the moment, using voice capture that would simply feel more natural and be far less intrusive. We also love the idea that it gives the consumer the power to decide how they want to share their experience with us. 

We are very excited to confirm that one of our upcoming consumer pilots will test voice capture as part of the Sensory Bot.  Should the results confirm that this is an engaging, valid and accurate way of collecting this data, we are looking forward to including this feature as part of our launch in September. 

Yes, we can!  This is a very good question indeed, as it means that we can combine the deeply nuanced and granular insight into the product experience journey with key performance metrics such as appeal, purchase intent or difference.  In the context of product testing – whether it be against ideal, concept, prototypes, or competitors – clients can still access the key metrics they value along with a deeper understanding of what makes for the very best product experiences that connect with consumers in a way that matters to them. 

As with all AI, we recognize the value of keeping the human in the loop!  For us, the Sensory Bot is almost like a double-edged AI sword – it has AI at the front, where the bot leads the dialogue and consumer engagement, and AI at the back, where the latest AI advances drive analysis and insights generation. While this accelerates speed to insight by processing such vast volumes of unstructured data more easily and meaningfully, the human expertise in interpreting these results in the context of a client’s specific business question is invaluable.  This is where MMR’s deep expertise in product and sensory makes the difference. 

Our Sensory Bot is built upon a robust Large Language Model that has undergone extensive training. Through such training, we are working to embed our specialized, deep sensory expertise into the model.  It is this kind of expert training – followed by experiments and mistakes! – that will make it possible for the Sensory Bot to recognize and elicit highly descriptive and tangible responses from consumers across the full product experience journey in a very dynamic and engaging way. 

The Sensory Bot is not designed to replace moderators but to democratize the skills they have by enabling sensory qualitative research to be done at a quantitative scale yet still on a one-to-one basis.  We are using AI at the front end of our Sensory Bot, to drive a dynamic and sensorially intelligent dialogue, and at the back end, to analyze the ‘big data’ that we access when conducting qualitative research at scale. 

The Sensory Bot has also been designed to work with consumers ‘in-the-moment’ when they are using products.  It interacts directly with consumers capturing the nuance and revealing the complex layers of product experiences as consumers (and sensory panellists) engage with your products.   

Just imagine the potential of a Sensory Bot in those more private or intimate moments when a moderator or interviewer would not be welcome.  As such, the Sensory Bot holds the potential to allow product owners to step closer to the real-world usage moment, even in those moments that were traditionally less accessible, without disrupting it and without losing the granularity of product understanding that is so critical to innovation, yet so often lost in recall. 

And whilst we are imagining, imagine the power of combining the consumer outputs with a sensory panel's objective deconstruction of the product experience. This is sensory ethnography at its best!

We are very excited about the promise we see in the Sensory Bot to bring us closer to the way consumers talk about their experience of flavor and other touchpoints – unearthing cultural cues and consumer signals as it helps us to get to the heart of their product experience, in their own words.  For example, an attribute such as “smoky” in the context of Scotch can signal so many things across different cultures – from heritage and traditional craftmanship to luxury and exclusivity.  For us, this is key to understanding how a product becomes an experience, how it becomes about more than shape and function, as the Sensory Bot steps closer to the lived product moment.

Yes! The Sensory Bot has been designed to moderate consumers through the steps involved in deconstructing the full product experience journey.  For example, with the Tony’s Chocolonely chocolate work, we talked about in the webinar, the Sensory Bot started with the appearance of the packaging, the feel, the sounds it made on opening right the way through to the aftertaste, and the aftereffects of consuming the chocolate.  Although this is a complex process, the Sensory Bot is being designed to work at a level that consumers will be able to readily interact with, while keeping the conversation engaging and dynamic. 

We see such potential in the insights we will collect in this way to deliver a holistic understanding of the total product experience journey that will in turn help us to unpack the sensory product experience per touchpoint, access how congruent it is across touchpoints, how different attributes ladder up to specific emotional and functional benefits, to better understand which experiences underpin such specific conceptual equities.  Altogether, this holds tremendous potential to help our clients optimize and innovate towards winning product experiences. 

The bot is being designed and trained through the involvement of MMR’s unique Sensory Innovation Panel.  Once the consumer Sensory Bot has been launched, we will move to tailor the Sensory Bot to work with sensory panels in deconstructing product experiences objectively (removing the sentiment and emotional elements of the consumer Sensory Bot).  The Sensory Bot could then also be used to train sensory panels, teaching them how to deconstruct product experience in-the-moment. 

We have developed and trained our Sensory Bot working very closely with MMR’s very unique Innovation Panel – a group of trained sensory panelists who support the development of new methods that take sensory research into the future.  As a result, we are able to shape our Sensory Bot conversation with panelists to only focus on the ‘what’ behind the product experience, rather than the ‘why’.  This is important in making sure we only capture the objective product experience from each trained sensory panelist. When guiding a sensory panelist through a Sensory Bot conversation, the bot would always follow a pre-defined order of touchpoints, rather than more dynamically adapting to what matters most to a consumer. 

We can internally set up each new Sensory Bot project using a dedicated chatbot platform.  Once ready to start data collection, the link can be shared with pre-recruited qualitative panelists, or we can share it with a consumer panel for data collection at scale.  In each case, the research participant does not need to download any additional tools, software, or applications.  It is as easy as opening the link on your desktop or mobile browser.

The Sensory Bot is being trained by drawing on MMR’s deep expertise in sensory and product.  Working closely with our internal and external AI partners, we are fusing sensory science and consumer insights, without depending on the data of any one project to support the Sensory Bot’s performance.  Data confidentially is assured. 

The Sensory Bot is being designed to deconstruct product experiences.  For example, if we take the chocolate bar deconstruction from the webinar, the Bot acted as a moderator taking the consumers step by step through the experience of eating chocolate from the first interaction with the packaging through to the aftertastes and aftereffects having eaten the chocolate. 

At each touchpoint, the Sensory Bot probes to elicit those highly descriptive responses that are so valuable, along with what it signals to the consumer, what it communicates to them about the product, and how it makes them feel. It is also learning to automatically detect the level of descriptive depth captured at each touchpoint so that it knows when to probe for more or when to move into the next touchpoint.   

Unbeknownst to most consumers these sensory experiences are very complex.  Yet our sensory-trained bot can moderate this experience in an engaging, responsive, and simple way.  As such, it is the Sensory Bot that needs to be proficient in sensory, not the consumer. 

As with all our solution development, we work closely with our clients in partnership to uncover their evolving needs and shape the parameters of a development. These conversations do not start with a tool or the potential of AI.  Instead, it starts with the client’s business needs.  As such, the Sensory Bot has been specifically designed to meet your research needs and inform your business decisions. 

At its heart, the Sensory Bot is crafted not to sell but to seek out the heart of the consumer experience. By deconstructing the total product experience journey, we will be able to identify the ideal consumer experience, shape product development strategies and optimization directions, and ultimately deliver superior products and product experiences for consumers. 

The honest answer to this is a lot more time than we envisaged!  Externalizing our sensory expertise and then teaching the Sensory Bot to apply it was a lot more complex and challenging than anticipated.  The journey so far has taken hundreds of hours across a diverse team – chatbot specialists, domain experts, and data scientists – not to mention our Innovation Panel. Testing, experimenting, iterating, and providing feedback is an intensive task but rewarding when we can look back to see the growing synergy between artificial and human intelligence. We also anticipate more testing to come as we wrap up our final consumer testing, which is geared to challenge and stretch different elements of this new solution before it is launched at the end of September.

Often when placing an order or contacting customer services, you’ll find yourself interacting with a service bot. Just like our research chatbots, these are designed to simulate a human conversation. But there is a key difference!  A service bot is optimized to close your query as swiftly as possible.  Our research chatbot – like the Sensory Bot – is the exact opposite. It is not only meant to feel responsive; instead, it is designed to be fundamentally exploratory, fostering a sense of rapport and encouraging open sharing.  Furthermore, a service bot is designed to replace a human, which from a user perspective, may be experienced as “trading down”.  A research bot, on the other hand, replaces a 20-40 question survey, which is certainly a “trade-up” for our research participants.