Back to Blog

Top Natural Language Generation (NLG) Software in 2023

Reading Time 16 mins | November 28, 2022 | Written by: Robert Weißgraeber

Natural Language Generation (NLG), the automated content generation, is a fast-growing subfield of artificial intelligence. This is no surprise, as the demand for written and audio content is steadily increasing, and human-generated content can be costly and time consuming to produce. The NLG software optimizes this process and saves time and money by analyzing structured data and translating it into text, like product descriptions.

What is NLG software, how does it work and where can it be used?

Natural language generation software can transform information into language and hence produce written and spoken content.

Its importance to many types of businesses, therefore, should not be ignored. It can automatically generate informative content, such as reports and summaries, with no need for a human to perform the time-consuming tasks of analyzing data and writing the content. 

Using NLG technologies is most useful for companies that use structured data as the basis for their content.

These are typically used in e-commerce, finance and banking, pharmaceutical, insurance, or media and publishing industries. Natural language content generation, however, can benefit a wide range of companies outside of these industries, as well.

Find out more about how natural language generation softwares work 

Comparison of the 10 best NLG Softwares

This article presents the 10 best solutions for natural language generation. (All information on the respective providers is based on their own statements.)

  1. AX Semantics
  2. Arria
  3. ReadSpeaker
  4. by retresco
  5. Yseop
  6. BeyondWords
  7. IBM Watson
  8. Narrativa
  9. United Robots
  10. Wordsmith by Automated Insights

1. AX Semantics

The AX Semantics natural language generation software is based on a self-service principle with e-learning modules, onboarding and live support. The software is based on artificial intelligence and is 100% Saas (Software as a Service), i. e. the software is accessible for the customer and no programming skills or IT consulting are required. AX Semantics has placed great value on UI and UX, so the software’s interface is intuitive and clear.

Automated content creation can begin within 48 hours. In the process, content is being generated for various industries, such as e-commerce, business, finance, pharmaceuticals, and media. 

Noticeable analysis functions include the grammar UI for proofreading and editing and the unique content check. The generation of content is available in 110 languages and used by more than 500 customers, among them world-renowned brands such as Otto, Euronics and AX Semantics was ranked by Gartner as one of the world's top five natural language generation platform providers and by Forrester as one of the leading rising companies on the NLG market.

With the help of AX Semantics' software, content can be automated based on structured data, using natural language content that is impossible to distinguish from hand-written text. 


  • Quick start & user-friendly software
  • Content generation in 110 languages
  • No programming skills required
  • All content types based on structured data possible


  • No predefined content templates
Website - AX Semantics

2. Arria

The Arria NLG platform combines the most modern technologies of data analysis, artificial intelligence and computer linguistics. 

The software analyzes big data sets and writes reports automatically about what's happening in the data at tremendous speed and scale, without human intervention. Arria's NLG technology is scientifically based on over 30 years of research and development.

At the core of the Arria NLG platform is the Arria NLG engine, designed to generate natural language. At its core, the key to scaling and deploying at an exponential pace is Arria NLG-Studio, which is the NLG artificial intelligence tool for developers. 

Supplied by the Arria NLG platform, the AI software system automatically writes rich and engaging content using insights gathered from datasets. 

The Arria NLG system can write like a human. This is why the composed content is impossible to distinguish from hand-written texts.

Customers of the company are banks, financial service providers, insurance companies, pharmaceutical companies, consumer goods companies, and news and media agencies. 


  • Content extraction in written or spoken form
  • Automatically generated content can hardly be distinguished from handwritten content


  • The variation of automatically generated content is limited
  • Multilingualism is only possible to a limited extent
Website - Arria

3. ReadSpeaker 

ReadSpeaker is a text-to-speech (TTS) software that reads aloud segments of selected written content or entire pages. The company delivers natural-sounding voices with its own technologies. With more than 200 natural-sounding voices in more than 50 languages, and over 20 years of experience in speech analysis and processing, it has 10,000 customers worldwide up to this date.  

ReadSpeaker as a SaaS solution uses Deep Neural Network (DNN) technology for structurally enhancing voice quality, resulting in natural and engaging conversational experiences. 

ReadSpeaker gives voice to businesses for online, embedded, server or desktop requirements, apps, voice production, custom voices and much more. 

With the help of ReadSpeaker's speech solution, websites, mobile apps, digital books, e-learning tools and online documents are given their own voices.

As a result, new opportunities are opened up for people who have reading and writing difficulties, learning disabilities, impaired vision, or language problems. In addition, by offering important services in the respective native language of the users, there is a notable competitive advantage.


  • Written content is being read aloud by people with a natural-sounding voice
  • The software offers great relief for people with reading and writing difficulties, learning disabilities and impaired vision


  • The focus is exclusively on reading aloud content
Website - ReadSpeaker

4. by Retresco is Retresco’s self-service SaaS platform that turns data into automated content. Retresco was founded in 2008 and is a company which develops solutions that are based on artificial intelligence. These solutions originate from the fields of Natural Language Generation (NLG) and Natural Language Understanding (NLU). 

The platform is based on AI-powered Natural Language Generation (NLG) technology, enabling users to generate unique content. It provides intelligent linguistic analysis, an intuitive user interface, SEO creation, limitless scalability, and personalized onboarding support.

The software is able to generate multilingual content, which is used to easily reach new markets and target audiences without the need for expensive language services. It is possible to have text structures translated automatically and thus boost internationalization. It can be applied in e-commerce, finance, tourism, HR and agencies.


  • High-quality content outputs with scaling options
  • Creation of SEO content


  • Relatively strict content specifications, e.g. product descriptions 
  • Wide variance not as easy to implement
Website -

5. Yseop

Yseop is a natural language generation (NLG) technology that delivers solutions which leverage artificial intelligence to make sense of complex datasets and generate high-quality written content at scale

The company concentrates on a components-based approach and SaaS solutions. In doing so, the software offers solutions for the financial and pharmaceutical industry and the area of Business Intelligence (BI). In addition, there are special tools for complex medical and financial reports, as well as for sales automation. Yseop generates texts from data in seven languages: English, French, German, Spanish, Portuguese, Dutch and Japanese.


  • Multilingualism is guaranteed, although to a limited extent


  • Financial and medical industry focus - no approach to e-commerce content, for example
Website - Yseop

6. BeyondWords

BeyondWords instantly adds text-to-speech audio to web pages, and enables customers to engage with alternatively written content. By using the software, it is possible to personalize audio, using the technology to create a synthesized version of one's own voice or a person's voice that will resonate with the appropriate audience.

All audio content can be infinitely scaled by using SpeechKit editor, API, RSS extractor or WordPress and Ghost CMS integrations. With a dedicated audio ad manager and audio advertising integrations, audio content can be monetized via audio ads.

This software connects with Amazon Polly and Google WaveNet to provide a wide range of voices and languages. The generated audio can be published on platforms like Apple Music, Google Podcasts and Spotify.


  • Generation of personalized audio content in a matter of seconds
  • Scaling of audio content


  • Audio-only content can be generated
Website - BeyondWords

7. IBM Watson

IBM Watson Text to Speech is an API cloud service that generates human-like audio content out of written content. The service does this by offering customers a wide range of languages and voices in natural-sounding audio. 

Its underlying algorithm is able to interact with users in multiple languages and tones. IBM Watson enables customers to develop interactive products for call center interactions, hands-free communications and automated customer service interactions. 

A variety of additional services are offered by the software: Among other things, a self-service feature allows customers' questions to be answered on the phone by a virtual assistant. Moreover, AI-powered documents and intranet search can be used to support service agents in real time during phone calls.


  • Enhanced user experience
  • Significant case resolution potential 


  • The solution is a text-to-speech generation solution only
Website - IBM

8. Narrativa

Narrativa is a Natural Language Generation platform that generates and automates content for a wide range of domains. Among them are life sciences (pharma, biotech and healthcare), financial services (banking, insurance, investments and cryptocurrencies), marketing (CX, advertising, e-commerce, communications and sales) and entertainment (gaming, gambling & media).


  • Widely available for a wide range of potential content
  • Uses large language models for suggested texts 


  • No self-service, relatively long time from conception to publication
Website - Narrativa

9. United Robots

United Robots focuses on automating news at scale through AI and data science, providing a world-leading text-as-a-service solution for robotic journalism. The software can be used to automate sports coverage, real estate and business content, and daily reports including local weather forecasts and up-to-date traffic reports for any geographic location at any time in as many as seven languages, among others.

The software analyzes structured data, finding correlations and creating high-quality content at scale and speed. Then, an intelligent distribution platform is used to deliver the right text to the right audience at the right time.


  • Generation of high-quality content in large quantities and at high speed
  • Smart distribution platform 


  • Focus exclusively on automating messages
  • Automation in only seven languages
United Robots - Website

10. Wordsmith by Automated Insights

Wordsmith by Automated Insights is a self-service platform based on Natural Language Generation (NLG).

Users upload their own structured data. This can be done via an API using JSON. The software generates content based on a narrative design that is written in the company's tone and voice. The narrative design consists of conditional logic rules that trigger various outputs based on the data set behind the content. Content auto-generated with Wordsmith sounds as if a human wrote it manually. Consequently, NLG enables companies to scale content creation and provide written analytics. A unique feature is that texts can be updated in near real-time. 

Wordsmith has hundreds of customers in more than 50 industries, covering e-commerce, financial services and business. Among them are Yahoo Sports and Associated Press. Potential customers can take a demo to familiarize themselves with the software.


  • Real-time analysis are possible
  • Scaling of content
Website - Wordsmith

Overview of all NLG providers

  AX Semantics Arria ReadSpeaker by Retresco Yseop
Free Trial/Demo Free Trial & Demo Free Trial & Demo Free Trial & Demo Demo,
no free Trial
not specified
Industries E-commerce, finance, pharma, business, media & other industries Banking, financial services, insurance, pharmaceuticals, consumer products, news and media agencies.
Government, Education, Automotive, Healthcare, Voice Response Systems, Public Announcement and Emergency Notification Systems, Entertainment, Publishers
E-commerce, tourism, finance, HR, agencies Finance, Pharma, Sales
System 100 % SaaS 100 % SaaS 100 % SaaS 100 % SaaS
Component-based approach and SaaS
Multilingualism 110 languages not specified 50+ languages and 200+ voices English, German, Dutch, French
English, French, German, Spanish, Portuguese, Dutch and Japanese
e-learning, meetups, webinars
Documentation Documentation
not specified


  BevondWords IBM Watson United Robots Narrativa Wordsmith
Free Trial/Demo free basic version not specified no 30 minutes Demo Demo
Industries not specified
IT, finance, customer service,
healthcare, risk & compliance
sports, real estate, business, daily reports life sciences, financial
services, marketing, entertainment
50 industries
System SaaS SaaS SaaS SaaS 100% SaaS
AI voices from
64 languages
14 languages, different voices of men and women per language, dialects, 7 languages not specified not specified
Trainings Guide not specified not specified not specified not specified

Taking into account the latest developments, another tool should be mentioned here: ChatGPT.

11. ChatGPT

GPT-3 was used as a basis to build ChatGPT.
ChatGPT is an AI-chatbot. It’s designed to generate natural and human-like conversation responses in real-time. It can do so in many different languages.
ChatGPT was designed to give users quick, precise, and helpful answers to their questions. The user enters a question and receives an informative or entertaining answer. ChatGPT has been trained until early 2022. This means that it can only generate texts about events and developments up until this point in time and won’t be able to give coherent information about recent events.



  • wide variety of conversational tasks: from customer service to online dialogue recommendation
  • can be used to build virtual assistants and chatbots
  • understands and writes code in different programming languages
  • answers questions in easy-to-understand terms



  • can give incorrect or nonsensical answers
  • answers may vary depending on the way users formulate their questions
  • ChatGPT’s output needs to be fact-checked
  • trained only with information until early 2022

You want to learn more about the different types of Natural Language Generation and what text generator is the right one for you? Then read this article about the difference between GPT and data-to-text softwares.


NLG software providers will continue to play a significant role in the future. Because NLG software will be able to meet the growing demand for written and audio content.

All the mentioned top 10 solutions for Natural Language Generation have several things in common: The softwares work on the basis of science by processing speech, learning from data and creating written or spoken content that can be used by different companies and customers. In the overview, the individual software providers and their advantages and disadvantages are compared. With this, you can weigh which provider is the best fit for you and your business. 

FAQ Whitepaper Title

Open questions around automated content generation?

Download our free white paper that answers frequently asked questions around content automation and the AX software.

What is Natural Language Generation (NLG)?

Natural Language Generation (NLG) refers to the automated generation of natural language by a machine. As a part of computational linguistics, the generation of content is a special form of artificial intelligence. Natural language generation is used in many sectors and for many purposes, such as e-commerce, financial services, and pharmacy sector. It is seen to be most effective to automate repetitive and time-intensive writing tasks like product descriptions, reports or personalized content. 

When is Content Automation Worthwhile?

The use of content automation is always worthwhile when:
1. High quantities of similar content is needed (such as thousands of product descriptions, hotel descriptions, destination descriptions, etc.) or
2. similar content is required on a regular basis (e.g. financial reports, soccer reports, reports/forecasts, etc.). However, there is no standard answer to the question of when content automation is worthwhile. Most of our customers need at least 1,000 content pieces, and the possibilities are endless. Some of them generate more than one million texts per year. Due to our different license plans, we find a suitable package for each customer.


Who Can Benefit From Natural Language Generation?

Natural Language Generation (NLG) software has been increasingly used worldwide by companies and individuals across a variety of fields. In the process, the content to be generated can be aimed at any target groups. Particularly, online marketing experts with large online stores use NLG software to automatically generate product descriptions or content for category pages.

Robert Weißgraeber

Robert Weissgraeber is Co-CEO and CTO at AX Semantics, where he heads up product development and engineering and is a managing director. Robert is an in-demand speaker and an author on topics including agile software development and natural language generation (NLG) technologies, a member of the Forbes Technology Council. He was previously Chief Product Officer at aexea and studied Chemistry at the Johannes Gutenberg University and did a research stint at Cornell University.