Improving Employee Engagement and Wellbeing Using Conversational AI

Context

As a non-profit dedicated to applying behavioral science for social good, creating positive impact is our raison d’etre. For this reason, leveraging behavioral science principles to design better products is something that is particularly exciting to us. Hikai, the first of our product ventures, is a conversational artificial intelligence which aims to improve employee wellbeing through targeted cognitive behavioral therapy interventions.

Currently, over 70% of employees in North America are either not engaged or actively disengaged at work. This results in an enormous amount of lost productivity, higher turnover and, in some cases, escalates to mental illness.

Approach

In an effort to tackle this problem, we worked with mental health practitioners in one of Canada’s largest institutes to develop a tool that leverages Cognitive Behavioral Therapy (CBT) techniques in order to address employee engagement. Hikai keeps a pulse of an organization using e-mail based surveys and then offers personalized CBT training modules using a chatbot format. Interestingly, while Hikai has a significant effect on engagement, the fact that it uses clinically validated CBT exercises means that it also has a positive spillover effect on employee mental wellbeing.

Stress and Financial Decision-making

Context

The average American holds about 38,000 USD in debt, excluding their home mortgage. For the majority of the population sustainable borrowing and making consistent repayments is a major source of personal stress. High levels of stress can impact the distribution of cognitive resources. Financial decisions are some of the most stressful, complex, and impactful decisions individuals make, and thus lent itself to an area of inquiry for potential behavioural interventions.

Approach

A top-10 US bank approached us to perform research and develop interventions that can improve the quality of financial decision-making for Americans. We identified stress as a particularly important topic that has thus far been unaddressed and decided to ask the following questions:

  1. What are the negative psychological consequences of stress on financial decision making?
  2. How do individual characteristics such as age, gender, education level, and household income mediate stress levels and responses to behavioural interventions?
  3. What simple and actionable interventions can have significant positive effects on financial behaviours and attitudes?

Impact

Although we cannot yet reveal the results, we did find that:

  • Stress is a strong predictor of poor financial decision making and attitudes
  • A quick 30 second exercise allowed us to create disarm the effects of stress on financial decision making, improving certain financial behaviours and attitudes, even without addressing stress levels directly.

Launching a Behavioral Science Unit Focused on Citizen Engagement.

Context

For societies the world over, the pace of change is accelerating. This rapid evolution is straining legacy government institutions, which struggle to keep up in this agile, iterative landscape. These challenges have been compounded in recent years as governments are expected to do more with smaller and smaller pools of resources from which to draw. In response to these challenges, behavioral science has been proving itself a valuable tool in the public sector. With its experimental, data-first approach, it offers a model for new ways of working within government, ways that are well adapted to the rapid evolution of society. Furthermore, by focusing primarily on communication rather than program structure, behavioral science has offered tangible increases to government effectiveness without the heavy costs that are often associated with such improvements. Together, this new way of working and the momentum built up behind a string of early wins are helping to propel behavioral science as a tool for modernization for governments worldwide.

Approach

When the City of Rome decided to launch an internal behavioral science practice, it approached The Decision Lab as a premier partner to provide guidance in a variety of thematic contexts, including: environment, healthcare, political participation, energy and public services. In addition to helping with individual thematic projects, The Decision Lab has also contributed key inputs into the organization of the internal unit, helping it to establish itself as a light and effective unit that can deliver value to the citizens of Rome for years to come.

Leveraging Nudges to Improve Mental Health in At-Risk Individuals

Context

A growing body of literature has shown that mental health issues among National Guard members are a serious issue – one that is often caused by the challenges that members face and are exacerbated by a culture of silence as well as possible career effects of revealing the issues. As suicides have reached epidemic proportions in the Guard, it is critical to understand the intricate system of beliefs and values that stops at-risk individuals from seeking help.

Approach

In working with the National Guard, The Decision Lab started with a thorough literature review of the topic. Although there is a wealth of information on the predictors of suicide in the National Guard, our unique approach combined insights from the mental health space with those from the behavior change literature in order to create a model explaining what factors can push someone to get help in one situation but keep them silent in another. We found several factors to be particularly predictive of help-seeking behavior and designed a series of interventions that take advantage of these findings in the real world.

Improving Adoption of Hearing Aids

Context

Nessa is a leading supplier of hearing aids in Southeast Asia. They are on a mission to address a deep-rooted public health problem. Although as many as 10% of adults need a hearing aid, only 15% of those that need one actually use one on a regular basis.

Challenge

We worked with Nessa’s scientific team to determine the exact roadblocks to adopting and adhering to a hearing aid. We then translated these roadblocks in behavioral terms, pinpointing possible bias explanations for each one and finally used the behavior change frameworks previously validated in the public health space to design interventions that address these biases in the SEA population.

One important insight that came from our work with Nessa is the importance of building a therapeutic alliance. In particular, we found that, regardless of the product being offered, a critical predictor of hearing aid adoption and uptake is the relationship that an audiologist had built with a patient and the extent to which a patient reveals their real reservations about using hearing aids (which can be very diverse between patients). In an effort to change this, our recommendations focused on ways to create an audiologist-patient alliance and shift the diagnostic sessions from transactional to collaborative.

Increasing Trust and Reducing Perception of Waiting Time Through Music

Introduction

Americans spend 37 billion hours a year waiting. Man Made Music (MMM) designs the sounds that many of them are exposed to while they wait – sounds that not only make time pass faster/slower but can also change their predisposition as they enter a room to see their doctor or call in to speak with a financial advisor. The total impact of soundscapes is enormous and yet the effect that sounds have on things like trust, likability, and openness are still poorly understood.

About Man Made Music

Fast Company named Joel Beckerman, the founder of NYC-based Man Made Music, one of the “Most Creative People in Business” and called Man Made Music one of the “Most Innovative Companies” in the world. MMM uses the power of soundscapes to develop signature sonic branding programs for global giants such as HBO (e.g. their iconic static intro sound), Disney, AT&T, and Southwest Airlines. Recently, MMM took a scientific direction in order to explore the psychological and physical benefits of different soundscapes – especially in the context of waiting.

The Ask

MMM approached The Decision Lab to conduct rigorous research on the types of sounds that make waiting times better. Since their clients include some of the largest consumer-facing companies in the world (including several big banks), the total impact of their sound design is enormous. The final goal was to create insights that serve a threefold purpose: to be mediable, actionable and scientifically rigorous.

Approach

Given these goals, we went through a similar alignment process that we have described in this proposal and came up with concrete research questions that fit gaps in existing research, was likely to produce interesting output for consumers and could produce actionable recommendations for MMM. Based on these research questions, we designed a study investigating the effect of a wide range of sounds on people’s perception of their environment. In particular, we were interested to compare how different sounds scapes (happy music, sad music, white noise, coffee shop sounds, a story, and a repeated message) would affect people’s ratings of:

  1. How much time has elapsed as well as their confidence level
  2. Feelings of happiness, sadness, annoyance, relaxation
  3. Propensity to trust the organization they are dealing with

Using the MTurk platform and PsyTurk, we randomly assigned a representative sample of 184 participants (87 females, 97 males, all American, ages 18-65) to various conditions. Participants were asked to listen to a 120 second sound and were then asked a series of questions. Results were analyzed using Matlab and Python statistical libraries and were then socialized with leading researchers in this space. Interestingly, we found that the best performing sound can ‘speed up’ time perception by 19% and increase trust by 40%. Furthermore, people had no idea this was happening and actually rated the sound relatively low on scales typically used to assess success. The final output of the study was turned into a whitepaper by our team and is currently being handled by MMM’s PR partners in preparation for publication on venues such as FastCo, HBR and TED.

Impact

  • We can take the simple case of financial advisors operating in a bank 
  • 45% of Americans do not think they have enough money to warrant making a plan, 20% do not know how to make a plan and 16% do not think they can stick with a plan
  • Level of trust towards an advisor is the biggest predictor of the likelihood of following their plan

Our experiment has shown that music can increase trust by 39%, impacting the 16 to 20 percent of Americans who encounter behavioral barriers toward long-term goals

Improving Financial Inclusion Through Digital Products

Context

Mylo is a leading Canadian fintech company with over 100M in assets under management and 400,000 users. They are tackling financial inclusion among younger populations in Canada. Typically, millennials are underserved by savings products – over half of them have less than 1000$ in savings. This results in serious setbacks when it comes to concrete financial goals (such as home ownership) and unfortunately introduces unsustainable borrowing behaviors that only exacerbate the issue.

Challenge

In an effort to improve their impact, TDL worked with Mylo’s product teams to better understand key user journeys – in particular, the decision moments that translate into key behaviors such as signing up for the app, linking it to a bank account and sharing it with others. After identifying key levers that affect these behaviors, we designed a system of social proof tools and time-cued nudges that work to shift use behaviors toward KPIs.

Reducing Bias in Venture Philanthropy

Context

The Skoll Foundation, started by Jeff Skoll (eBay’s first President), aims to identify organizations that have a strong potential for equilibrium change and fund them to the tune of 1.25M. So far, the foundation has disbursed close to half a billion dollars and is looking to validate that its approach is indeed consistent with its values and intentions.

Challenge

Each year, Skoll must go through hundreds of applications and work as a team to make extremely complex choices about which organizations deserve more attention and, eventually, funding. To that end, we worked with Skoll to design various psychometric tools capable of identifying bias at various levels of their decision making process.

Increasing Clean Cookstove Adoption in Uganda

Context

Inefficient biomass stoves cost Africa over 50B USD per year and cause irreparable damage to the environment through pollution and deforestation. Efficient alternatives exist for <10$ – they use up to 80% less fuel and have the potential to save millions of lives. However, uptake of these stoves remains low due to negative consumer attitudes and behavioral barriers around the purchase process.

The World Bank operates a number of projects focused on economic empowerment in East Africa. One of the most impactful of these is their Clean Cookstove initiative in Uganda – a country where inefficient cookstoves cause more deaths than malnutrition or poverty. Although a number of clean cookstoves are available to consumers, there are issues of fraudulent practices and scams in the stove market, in addition to issues of ineffective marketing strategies adopted by Ugandan distributors. There is also a lack of alignment between marketing collateral and consumers, leading to low affinity.

Approach

Past work by the Bank and other para-governmental institutions focused on shifting health awareness. However, this approach has been ineffective – our team worked with the Ugandan government, distributors, manufacturers and the Bank to carry out research on hundreds of households in Uganda aimed at understanding stove usage habits, penetration methods, and biases involved in decision making during the purchase funnel. This fieldwork revealed that the biggest barrier to uptake is the perception that ‘investing’ (i.e. buying a more efficient stove so save money in the long-term) is seen as a luxury. Our recommendations use behavioural economics techniques (nudges delivered through a centralized SMS systems) to reverse this bias and reframe baseline stoves as the ‘luxury’ option. Based on experiments, this approach significantly increases willingness-to-invest. The project also resulted in the creation of reports that are currently being published by the Bank and used to inform decision making at the private and government levels in Uganda.