Chatbot for Health Clinic
- Custom Software Development
- $10,000 to $49,999
- Feb. - June 2019
- Quality
- 5.0
- Schedule
- 4.0
- Cost
- 5.0
- Willing to Refer
- 5.0
"They understood what needed to be corrected pretty quickly."
- Other industries
- Belgium
- 51-200 Employees
- Phone Interview
- Verified
Botodata built a machine learning solution providing medical recommendations based on patient response inputs. They reviewed raw patient data and created a functional algorithm based thereupon.
They delivered a proof of concept suitable for future mobile app development. They algorithm calculates recommendations correctly at over a 90% rate, which has improved internal resource efficiency. Botodata is a proactive team with a good technical understanding of their projects.
A Clutch analyst personally interviewed this client over the phone. Below is an edited transcript.
BACKGROUND
Introduce your business and what you do there.
I’m the CEO of a medium-sized, privately-held clinic. We have 50–55 health staff onboard.
OPPORTUNITY / CHALLENGE
What challenge were you trying to address with Botodata?
The primary challenge was to digitize our clinical pathway. We’d been doing it manually for the past five years, and it was getting cumbersome and outdated. The idea was to try to get the patients to do all of that online, without the need to physically go to the doctor.
We decided to develop a chatbot with a questionnaire that would normally have been done on location. The idea was to digitize that while keeping a comparable level of precision in terms of the answers given. We would offer the patients recommendations and potential prescriptions based on those answers.
SOLUTION
What was the scope of their involvement?
We had a considerable amount of written data from countless patient consultations, and we needed Botodata to process this raw data. There were patterns based on patient responses, along with the resulting recommendations. Botodata needed to review all of those and input each one into the system and create an algorithm based on those responses and the resulting recommendations.
They had a huge amount of raw data to process and then create an algorithm that would process entries in the future without any involvement from our side or theirs. It’s basically a learning system that takes the replies into account and spits out a result. The replies are within certain constraints. We can’t have random results that break the logic. The system aggregates all the replies.
They’ve completed the main chunk of work on the backend. We’re now looking for an app developer to implement it into a mobile app.
What is the team composition?
The bulk of my communication was with their main chatbot developer. They had one other team member who was pretty helpful with recommendations and input. I mainly worked with those two people, and I don’t think there was a project manager involved on Botodata’s side. I don’t think they had a sales manager, either.
How did you come to work with Botodata?
We did a website project with another company, and they recommended paying attention to Botodata if we wanted to build a chatbot and digitize the clinical pathway we had. They didn’t have the expertise in machine learning and big data to do it, so they recommended this new team. We were pretty hesitant at first, because they were a relatively new team, but the prices were decent, and the overall presentation during our first call was pretty good. They’re more of an enthusiastic, “we’ll do this all ourselves” type of company, which was another reason why I found them interesting.
How much have you invested with them?
I think we ended up paying $33,000–$35,000, which was over the initial estimate. I knew that the initial estimate can be changed or adjusted, and we did ask for additional features, so I have no gripes with that. The guys did their job well, and we paid what we paid.
What is the status of this engagement?
We worked with Botodata between February–June 2019.
RESULTS & FEEDBACK
What evidence can you share that demonstrates the impact of the engagement?
It’s hard to give feedback without us actually being able to physically interact with it. Right now, it’s a proof of concept with a very simple frontend site where we can input replies to questions and it gives us recommendations. It’s made purely for us to see that it actually works, but the end goal is to have it on mobile. We’re still trying to find a good mobile company that will implement that.
Based on what we can see from the frontend side that they have connected to the backend, it does indeed calculate what we need correctly at over 90% accuracy. At least from this perspective, it’s a very good indicator that it does indeed do some calculations. We have to review them sometimes, but the main idea is that we no longer have to do them manually all the time.
How did Botodata perform from a project management standpoint?
It wasn’t anything spectacular or out of the ordinary, but it wasn’t bad. They were okay at managing the project and got the job done. No out of the ordinary issues happened, and there were no problems. Everything was done more or less on time.
There wasn’t a need for a dedicated project manager, and I didn’t request one from my side. The developer notified me if any important issues arose. I couldn’t care less if there was a dedicated project manager, or if it was the developer himself, as long as the information was properly relayed, and in time.
What did you find most impressive about them?
They seem to be pretty proactive. I liked that if there were any potential shortcomings when I checked the results of the recommendation the solution provided, they reacted quickly, and I didn’t have to explain the same thing several times. I’d experienced this before with other companies. Those other companies didn’t bill me for the time spent explaining things over and over, but I still was wasting my time with them.
With Botodata, the level of English was pretty good, so I didn’t have to explain a lot. They understood what needed to be corrected pretty quickly. Even if it was above the estimation provided initially, they took it up upon themselves. When they screwed up, they admitted that they screwed up, and took responsibility for it.
Are there any areas they could improve?
It would be following the timeline more closely. I have no qualms with the quality. Even if some features weren’t right at the start, they made the corrections, and it’s now up to the standard.
Timelines weren’t an issue for me, but I can see how a company in the corporate sector might find that Botodata doesn’t behave professionally or in a very business-like manner. They practice more of a chilled-out atmosphere. Of course, I didn’t mind that. I even appreciated their friendly atmosphere at times.
Do you have any advice for future clients of theirs?
Clients shouldn’t be scared by their relaxed and very friendly attitude. They tried to befriend us right from the start and treated us not as colleagues, but as someone they’d known for a long time. I believe they even position themselves as enthusiasts that don’t want to look overly professional.
Approach them with this in mind and treat them with respect, and they will perform to your expectations. And don’t be afraid to point out their mistakes or delays, if there are any. They will admit it if anything is wrong.
RATINGS
-
Quality
5.0Service & Deliverables
"There were some issues, but they corrected them all."
-
Schedule
4.0On time / deadlines
-
Cost
5.0Value / within estimates
"It was one of the reasons I went with them, and the cost reflects the team."
-
Willing to Refer
5.0NPS
"I’m very likely to recommend them, with the note to keep in touch with Botodata with regard to the schedule.