Huckleberry

Our mission is to democratize access to child behavioral specialists. Every family should be able to receive expert & specific guidance so they can thrive and live their best life.

Every family with young children has a need for behavioral guidance, whether related to sleep, tantrums, positive discipline, eating, or speech & language. Specialists exist for each of these areas, but many families do not access them due to cost, inconvenience, stigma, feeling that their issues do not warrant a specialist, or being unaware that such expertise exists.

Huckleberry augments human specialists with AI, thereby reducing the cost to receive an analysis and plan for most families by 90%. Our AI writes the behavioral plan for the specialist, who then checks and confirms the plan, rather than needing to write it from scratch. 

Through this global service, we can help every family thrive and provide early intervention to all.

If a parent is too tired to do brain-­building activities, or a child is in no mood to learn, you can have all the books and learning resources, but they would be going into a "leaky bucket" if the child is not sleeping well, or there are other behavioral issues going on.

Research shows chronically disrupted sleep is a widespread problem with many ramifications, including school performance, cognitive development, hyperactivity, obesity, maternal depression, and family disruption.

Parents want their children to sleep better, eat better, and engage in enriching activities, but often the reality is that they just don't know how to improve sleep, their child won't eat the vegetables, and is throwing tantrums. This takes its toll on the family's stress level and parent-child relationship.

In moments like these, generic tips or parenting lessons are either ineffective or too much for the parent to handle.

There are child specialists who are able to help in these situations: pediatric sleep consultants, occupational therapists, applied behavior analysts, child psychologists, speech & language pathologists, and more.

However, they are expensive or limited to those with more acute issues.

Huckleberry uses AI to make these specialists accessible to all.

Our target population is parents of 0-5 year olds, and their children. With Huckleberry, it costs less than a co-pay to receive an analysis & plan from an expert, created just for them. We also have a free tracking app with predictive algorithms.

Over the years of working directly with tens of thousands of parents, and indirectly with over 150,000 through our app, we have come to understand that everyone has different needs and situations. They need very specific guidance accordingly. Both our human experts and AI take into account schedule constraints, housing constraints, child disposition, parenting philosophy, and more. 

We have created sleep improvement plans for a student mom whose dad came to care for her baby while she was in class. We helped grandparents who found themselves with custody of their twin grandchildren. We helped a family who all sleeps together in the same bed. 

We continuously gather feedback from the experts and the families' progress to improve our system and processes. By combining human experts with AI, we are able to build a solution that works for every situation in a scalable way.

We augment human experts with AI by drastically improving the time it takes to write up a report, thereby cutting costs.

Normally when a therapist works with a client, they have in their head what they would recommend. Then they spend 60-90 minutes putting a plan together for the client.

Huckleberry uses AI to do a first pass on generating that plan. It is not simply choosing from a set of pre-canned plans. Rather, the AI system builds each plan from the bottom up, taking into account over 50 metrics, and writes it as a human would. Now the therapist just needs to check, confirm, and modify where needed. Any changes the specialist makes is then fed back into our system, along with progress data from the family.

A single plan could be 10-15 pages long if it were all in one document, but it is presented to the parent in an easy-to-digest hierarchy, broken down by steps that start with easy wins. Parents can indicate whether they started on a recommendation, and how it went. They can get modifications or ask follow-up questions as needed.

Parents access Huckleberry through an app, which increases accessibility. Rather than needing to set up and travel to an appointment, parents provide information through the app and answer an adaptive questionnaire that mimics speaking with an expert. 

We have started with sleep, and will be expanding into additional areas, including feeding/eating, behavior (e.g., tantrums and positive discipline), and speech & language development.

Why is this level of specificity needed, and why do general parenting tips not work? Let's take tantrums as an example. A time-out can be effective (if done properly) when a child hits because he wanted something he couldn't have. However, suppose you asked the same child to clean up, and then he started hitting and threw a tantrum. A time-out in that case reinforces the negative behavior because he just got out of what he didn't want to do. Our experts and AI would be able to look for patterns, codify the root of the tantrums, and provide guidance on how to prevent the tantrums on the positive side, and negative consequences that fit the scenario, child, and parent.

By providing guidance that respects both child and parent, we help to establish secure caregiver­-child attachments, increase caregiver knowledge, increase children's social/emotional competence, and reduce the risk of abuse and neglect.

The key to our innovation is in combining human experts with AI. The human experts allows for differentiated and specific guidance between a single mom who works late at a restaurant and picks up her kids late from her parent's house, versus a new mom with a colicky baby who has reflux (both real examples). The AI makes it scalable and affordable.

We augment human experts with AI by drastically improving the time it takes to write up a report, thereby cutting costs. Normally when a therapist works with a client, they have in their head what they would recommend.

Prior to building out our AI system, it took the human experts an average of 60-90 minutes to put an analysis and plan together per family. Our AI has improved to the point that it now takes an average of 10-15 minutes to check and review the AI-generated plan.

Any changes the specialist makes is then fed back into our system, along with progress data from the family, to continuously improve our AI system.

Normally it costs a family $100+ for a session with a specialist. Through Huckleberry, it costs $15 to get a full analysis and custom plan.

The core technology being used are AI (symbolic AI, expert systems, and machine learning), a robust portal we developed for the human experts to use that aggregates data into easy-to-understand metrics and flags, cloud computing, and our apps on both iOS and Android.

We started with sleep, which is the #1 challenge for new parents (According to the National Sleep Foundation, 76% of parents want to change something about their child's sleep), and will next expand into tantrums/positive discipline.

Research shows chronically disrupted sleep is a widespread problem with
many ramifications, including school performance, cognitive development,
hyperactivity, obesity, maternal depression, and family disruption.

85% of our users report that their child's sleep has improved as a result of Huckleberry, with another 10% either not having gotten around to working on sleep, or citing external circumstances such as illness or travel as preventing sleep improvement.

We have not ourselves measured the secondary benefits, such as a decrease in family disruption or maternal depression, but existing 3rd-party research has already shown the link.

By providing situation-based guidance that respects both child and parent, we expect to establish secure caregiver­-child attachments, increase caregiver knowledge, increase children's social/emotional competence, and reduce the risk of abuse and neglect.

Current: 150k cumulative

In 1 year: 1.5M cumulative

In 5 years: 50M cumulative

Our goal within the next year is to reach 1M families, and replicate what we have done with sleep, into tantrums, feeding, and speech & language development.

In 5 years, we aim to continue beyond the first 5 years of childhood. The early elementary age is when most ADHD is diagnosed, as children are now expected to sit in a classroom and do homework.

N/A

14 total:

• Full-time: 4 (with 1 more joining soon)
• Contracted child specialists: 7
• Contracted software developers/designers: 3

The CEO started started this venture out of personal experience, after her son woke every 2-­3 hours for 20 months. With a triple-major in Electrical Engineering & Computer Science, Math, and Statistics from UC Berkeley, MBA from the University of Cambridge, and startup management experience, Huckleberry is the culmination of her academic, professional, and personal experience.

Our team now also includes 2 Berkeley Ph.D.'s in computer science and predictive analytics in biological systems, a sleep researcher who was 10 years at Harvard Medical School, and 7 experienced child specialists in sleep and applied behavioral analysis.

This unique combination of AI expertise and subject-matter experts in early childhood allows us to develop this fresh approach to children's behavioral medicine. Rather than merely providing tips to parents, we're able to mimic the thought process of human experts and provide the same level of specificity and guidance, at a fraction of the cost.

• Promise Venture Studios and the Harvard Center on the Developing Child: (through their summer 2019 fellowship) to better understand the needs and delivery methods to low-income populations in the US.
• Baby Sleep Science: we refer families who would like a one-on-one video consultation with pediatric sleep experts, and they refer families to us for data capture and a lower-cost way to access personalized sleep guidance.

Huckleberry currently has a B2C freemium revenue model, with future channels into the workplace, pediatrician offices, and government organizations.

The end-customer and beneficiaries are parents and their children. Currently they sign up directly through our app or website, and interact with Huckleberry digitally. They have the option of using just the free app for data logging and algorithmic guidance (e.g., SweetSpot: the world's 1st real-time predictor of the optimal next time to sleep--user reviews have literally compared this to Nostradamus and the best thing that happened to them as a parent), or upgrading to get a full analysis and custom plan by our AI-augmented human experts.

In the future we would like to expand to government & impact organizations to better serve the low-income community. It is an important part of our mission that we be affordable and accessible to all families. We would also like to reach families through workplace benefits and pediatrician offices.

This way we can maximize our reach, to strengthen families by increasing knowledge of parenting and child development, and increase the social and emotional competence of children.

We have raised investment capital and are also generating revenue.

We are most excited about being a part of the Solve network to contribute however we can, and to establish connections into impact organizations in the early childhood development space to better understand their needs and ultimately lead to pilots or distribution. We would also like to gain input from AI technologists as we expand our behavioral services.

N/A

We would like to partner with impact organizations who already serve families with young children, such as Help Me Grow and their affiliates, Early Head Start, First 5, and nurse-family partnerships.

We would also like to partner with children's behavioral health organizations to fit into a system of care for families.

We would use the prize to advance and accelerate our ability to ramp up new services, so that we can make more types of child specialists affordable and accessible to all families.

When we developed the solution for improving sleep by augmenting human experts with AI, much of the initial work was through designing expert systems and symbolic AI, with machine learning being used to determine thresholds. 

As we expand into new services where there is an evaluation & plan provided to a family, including tantrums (using applied behavioral analysts), feeding (occupational therapists and nutritionists), and speech & language development (speech & language patholigists), we will be developing AI methods to more quickly parse human-generated plans into our system and determine when certain wording and advice is relevant.

We would use the prize to bring on a machine learning expert full-time who we know well and has also done some consulting work for us in the past. He is currently at a Fortune 50 company, has 10+ years in algorithm development and machine learning, is incredibly bright and as a parent, is ready to join Huckleberry's mission of democratizing access to child specialists to all families.

N/A

We would use the prize to advance and accelerate our ability to ramp up new services, so that we can make more types of child specialists affordable and accessible to all families. This would decrease the resource and achievement gap among families.

When we developed the solution for improving sleep by augmenting human experts with AI, much of the initial work was through designing expert systems and symbolic AI, with machine learning being used to determine thresholds. 

As we expand into new services where there is an evaluation & plan provided to a family, including tantrums (using applied behavioral analysts), feeding (occupational therapists and nutritionists), and speech & language development (speech & language patholigists), we will be developing AI methods to more quickly parse human-generated plans into our system and determine when certain wording and advice is relevant.

We would use the prize to bring on a machine learning expert full-time who we know well and has also done some consulting work for us in the past. He is currently at a Fortune 50 company, has 10+ years in algorithm development and machine learning, is incredibly bright and as a parent, is ready to join Huckleberry's mission of democratizing access to child specialists to all families.

Data: data is ethically sourced as it is 100% voluntary and input by parents. It is stored encrypted and follows GDPR regulations. It is never shared with a 3rd party unless the user grants explicit permission to do so (e.g., with a pediatrician).