Quantified Communications – Can AI accurately assess human communication skills?
Quantified is revolutionizing communication skills assessment and coaching through AI and analytics. Can machines really assess human soft skills? What challenges and opportunities lie ahead?
We spend roughly 80% of any regular workday communicating, whether it is in person, over the phone, texting or via email. We all spend much more time than we would like sitting in meetings and making presentations. Keeping your audience captive is no easy task, especially in our day and age where people have very limited attention spans. This said, most would agree that being a good communicator is a key factor of success in the business world (and life in general). What if every word and gesture you used impacted the market valuation and reputation of your company? This is most probably the case for C-suite executives of publicly traded companies.
But how do we get better at communication skills? The most traditional approach would dictate hiring a personal coach. However, this poses serious limitations with regards to offering hard, empirical data to analyze performance and recommend a course of action. A coach’s analysis and pointers would be mostly based on subjective observations and, in any case, it is hardly possible to offer a personal coach to every person in your organization in order to improve communication skills.
Enter Quantified Communications, the communication skills platform that aims to help organizations cultivate extraordinary communicators at every level. Technological advances have made it possible to accurately measure how well someone communicates. So, through their work with over 100 Fortune 50 C-Suite leaders and more than 160,000 individuals, Quantified has developed a AI powered technology platform that accurately assesses, develops, and improves an organization’s employees’ communication effectiveness, influence, and performance at any level. Quantified’s human-trained AI technology uses sight, sound, and language understanding to refine people’s communication skills. They index individual strengths, weaknesses, and opportunities for improvement, in addition to providing individualized development plans alongside real-world examples of great communication in action. If that is not enough, they also offer one-on-one coaching sessions. 
Quantified’s value creation formula is very simple: It offers a scalable, personalized, and measurable alternative that traditional communication skills training could never match. Their process is also very simple: Assess, develop, improve. To start, they give you a prompt (e.g. It is your first day at work and you must an introduce yourself to your team at a meeting) and ask you to upload a 2-10 min video. Quantified’s AI technology evaluates your video and produces a personalized report including individual strengths, weaknesses, and opportunities for improvement into personalized development plans. Finally, you improve by completing the online courses tailored to the competencies you need to develop. The results have been encouraging so far: More than 162,000 participants have averaged a 17.4% increase in their ability to influence audiences (QC Score), in just 6 months.
With regards to value capture, Quantified offers three types of solution tailored to different audiences: Enterprise Learning, Executive Communication, and People Science. While the Enterprise Learning solution targets the broader organization and its employees at all levels, the Executive Communication solution targets senior leaders with more relevant skills such as media training, investor communication and strategy communication. Finally, the People Science solution goes after a different market altogether: It offers organizations such as Harvard with an alternative to assess and select candidates, in addition to predict performance and provide comparative benchmarking. By following this approach, Quantified is aiming to capture a considerable part of a large but highly fragmented market estimated to be worth over $10 billion.
Given that this space is new, the challenges and opportunities present are numerous. AI and algorithms are still considered by many as a “black box” that is not to be trusted when the study subjects are human beings. Can machines really understand how we communicate, evaluate our performance, and provide tailored recommendations to improve? Quantified claims this to be the case, and is arguably producing good results, but they are not doing a good job at generating trust. The way their algorithms are built is pretty gray and the inputs that go into them unclear. Yes, they mention that their AI is powered by their “work with over 100 Fortune 50 C-Suite leaders and more than 160,000 individuals”, but machine learning is only as good as its inputs. Who are these people? What are their demographics? Which industries? Why are they considered good communicators? All of this information remains unclear, which is why the solutions offered might produce skepticism with potential clients. Is there bias baked into the algorithm? The only way to assess this would be to really understand what is powering this technology. Quantified should do a better job in educating customers about this, and this would in turn result in a higher level of trust and adoption.
Quantified is paving the way for the future of communication skills training, but the road ahead remains long and windy.
Student comments on Quantified Communications – Can AI accurately assess human communication skills?
James, thanks for sharing a little about this interesting company!
I completely agree with your concerns about the “black box” nature of Quantified’s approach. It sounded to me like there are two separate ways the tech can be used – coaching (for personal improvement) and evaluation. In the case of coaching, I’m less worried… but the ethical implications of evaluating a human being using AI are dubious.
For example, a company using Quantified to analyze a job candidate’s skill level purely from an AI-generated “QC” score raises some serious questions. As you rightly point out, algorithms are only as good as the data they learn from, and if there is bias in the training data (e.g., 99% of “100 Fortune C-Suite Leaders” are men), then scoring can be similarly biased (e.g., against women communicators).
James, thanks for the post. This is very interesting!
I found fascinating how the company creates value in a world where communication is essential. For me a point that is critical is the fact that the process is simple (assess, develop, improve). The company should maintain that important characteristic. A more complex process would reduce the number of people that use the platform.
Going forward, I think it will be interesting how the company tailor the product for different countries. The basics of communication depend on the culture. The AI algorithm will have to take into account that factor when assessing the communication performance of that individual.
Amazing post, James!
The idea seems great on paper, but as you point out in the last piece, it is worthless if they can’t provide clear reasoning behind the AI black box. Even when their algorithm is correct, they are in a complicated place where they need to engage the community of communication researchers to can support the effectiveness of the QC score…and at the same time are introducing a tool that wipes out the jobs of the people from whom they need support.
It’s like the many failed Zebra alike start-ups that positioned themselves as eliminating the job of radiologists.
Great piece. Thank you Hames.
I had the chance to interact with QC a couple of time in the past. I think your articles did a great job in capturing the strengths of the tool. I particularly like the precise, data-driven and bench-marked analyses.
For its weaknesses, I agree with your assessment. I did not like QC inability to take into account contextual differences, spontaneous conversations, cultural differences, biases, etc.
I personally do not see many high-profile executives substituting a human coach (that will tailor made the training to the contextual differences mentioned above) with a product like QC.
Great posting. Thanks!
I do see the value of education piece because, I think, the solution is affordable and convenient for individuals. Alternatives for this service would be personal coaching or self-researching which are expensive in terms of cost and time.
However, as you mentioned, assessing and selecting candidates seem alarming. As input might be biased and have only a limited number of best practices, the result would be fueled by biases. There is no clear cut in why some people “perceived” as good communicators. If the perception is influenced by gender, race, and appearances, I don’t think we can trust the algorithm to replace human judgment.