Scaling Internal Static Crowd by 1400% for LogMeIn

Growth pains: not scaling fast enough

“We were starting to get frustrated, just processing meeting data with our own team. We tried to take technical shortcuts or compromise the user experience. But we realized we needed to scale our ML and human computation efforts.”

LogMeIn is a $1bn provider of software as a service and cloud-based remote connectivity services for collaboration, IT management and customer engagement. One of the services they wanted to offer to their customers is Scopus.io, a meeting bot solution that would automatically summarize and share meeting notes with the attendees. They were trying to automatically understand and summarize natural language in their meetings to automatically take notes. Internally they spent most of their time processing the data themselves. As a result, they had to take technical shortcuts or compromise user experience. They realized they needed to scale their AI and human computation efforts.

Scaling with super.AI

"Our solution was missing the mark. We realized we needed to scale AI and human computation. Our technology wasn't there yet."

LogMeIn wanted to improve the output of their solution and reduce the team's time in output data verification before scaling the product to customers. Their concern was that by fully transitioning the product to AI, in order to achieve their desired turnaround times, they would have to sacrifice on the quality of the data output provided by the AI. At the same time, they knew a 100% human based solution would not significantly accelerate the scaling of the product compared to the throughput of their own internal team of reviewers. They reached out to super.AI because they know we provide a humans & AI integrated solution. Thanks to our humans in the process solution, they could scale with the right amount of Human intelligence needed. Thanks to our AI, they were also able to meet the latency requirements they had in mind.

Lastly, LogMeIn was excited about the ability to directly integrate the super.AI solution to their product via our API, which allowed them to get real time results.

super.AI took the data provided via the API and run it through our proprietary data program, processing the unstructured meeting recordings into structure meeting notes.

Results

Thanks to its guaranteed quality and speed deliverables, super.AI provided LogMeIn with elastically scalable human compute, which they could scale up or down as needed. The production API with a guaranteed latency was used to scale their production traffic by 1400%. Thanks to the API, they were also able to fully transition the product to the super.AI platform. This allowed the LogMeIn team to free up valuable time to develop new product features and prepare for the solution rollout to its customers.