Utilizing AI and other automation tools to enhance the delivery of human services can make exclusive services like career coaching available to more people. But for this to happen, the human and tech aspects of a service must be balanced.
My study of 65+ startups operating as tech-enhanced human services (TEHS) suggests that while companies should be ambitious about what tech can do for scaling a human service, they should also ensure that the human component of the service is not compromised.
The rise of tech-enhanced human services
Many startups are automating parts of tasks historically performed by people while holding on to the human quotient for the parts that provide the best outcomes for clients.
This is based on the premise that because certain services, like concierges, require people as part of the solution, fully automated apps are not enough. But, human time is often expensive and limited.
By integrating technology with human expertise, companies can render their services more affordable and accessible.
TEHS companies can be found in numerous industries — I’ve identified 66 such companies, of which 19 are unicorns operating in over 10 verticals. These companies all use various technologies to streamline parts of their service that require the time of a human expert.
Image Credits: Alon Laniado
For example, Wishi connects people with experts who help them shop for clothing and other personal items online. It uses surveys and algorithms to match clients’ preferences to clothing inventory. This is further curated down to a final product list by style experts who chat with clients about individual needs, including specific attire for events or incorporating the requirements of a medical condition.
By reducing the time a stylist needs to spend on the process, Wishi is able to charge $40 to $90 per client. The company’s co-founder, Clea O’Hana, says partners like Farfetch and Saks Fifth Avenue who use Wishi have seen a sharp increase in transactions.
“[That’s] thanks to the more granular data we capture from clients and the relation of trust they have with their stylist,” says O’Hana.
Is my service “tech-enhanceable”?
Three conditions appear necessary for a company to enhance their service with tech.
- A broad segment of the population could benefit from the service, but it is currently inaccessible or cost-prohibitive as a result of the experts required to deliver it.
- Some human participation is essential to delivering high quality of service and cannot be fully automated.
- Some tasks can be automated without affecting quality. For example, solutions for losing weight are limited enough in scope for AI to replace parts of the conversation between coaches and clients.
Example: Talent attraction
I first started working in TEHS when I co-founded PathMotion, which connects job candidates with employees of organizations they are considering applying to.
Candidates prefer to ask questions to employees of prospective employers rather than receive marketing information from that company. While companies also welcome the idea of their employees acting as ambassadors, they are concerned with the amount of time an employee would have to use to chat with potential candidates.
We realized that only a fraction of candidates asked a new question to employees. Others relied on a database of previously answered questions.
We capitalized on this insight by ensuring the candidate search journey included a variety of automated ways to find relevant answers to their questions. This included AI tools that categorized thousands of Q&A topics, as well as tools that provided candidates with closely matched answers before they submitted their questions.
An employer can have 20 “super ambassador” employees serve thousands of candidates by answering a couple of new questions per month.
Leverage tech to replace human tasks
Low-tech automation tools can save significant time spent on repetitive tasks.
Companies like Rocket Lawyer are democratizing access to legal services by reducing costs through the use of chat tools to answer common questions. When that is not enough, they redirect clients to live legal advisers to respond to nuanced inquiries.
Others are taking it a step further.
Mental health
While most mental health companies use low-tech tools to match clients with the right therapists, Talkspace uses machine learning (ML) to identify behavioral patterns and potential harm risks, sending push notifications to therapists in real-time if an elevated risk is detected.
It also provides therapists with recommended tips and actions to avoid early dropout and retain patients, as well as space to reflect and adjust treatment courses.
Accounting services
Startups like Pilot, Bench and Taxfyle use technology to automate the rote and data-sorting processes required to accurately file thousands of tax returns and conduct proper bookkeeping.
In this vertical, AI is also used for error visibility, predictive insights and context-specific reporting. The time spent by accountants is drastically reduced, which allows them to focus on solving problems and supporting companies’ growth.
Here’s a snapshot of the human tasks TEHS startups are tech enhancing and their different levels of complexities.
Image Credits: Alon Laniado
Don’t over-automate
To scale rapidly and efficiently, many companies are tempted to optimize for automation. However, as the examples in the table above show, there are tasks that, at least today, can only be performed by people. In most cases, if these tasks are automated, they can comprise the quality of service.
In a more complete study on this topic, McKinsey & Co identified 56 human skills that can add value beyond what can be done by automated systems and intelligent machines. These range from structured problem-solving and creativity to motivating different personalities.
Example: Health coaching
Coaching has been proven to help people achieve life-changing goals provided evidence-based methods like motivational interviewing are applied. However, these techniques are usually high-touch in terms of coach/client dialogue.
For example, if someone has a goal of going to the gym twice a week after work, but ends up working late, a health coach would ask open-ended questions aimed at helping the client find their own solution. They might say, “So going after work really doesn’t work — that’s great to know. What’s another way to get in those two workouts?” These direct connections can then help the client talk themselves into change.
Despite the proven benefits of health coaching, it traditionally costs a lot, at up to $500 monthly. That explains why less than 2% of overweight Americans use a coach, even though 95% need the support. Poor access makes health coaching a prime candidate for being tech enhanced if it can streamline experts’ time.
You can’t apply automation to core coaching functions
Some coaching tasks cannot be automated, particularly the persuasive aspects of coaching. Also, conversational AI cannot effectively replicate all coaching dialogue required for behavioral change, such as addressing client setbacks. That’s because the nature of these conversations is wide-ranging and difficult to template.
This explains why the 350,000 automated digital health apps available now usually apply a fraction (less than 20%) of behavior change techniques used by health coaches. It also helps explain why 90% of their users usually churn in the first month and why most overweight people still can’t lose weight.
Fitmate Coach has staff coaches who deliver core behavioral support. At the same time, we streamline significant coaching time by cataloging repetitive dialogues and answers in FAQs.
The good news is that this coaching delivery model is likely to get better as AI technologies advance. That advancement could allow for increased automation of more behavioral support. But for this to happen, a sufficient amount of human-led conversations are required for machines to learn from.
Insufficient human involvement at the start, therefore, limits opportunities for more automation. That means that humans must be in the front seat of any company aspiring to automate more of the coaching process.
The future of TEHS
Human services enhanced by tech democratize delivery of service. More people everywhere will be able to utilize services previously believed to be beyond their reach.
The fascinating question facing TEHS is finding the fine balance between the person and automation. One thing seems certain: The human touch isn’t going anywhere anytime soon.