Show your colleagues:
AI Workflow: Press Release Generation
Automatically generate press releases with company-specific preferences.
- 8:18min

Goal:
Create a press release draft for a new company announcement, optimized for your company.
Description:
Every company has a specific way of communication – custom wording preferences and branding guidelines that decide whether a text is good in the eyes of the company, or rather not.
This flow allows you to save all of your company’s custom communication rules and considers them, when announcing new milestones and company updates.
Benefits:
Actively guide and control your AI outputs.
Automatic evaluation and correction of your criteria.
Easily refine your results.
Analyze your press release draft & refine it until it is great.
Simlpy duplicate your previous configuration and quickly adjust last details to find the settings that work for you.
- 5:06min

Create an AI Template for your repetitive Press Release Tasks.
Your company-specific communication preferences will not change much over time. Build yourself AI templates to get your next AI-based press release even faster.
- 3:58min

Frequently Asked Questions
What other users already wanted to know.
About this flow:
This flow uses the translation capabilities of the chosen underlying AI model. Currently the default model is GPT4o, which officially supports 58 languages, but its proficiency extends to over 95 languages. Most widely spoken languages like English, Spanish, French, German, Italian, Dutch, Portuguese, Chinese or Japanese will work fine.
Company-specific translation preferences can easily be used in this press release generation flow. Within the flow configuration, you will be able to insert information about the desired details. For repeated usage you can create an AI template, so that consistent preferences are saved for the next translation task.
Within the flow configuration it is possible to name and explain all company-specific terminology, so that your generated text can optimize all such terms, without manual corrections.
If a press release output is not yet satisfactory, you can simply duplicate the flow configuration and adjust some of your inputs to slightly improve the output into the desired direction.
Take a look at the video on “Refining your output until it is great” (second video on this page).
Yes. ValueFlow offers affordable and monthly cancelable subscriptions for its AI studio (of only 25€). With this subscription you get access to this translation flow, but also many more useful flows and AI capabilities.
Yes. There is a 7-day free trial. Simply click on “Start your trial” on the top right. Also, ValueFlow additionally offers a 30-day money-back guarantee. You can therefore try this flow risk-free.
About ValueFlow:
No. You do not need any technical skills for using ValueFlow and this translation flow. ValueFlow is building functionalities like this press release generation flow for you, so that you can get an easy entry into the topic of AI.
We recommend purchasing a managed seat which only costs 25€ per month ($25,94 USD). If you choose a yearly license then you get 3 months for free, every year. You not only get access to this press release generation flow but to many more flow and the complete ValueFlow AI Studio.
Yes. There is a 7-day free trial. Simply click on “Start your trial” on the top right.
If you would like a personal demo, then click on “Book a demo”. A member of our sales team will be happy to show you ValueFlow in detail.
Yes, ValueFlow offers a 30-day money-back guarantee. If you are not convinced about the software, then you will get your money back – no questions asked.
Yes, you can change your plan each month. The payment provider Stripe will take care of it.
No, there are no hidden fees like AI models costs. With your managed subscription you purchase the right for unlimited but fair usage of ValueFlow and respective AI generations.
Yes, if you choose the yearly plan, then ValueFlow will be happy to give you three months for free – every year.
It is secure. ValueFlow uses Microsoft’s servers that fulfill enterprise-level security standards. For the business package, every organization receives its own server instance, which completely separates their server and database from other customers. Furthermore, no customer data is used for training AI models.
Yes.
If you’re selecting the managed package then you will be provided an OpenAI AI model, which is hosted on one of ValueFlow’s servers, located within the EU. Also, your ValueFlow server instance will also be hosted within the EU. Those hosting aspects in combination with the fact that none of your data will be used for model training, make ValueFlow GDPR-compliant.
This allows you to interact with internal and customer-related data within your flows.
All Fields & Details about this Flow
Here you can see a list of all available fields for this specific flow and some examples on how to make the best use of them.
Field | Description | Example | |
---|---|---|---|
![]() | Content | Describe what the press release should be about. | • [company X] announces that ... |
![]() | Quotes | You can list some quotes by relevant persons that may be included in the press release. - Full Name - Job Title - Quote | • Jane Doe • CEO • "We are very happy to announce this milestone." |
![]() | Press contact | Contact details of a company employee for journalists. - Full name - Job title - Contact information | |
![]() | Outlet | The outlet (platform) where the publication will be published. | Company Website |
![]() | Target length | How many words the publication should roughly have (without finalizing instructions). | 1000 words |
![]() | Audience | Describe the typical audience who will consume the publication. | |
![]() | Brand information | Give details about the brand that will release the publication. | |
![]() | Intention | What do you want achieve with the publication? Describe here its intention/call-to-action. | • Reach out to us, if you would like to speak to us about this announcement. |
![]() | SEO keywords | If you want the publication to be SEO-optimized for certain keywords, you can list these keywords here. The first keyword will be considered the most important one, and so on. | |
![]() | Language | The language the publication should be written in. | |
![]() | Format | Which format should the publication have? | |
![]() | Specific Instructions | Here you can give specific instructions on how the publication draft should be written, relating, e.g., to the brand, the branch, the language, or similar. | #Company-specific terms with a certainly defined spelling: • Write the brand ebay, always in lowercase letters; correct: "ebay", wrong: "Ebay", "EBAY" # Company-specific linguistic preferences: • in German, do not use a comma before 'und' • in German, if there is a combined term that demands the connection through a '-', then use the '-' in all connection possibilites. Example to make it clear: English -> "machine learning solution“; correct in German -> "Machine-Learning-Lösung“; wrong in German: "Machine Learning-Lösung". |
![]() | Examples | You can provide examples of similar publications for reference how the output should look like ("few-shot learning"). | • [historic press release #1] • [historic press release #2] • [historic press release #3] |
![]() | Finalizing instructions | Instructions you give here will be returned in the flow output. They are intended for the person who finalizes the publication afterwards. | • if a quote is used: add an image of the quoting employee |
![]() | Evaluation Criteria | The criteria to evaluate the text against, each given as a description. | • Ensure that company-specific spelling preferences are met. • Ensure that company-specific linguistic preferences are taken care of. |
![]() | Minimum degree of compliance | A degree between 0 (0%) and 1 (100%) to which each criterion - when interpreted quantitatively - must be fulfilled by the text. Default: 0.8. The higher the value the stricter you are. | |
![]() | Max tries | How often the flow can try to improve the text until giving up. A Good rule of thumb: The amount of evaluations * 3. This way you allow to improve each criterion 3 times. |