Many Nuance customers are looking for an alternative in the wake of Microsoft’s acquisition of the longtime speech leader.
Some are trying to get ahead of what they view as an inevitable issue with migrating to Microsoft’s Azure cloud. Others have been dissatisfied with Nuance’s round-up pricing or expensive professional services and are finally pursuing an alternative in the wake of the change in ownership.
Switching ASR providers, and migrating your data, is not something to take lightly. While there are tools that make it easier than you might think, it’s certainly not something you’d want to undertake multiple times.
So if you are thinking about migrating off Nuance, you’ll want to do a careful comparison of the capabilities of alternative providers.
For this post, we compared Nuance to three major speech recognition alternatives – Google Cloud, Amazon Transcribe, and LumenVox – across several key categories, including accuracy, price, and support.
For a deeper dive into the landscape of Big Tech ASR providers, download our eBook.
Nuance vs. Google Cloud
Google Cloud’s ASR is very comparable to Nuance and may be a solid alternative if you prefer Google’s cloud to Microsoft’s Azure.
Nuance has slightly more supported languages with 86 compared to Google’s 75, but the difference is small enough that it likely won’t make much of a difference unless you are doing business in one of the few markets with Nuance language support that Google lacks.
Although with Google Cloud you will still pay for rounding, it’s only up to the nearest 15 seconds, as opposed to Nuance where you are paying for rounding up to the nearest minute. That can make a big difference at scale, especially if cost is one reason you are considering leaving Nuance.
When comparing transcription accuracy using a challenge dataset of short phrases, Google Cloud was a step down with 73.9% accuracy compared to 86.3% for Nuance.
Overall, Google is very similar to Nuance in most facets. If you are dissatisfied with Nuance, Google might be an easy transition, but it also may not solve all of the problems that led you to want to switch providers in the first place. Specifically, Google Cloud has limited support for grammar ASR, and that could be a hurdle for enterprises that have heavy investments in grammars.
Nuance vs. Amazon Transcribe
If you are using Nuance for medical transcription, Amazon Transcribe may offer an alternative.
Amazon Transcribe’s base language models include general and medical, so it should be able to handle many of the technical terms that have made Nuance a leader for medical ASRs.
Amazon Transcribe has slightly better speaker diarization, with up to 10 speakers compared to Nuance’s 6, so it could be a fit in environments where you have many voices.
Amazon Transcribe will be slightly cheaper than Nuance. Amazon rounds up to the nearest 15 second increment as opposed to Nuance, which bills you for the full minute.
Unfortunately, you will be sacrificing accuracy with Amazon Transcribe. At 71%, Amazon Transcribe had the worst accuracy of the ASR platforms we reviewed when using the challenge phrases dataset, compared with 86.3% for Nuance.
Overall, Amazon Transcribe could be a workable alternative to Nuance, especially if your business is in the medical field.
Nuance vs. LumenVox
If you are concerned about Microsoft owning Nuance – and the possibility of being forced to migrate to a proprietary cloud – neither Amazon nor Google will solve your problem.
In both cases, you are changing out one Big Tech provider for another, and each have proprietary clouds that will limit your ability to use a competitor’s solution.
LumenVox is a rarity insofar as it is an independent, speech-focused company with a long history of innovation in an industry that has increasingly been dominated by Big Tech.
In our eBook, we discuss some of the problems common across all Big Tech providers. That includes losing control of your data, as well as partnering with an 800 lbs. gorilla who likely competes with you for business.
Big Tech also mandates one-size-fits all pricing models (no flexible licensing) which creates a problem with operating expenses as you scale.
Another aspect to consider is the implementation process. You won’t get much support if your unique situation isn’t covered by an FAQ. LumenVox stands out for having the best support and the most detailed resources for our partners and clients.
LumenVox also stands out in terms of accuracy across ASR providers. Thanks to an end-to-end, deep neural network, LumenVox achieves a 98.7% accuracy for phrases using grammar ASR, as compared to Nuance’s 86.3% (the second best next to LumenVox).
LumenVox is also roughly one third the price of Nuance. A big contributor to that savings is that Nuance requires professional services for implementation, which can cost as much or more than the actual software.
The other source of ongoing cost savings is that LumenVox charges only for the time you use and does not round up pricing. If you are used to paying up to the nearest minute for Nuance, you’ll notice the reduction in operating expenses at scale.
LumenVox makes it easy to migrate from Nuance. We have tools for grammar migration and tools to ensure that the accuracy and performance of your speech applications is exactly where you want it to be.
Which ASR provider is best for you
There are many factors to consider when choosing your ASR provider. No one blog post can cover all the unique circumstances of your business.
You’ll weight capabilities differently based on whether accuracy, cost, or ease-of-use is more important to your business. The volume of calls you’re handling as well as the pricing model you use with your clients may disqualify some of the Big Tech vendors from consideration.