Simpplr Writing assistant

Table of Contents


Intended use cases


Getting started

Modify or summarize text



Known issues

Best practices



Writing content can be a lengthy task, and for many content creators, it's added work that's not part of their normal scope. In order to help with this, Simpplr has introduced an AI-powered smart writing tool for content creators called Writing assistant.

As part of the 23.09 release, Content managers will now be able to generate "smart content" based on prompts entered. Users can decide the tone, format, and intent of the content, as well as tie any applicable company values to the messaging. 

Intended use cases

The Writing assistant is intended to cut down on time it takes to write and approve content. Often, our customers will publish long-form content that takes time to go through the proper writing and review stages. This feature allows creators to simply enter values for how they want their content to be read, and in a matter of seconds, the smart content is output. You'll then have the ability to edit the output however you'd like. 


The Writing assistant can be controlled by the App manager in Manage > Application > Privileges. Scroll down until you see the Generative AI section. Here you can decide to have the feature (along with Smart answers) on or off for your users. Note that once the feature is enabled, any user who has access to publish content or submit content for approval will be able to use Writing assistant for their content.
Writing assistant enablement.gif

Getting started

To get started:

  1. Add a new piece of content as always.
  2. From the text/content editor in the body of your content, find the magic wand icon and open Writing assistant. Here you'll have the option to:
    1. Generate new text: This will be the only option with a blank page, as there's been nothing written yet. 
    2. Modify all selected text: This opens the option to have the AI modify or summarize only the text you select, while leaving the rest of the text alone (more info below).
    3. Modify all text: This opens the option to modify or summarize all text in the content body, including your manually input text (more info below).
      Writing assistant new location.png
  3. Enter the prompt you'd like the assistant to expand upon. This can be anything as simple as, "Make an announcement for the new lunch schedule. Lunch will be pizza on Mondays, tacos on Tuesdays, wings on Wednesdays, salads on Thursday and BBQ on Friday." Or you can get granular with in-depth details around the content you'd like to publish. Best practice here is to enter a prompt, generate the text, then edit and articulate your prompt as needed to get the message you'd like if you're not receiving the desired outcome. Remember, you can also always edit the text after it's been generated to add any information you want.
  4. Next, scroll down and choose values for each of the dropdowns.
    1. Company values: Tie your applicable company values to the messaging.
    2. Tone: This will indicate the tone of the message. Depending on the context, this is likely to change from content to content. For example, updated benefits information will probably be best written with a formal tone, while a new lunch announcement can be more casual. 
    3. Format: Choose how you'd like your content to be presented. Again, this will change depending on the content. Long-form information is likely going to work best in Article or Paragraph format, while a list of new team members could work best as a Bullet list.
    4. Intent: This is where you establish the content's purpose. Let's say your org is undergoing a restructure, and you know some employees aren't happy about the recent changes. Maybe your content is meant to convince teams that these decisions are good for the company. If so, choose Convince. If you choose Educate, the content will be written in such a way that users will come away with information that's meant to be retained; whereas Inform will adjust the content to simply inform users of something happening. Best practice here is to try different variations of your content and see which output works best for you and your org.
  5. When ready, click Generate. The image below shows the first generated output from the prompt above:
    Smart writing prompt output 1.png
  6. Once the text has been generated, you'll have the option to read it, then improve it by adding a clarification to the initial prompt. In How would you like to improve this text?, you can include and challenges related to the topic, customize the message for certain employee types (remote, for example), or just any clarifying information that was left out of the original prompt. You can also just regenerate the content if you don't like the first output. To do so, just click the refresh icon.
    Regernerate smart writing.png
  7. Note that you can head back to a previous version at any time with the Previous version button at the top. Similarly, if you want to use a regenerated version, once you create more than one output, you'll have a Next version button.
  8. The new output will have some added information:
    Smart writing improvement 1.png
  9. Finally, note that you can make the generated information longer or shorter without having to add more of your own prompts to it by simply choosing Make longer/shorter. This will communicate the same message, just in more or fewer words as you'd like.
    Make longer or shorter smart writing.png
  10. When satisfied with the output, click Add to draft. The content will be added to your page, where you can edit it accordingly. Formatting and editing content will work the same as it always does with non-AI generated content.

Modify or summarize text

Once there's any form of text input to your content body, you can have the Writing assistant summarize it; either only the AI-generated text, or all text, regardless of whether you manually input it or if the AI generated it. From the magic wand icon, click either Modify selected text or Modify all text. This will open the modal that allows you to regenerate whatever selected text you want. Then there's a new button that allows you to summarize the text.
Smart writing modify text.png

Here you have three options for a summary:

  • TLDR (Too long, didn't read): This will provide a brief summary of everything contained in the content. Often this is helpful to place at the top of the content with the heading "TLDR", especially if the content is longer than usual.
    Smart writing summarize TLDR.png
  • Conclusion: This is a great way to conclude all the points/highlights of your content. We recommend using this for more story-based content.
    Smart writing modify text conclusion.png
  • Key takeaways: With this option, the AI will highlight the key points and facts made throughout your content in a listed format.
    Smart writing summarize key takeaways.png

Results with any of these summaries will differ greatly depending on the content subject, audience, and all the factors like tone, format, and intent. Experiment with different results to determine what will be the best fit for your team and audience.


As is the case with many AI features recently, the big question is around security. How can you be sure what you're writing is not being stored or shared somewhere without your explicit consent?

The Writing assistant tool is built on Simpplr’s secure platform (click this link for more on Simpplr Security Compliance). Additionally, Writing assistant utilizes LLMs through Microsoft’s Azure platform. We have chosen to utilize Azure because they provide a higher level of security and trust, and more control over the data that we pass to them. This includes any input and output to both Simpplr and Azure’s platform not being available in any other place than our private, secure instance, and is not accessible to others. Additionally, the access to this Simpplr resource is controlled and restricted.

What's happening on the backend? 

  • The feature is contained and controlled entirely by Microsoft Azure on Simpplr’s private instance. This means that:
    • No data is available to others
    • No data goes to OpenAI or any services operated by OpenAI (i.e., ChatGPT or OpenAI API)
    • No data is used to train any public models
    • No data is used to improve any other Microsoft or other 3rd party products

Click here for more information on the Azure OpenAI service.

Since Writing assistant is operated on Simpplr’s private Azure instance, this means that only Simpplr has access to this instance, and no data from it can be used for any other purpose. Additionally, the inputs and outputs from your organization cannot be used for training a model that would be used for other organizations within the Simpplr instance. Additionally, none of the inputs or outputs from your organization are stored in the model itself or used to train it.

Data privacy and retention

All data is covered under DPA, separate instance, is GDPR and CCPA/CPRA compliant. No personal data is transferred outside of the above noted areas, and none is used for anything other than the output for your input. All other standard Simpplr data privacy and security considerations apply to data kept by Simpplr through this process, and are not shared with or used for any other purpose.

Security certifications

Simpplr has received the following security certifications:

  • ISO 27001:2013 certification
  • SOC 2 Type 2 attestation (available under NDA)
  • SOC 3 report (pending)
  • EU-U.S. and Swiss-U.S. Data Privacy Framework (DPF) certification 
  • TRUSTe Data Privacy Framework (DPF) verification
  • TRUSTe Privacy Dispute Resolution program


Q: Are all users allowed to publish content with Writing assistant?

A: Any user who has access to publish content or submit content for approval will be able to use Writing assistant for their content.

Q: Is Writing assistant compatible with all supported languages? 

A: While you can technically prompt the AI to write and publish in other languages, it is not very intuitive. We are working on an enhancement this next release that will:

  • Keep the user's selected language in mind
  • Give them the option to select a specific output language (i.e. I am comfortable writing in English, but once I'm done checking my prompts I know I need to publish in Spanish)

Q: How does Simpplr store the AI generated content? Is the information sent to any 3rd party vendor?

A: No. All content generated is locally hosted on Simpplr's own instance of Microsoft Azure. Our AI modules are not shared outside and are used internally only. We joined the Responsible AI association to ensure and confirm we are not using AI for anything but the benefit of the customer itself as it is not shared nor abused. See the Security information above for more detail.

Q: Am I able to edit already published content with the Writing assistant?

A: Any published content that was published as of the 23.06 release or later will be able to be regenerated with the Writing assistant. For older content, the previous content editor is still in use, and the Writing assistant is not applicable to that content. 

Q: Is generative AI exclusively used in the Writing assistant feature? If not, are there other features within Simpplr that leverage Generative AI?

A: Currently generative AI is used in the Writing assistant as well as the Smart answers feature, which shows a natural language, most-likely answer at the top of the search results.

Q: Has a penetration test been conducted on this feature before full release? If not, when will it be included in a future penetration test?

A: No. The feature has been through our SDLC, security reviews, automated security testing, and internal QA. Penetration tests are conducted on the complete application quarterly. The next test will be in December/January timeframe.

Q: Has the feature been tested to prevent prompt inject attacks or tested for potential harmful output?

A: Yes. The system includes guardrails to reject suggestions to generate potentially harmful output and to protect it against prompt injection attacks. However, in the same way that guardrails can improve but not guarantee motorist safety, no AI system is completely invulnerable to prompt injection. That said, we are continually testing and improving our product to ensure quality.

Q: Will my company own all the content generated by the AI Writing assistant?

A: Yes.

Q: It was noted that data is not used to train the model. Please clarify on which data sets that comprise the LLM uses.

A: Azure OpenAI GPT Turbo 3.5. Simpplr uses the Microsoft Azure OpenAI platform. Simpplr AI features operate entirely within Simpplr’s private instance on Azure. We are working on some enhancements and will be testing them with the newer models as well. We will plan to upgrade to a newer model once we are satisfied that it improves the quality, which will take some testing. This will be the pattern going forward for all new releases of models.

Q: How long does Azure store our data (input and output)? Do customers/users have the ability to set a data retention timeframe? Who has access to the data that is stored?

A: Azure is storing it for 30 days for abuse monitoring only, and yes, we can ask to lower it. We are storing it as long as their contract states data can be stored currently on our end.

Q: Does Simpplr have access to the data or use the model for any product/service improvement for itself or other customers?
We have access to the data but do not use it for improvement for itself or other customers.

Q: Is this an Open AI instance just for my company/instance or does Simpplr use the same Open AI instance for all customers?

A: We are not using OpenAI; it is Azure, and is a Simpplr private instance for all Simpplr customers, but no data is exchanged between customers.

Q: Is there confirmation that the data is encrypted in transit and at rest (inc. to/from Open AI)?
A: Yes

Q: Does Simpplr save backups of the model in the event of corruption in case my org needs to “roll back” to a prior version?

A: We are not currently updating versions. We will let customers know if and when we do.

Q: Does Simpplr take any steps to prevent the Writing Assistant tool from generating unlawful content or content that infringes the rights of third parties?

A: We run moderation checks for inappropriate content and will be updating the model shortly for copyright best practices.

Known issues

  1. If your org has not set up Company values, the default values will appear in the Writing assistant. You must add your own company's values, or disable the Company values feature to make the default ones not appear.

Best practices

Creating effective internal communications content using Writing assistant requires careful consideration of content, tone, format, and intent. Here are some best practices to keep in mind:

Importance of content:

  1. Clarity and relevance: Ensure the content being generated and/or written is clear, concise, and directly relevant to the recipients. Avoid jargon or overly technical language that might confuse employees. Just because you have the option to make your content "formal" doesn't mean it should be overly complex.

  2. Timeliness: Deliver information promptly, especially for time-sensitive matters such as policy updates or project deadlines. Don't spend too much time debating on the best tones, formats, etc. when you have a clear message you want communicated. If the message isn't clear, or you don't fully understand what you want said yet, hold off on using Writing assistant until the message becomes clear. 

  3. Accuracy: Verify the accuracy of the information before publishing. Incorrect information can lead to confusion and misunderstandings among employees. The Writing assistant tool is not perfect; no AI is. It is still up to oyu as the owner or author of the content to ensure accuracy.

Examples of Tone:

Choosing the appropriate tone is crucial for effective communication. Here are examples of different tones and when to use them:


"Dear Team, I hope this message finds you well. We are pleased to announce the launch of the new project management system..."

Use formal tone for official announcements, policy changes, and important updates.


"Hey everyone, exciting news – we just hit a major milestone in our sales target! Kudos to the entire team for your hard work and dedication..."

Use a casual tone for positive announcements, team achievements, and casual updates.


"Attention all employees, effective immediately, the company's work-from-home policy has been revised. Please review the updated guidelines attached to this email..."

Use a direct tone for urgent matters, policy enforcement, or critical updates.

When choosing the Format of your generated content, consider:

  1. Consistent structure: Maintain a consistent format for different types of communication. This could include using headers, bullet points, and numbered lists for easy readability.

  2. Visual elements: Use visuals like charts, graphs, and images where appropriate. Visuals can enhance understanding and engagement, especially for complex information.

  3. Make it mobile-friendly: If your org is using the Simpplr mobile app, ensure the content is optimized for viewing on various devices, including smartphones and tablets, as many employees will access the intranet and your content on mobile devices.

What makes good Intent?

This goes back to what we mentioned earlier about understanding what exactly you want to communicate. Without a clear and concise purpose, intent won't matter. Consider these points when choosing your intent:

  1. Clear purpose: Clearly state the purpose of the communication. Employees should understand what is expected of them or what action, if any, they need to take after reading the message.

  2. Engagement: Encourage engagement by asking for feedback, opinions, or suggestions. This could be in the form of surveys, polls, or open-ended questions. If your intent is to educate, this is a great way to incentivize engagement. You may even want to use Simpplr's Recognition Suite or Employee Listening add-ons if you're not already. 

  3. Always try to reinforce company culture: Align the content with the company's values and culture. Use language that resonates with the organization's ethos to strengthen the sense of belonging among employees. This is a great opportunity to tie in your Company values to the content.

  4. Encourage collaboration: Use communication to foster collaboration among teams. Highlight successful collaborative efforts and encourage cross-departmental discussions. You don't have to be the only communicator here. Cross-team collaboration can serve to increase company culture. 

In summary, effective communications, either written by humans or AI, should be clear, appropriately toned, well-formatted, and aligned with the company's goals and values. Tailoring the tone to the specific message and being mindful of the intent can significantly enhance the impact of the communication on employees.


Was this article helpful?
1 out of 1 found this helpful
Have more questions? Submit a request


  • Am I able to add Company Values even if we do not subscribe to the Recognition module?

    Comment actions Permalink
  • Hi Maureen, yes. Company values can be selected in the Writing assistant even if you're not using the Recognition suite. 

    Comment actions Permalink
  • Hello, do we have to pay to use that?

    Comment actions Permalink
  • Hi Karim. No, Writing assistant is a free feature. It just needs to be enabled in Manage > Application > Privileges by an App manager.

    Comment actions Permalink

Please sign in to leave a comment.

Articles in this section

See more