Health Chatbots Are Here, Heavily Funded, And We’ll Have To Get Used To Them

Gratitude journaling. Just the thought of it makes me grumpy. To make matters worse, it’s the advice that not one, but two separate mental health chatbots offered after a diagnostic on my state of mind.

Now, don’t misunderstand. I have nothing against gratitude. Or journaling, for that matter. Most of us could probably benefit from time spent reflecting on what makes us thankful. 

But is it really appropriate to hear that from an AI-enabled software platform filling in as therapist? Shouldn’t it come from an actual human?

Subscribe to the Crunchbase Daily

Probably in the future it will all seem normal. That’s because however one may currently feel about chatbot health advice, startup investment data indicates it’s only getting more widespread.

Over the past two years, investors have poured more than $800 million into an assortment of companies developing chatbots and other AI-enabled platforms for health diagnostics and care recommendations, per Crunchbase data. We put together a list of 14 funded players below:

Babylon poised for $4.2B exit

The space is poised to see its first big exit shortly. Telehealth startup Babylon Health announced plans to go public at a $4.2 billion equity valuation through a merger with blank-check acquirer Alkuri Global Acquisition Corp.

London-based Babylon is by far the most heavily funded player in the AI-enabled health chatbot space, having previously raised more than $630 million in known funding, per Crunchbase data. Its service does not purport to take the place of a doctor, but does seek to assist in diagnosing symptoms and helping users figure out next steps. 

Overall, the content isn’t too different from an initial intake interview at a medical provider. A trial run I did for some cold-like symptoms, for instance, went through several Q&A screens before offering up some potential diagnoses (respiratory infection, flu, or COVID-19) and a link for a video consultation. 

The experience is fairly typical of what I’ve sampled in the AI-enabled health and chatbot space. It’s fast, data-driven, and seems to produce sensible suggestions. 

While the term chatbot evokes images of a cheery, human-like avatar, most health offerings look like plain Q&A screens, more reminiscent of a standardized test than a friendly chat.

Horses, zebras and venture capitalists

It doesn’t seem like anything too fancy. But what the chatbot approach does do cost-effectively is direct users to the more obvious diagnoses for the symptoms at hand

As we observed last time we wrote about this space, there is a popular saying in medical diagnostics: “When you hear hoofbeats, think of horses not zebras.” That is, if you have the symptoms of a cold, it’s probably a cold. And that mild rash is probably a minor case of dermatitis that will go away soon.

This all sounds logical. But is a far cry from the diagnosis you would likely make Googling symptoms, which commonly leaves one convinced it’s a case of bubonic plague. 

It’s not just venture-backed startups that find chatbot and self-diagnostic interfaces helpful. Mass General Brigham, a nonprofit hospital network, is one of several regional health providers to roll out an AI-based chatbot to help differentiate between potential COVID-19 cases and less threatening illnesses. The CDC has a similar chatbot offering.

But venture capitalists are the ones putting the most money into the space. In the past couple months alone, two prominent chatbot developers have raised big rounds.

Berlin-based Ada Health, best known for its Ada diagnostic app, pulled in $88 million in a Series B round led by Leaps by Bayer. To date, the decade-old company has raised close to $160 million.

Last week investors poured $90 million into San Francisco-based Woebot Health, developer of an AI-powered app that dispenses mental health advice (including the aforementioned journaling suggestion). The company also recently published a study seeking to demonstrate that its bot “establishes a therapeutic bond with users—a construct long thought to be the unique domain of human-to-human interactions.”

The automation of everything

At this point in the industrial age, we expect machines to do most of the work when it comes to making cars, churning out microchips and processing packaged foods. For other areas, health care in particular, we’re more accustomed to humans. 

Trouble is, there aren’t enough other humans to do all the work we want. We’ve been writing a lot lately about labor shortages, and how automation startups are raising more capital for technologies allowing businesses to keep up productivity with fewer workers. The health care industry is no exception when it comes to staffing shortfalls.

Older adults seeking home health care services are finding long waits and rising costs. Nurses are perennially in short supply.  And in the field of mental health, for instance, Woebot concludes that: “There are far more patients in mental distress than there is time available for mental health professionals to support them.” 

Against that backdrop, it seems health chatbots have a role to play. They’re not likely to replace health care professionals in the foreseeable future. However, by taking on some routine tasks, such as gathering initial symptom lists from patients, perhaps they can free up time for doctors and nurses to focus on other essentials.

If so, then maybe I am a little grateful to have that chatbot around. Just don’t ask me to journal about it.

Illustration: Li-Anne Dias