There are a lot of AI-related technologies out there. I mean, really a lot. If you happen to glimpse Matt Turck’s annual AI and Data Landscape, you might even feel a bit overwhelmed by just how diversified and crowded that field is! Though one sub-field seems to stand out: Natural Language Technologies. All recent studies point to the direction that natural language technologies will drive the next generation of AI solutions. Even before the pandemic and the acceleration of digital adoption, Natural Language global market was expected to grow at a yearly 20% rate to reach USD 35 billion in 20262. And it won’t be slowing down, as each breakthrough in this field easily translates into market-ready solutions, and business outcomes can efficiently be measured.
Natural Language technologies can be classified into 3 main categories – NLP, NLU and NLG
Those 3 complementary technologies are often used together to create human-machines interactions or analyze how humans communicate with each other. For the rest of this article, we’ll simply refer to them as NLP technologies.
So how do companies use these technologies? How do they create successful business models and products that delight their customers? Let’s deep dive!
A closer look at the NLP market makes us outline 4 different types of organizations, each having a different focus to build their product and solutions:
Here we consider organizations focusing on developing new language models. Those pre-trained models are often open-source and for the most part publicly available. The two most popular are BERT, developed by Google, and GPT-3, published by OpenAI. Those companies are on the scientific and R&D end of the NLP market. They aim to push the boundaries of NLP capabilities before anything else. There are no clear business models associated with this approach.
NLP PaaS companies develop and monetize NLP models. They create frameworks for developers to build NLP-based solutions. Unlike NLP Model organizations, NLP PaaS have a clear for-profit positioning. They target AI and developers’ teams of large corporates to provide them with the best NLP tools possible. Two examples quite define the Platform-as-a-Service business model: Hugging Face and Rasa.
Hugging Face solutions are for any company that want to “build, train and deploy state of the art models powered by the reference open source in natural language processing”. Hugging Face provides tools that cover most natural language technologies, unlike Rasa, that chooses to focus on conversational AI models. Also based on open source, Rasa’s mission is to let AI teams build best-in-class human-machines conversations and customer experience.
Like those two companies, NLP PaaS put most of their resources into creating an NLP framework. They need to always be at the forefront of natural language technologies while deeply understanding a developer’s needs and pain points. Thus, most of these companies base their products on a strong open-source library and a large community of developers, that acts as technical buffer as well as a sales pipeline.
An NLP SaaS company uses natural language technologies as part of a larger set of solutions to address the needs of either a business function or a specific industry.
Let’s take Gong, the revenue intelligence platform for sales teams. Gong leverages NLP to optimize sales departments. Their technology screens all customer interactions to provide teams with insights on their pipeline and deals, understanding of staff performance and market intelligence. To achieve this, the company needs to first assess how a sales team usually operates and how their workflow can be streamlined. NLP technology only comes as a support of the solution, but the main effort is on creating a seamless experience for sales staff. Other companies replicate Gong’s approach for various departments (HR, Finance, Customer service, marketing…)
Thus, unlike NLP PaaS, most technical resources of an NLP SaaS go into software development and analytics with a strong focus on UI/UX. They sometimes use NLP PaaS models that they fine tune to their needs.
Other companies use NLP technologies to serve the needs of a specific industry. Linguamatics is doing just that for healthcare. The company has developed a Natural Language AI platform that uncovers valuable information through text mining. It applies its solutions to a large range of situations in life science, from supporting the drug discovery process to understanding the voice of patients. What matters for such a company is to be recognized as an industry expert, be able to understand every stakeholder’s challenge and offer relevant solutions. Their NLP models are trained with industry-specific data to gain competitive advantage over other technology providers.
This category comprises all consumer-facing NLP products. That includes smart home speakers and personal assistants, dictation tools, chatbots, text predictions etc… Those products are more often considered as a feature of another service and are mostly provided free of charge for unlimited usage. They act as enabler rather than as a stand-alone solution. The value of those products resides in the data collected.
Now, we’ve just seen 4 different approaches to building a business on NLP technologies. But what kind of problems do these companies solve? How do they use NLP to address pressing business issues? Let’s now delve into 4 different organizations that shed some light on how natural language technologies apply to business operations.
“How to improve customer experience?”: analyze what customer say, think and feel.
ASAPP is a company that uses AI and natural language technologies to help companies create better customer experiences. It has developed an NLP-based platform for customer service teams. It analyzes customer calls and interactions, transcripts and summarizes the calls, understands and classifies customer intent. This screening automatically transforms unstructured communications into valuable data.
When applied to the customer journey, NLP helps companies collect insights on customer interactions and better understand how they feel at each touchpoint. The outcomes are numerous: increase customer satisfaction, reduce churn, leverage upselling opportunities, or understand your staff performance… Because customer service is, in-itself, driven by natural language interactions, the benefits of those technologies are outstanding.
“How to better understand the trends shaping my market?”
Dataminr is a real-time event and risk detection platform. It screens social media platforms such as Tweeter to detect in real-time the trends that might have an impact on their customers’ business. Unlocking this kind of information is very valuable in industries that are highly sensitive to public, political and social events (finance, public sector, media, defense…).
Natural language technologies can help analyze market trends and competitive landscape, providing competitive advantage and better risk management.
“How to improve the efficiency of my organization?”
Eigen Technologies uses NLP to analyze written documents of financial institutions and turn them into actionable data. They aim at unlocking valuable information that is hidden within numerous documents in the form of natural language. Every company has bottlenecks in their operations. That is true not only in finance but in every industry. Information doesn’t flow seamlessly across all teams. Instead, employees often loose time and energy trying to recollect valuable pieces of information and improve data management systems. By making sense out of communications and documents within an organization, NLP can help businesses streamline their operations.
“How to put the right product in front of the right customer?”
Albert offers various AI marketing solutions. Some of which use NLP to optimize digital campaigns efficiency. NLP can be applied to understand how a target audience respond to specific campaigns. Combined with other AI technologies, it can advise and automate campaigns adjustments. With NLP, it is possible to better understand how an audience feels and what their buying decision process is. Albert, like other NLP-based marketing solutions, let their customers do a better targeting and drive additional revenue.
Those are only a few examples of how natural language technologies can benefit organizations. If you are looking to make sense of this crowded field, it is worth taking the time to assess your current situation and the exact pain points you are trying to solve. Do not try to implement NLP everywhere but rather start applying it to a specific use case. With everything that’s out there, you will surely find the right solution for you, or the tools to build it yourself.
Aiello has extensive expertise in Natural Language Technologies across various industries. We have developed an NLU SaaS Platform that provides organizations with a complete framework to build NLU based applications. We help our customers unlock the potential of NLP with customized solutions that fit their needs.
Do you want to learn more about what we’re building? Contact us to schedule a free demo now!