Workato, a low-code/no-code platform for business automation, has announced a strategic partnership with OpenAI to include numerous OpenAI AI models and future releases into Workato. The use of generative AI is at the heart of this collaboration, which seeks to streamline the process of developing automation and integrations.
Workato has said that it would be releasing a number of new features as a result of the partnership.
Workato Copilots are one example; they allow users to create automations and app integrations with simple English explanations. With Workato’s OpenAI connection and AI connectivity, customers may add generative AI capabilities to their automations.
“Constructed using OpenAI models, the Copilot is like having a Workato-savvy colleague in the office, capable of creating process recipes and data connections through dialogue. Workato’s founder and director of products and engineering, Gautham Viswanathan, told VentureBeat, “It has been trained on millions of data points from Workato’s public community.” As the developers of Workato Copilot put it, “We believe Workato Copilot will further lower the barrier of who can build within an organisation.”
Viswanathan elaborated that the Copilot will aid customers with things like getting started, learning new features, figuring out what to create next, getting suggestions, and solving issues quickly.
Through popular messaging platforms like Slack and Microsoft Teams, users will be able to have natural, conversational interactions with workplace applications and data using another feature called WorkbotGPT.
RecipeIQ, Workato’s proprietary AI/ML models, are already included into the platform and offer data mapping, reasoning, and next-step suggestions. Workato plans to make it simpler for organisations to embrace its technology by adding OpenAI’s models into its automation and integrations development process.
The partnership, according to the corporation, also assures strong security and governance skills, fosters trust between IT and business teams, and promotes effective scaling operations.
OpenAI model-based business process streamlining
According to Viswanathan, the process of incorporating OpenAI’s models into the Workato platform required taking into account a wide variety of use cases demanded by the company’s customers across a wide range of sectors.
Examples of such applications include meeting summaries, automated email sequences, and virtual assistant creation.
The business chose its LLMs by analysing the automations its clients are already creating on the Workato platform and imagining how they may be improved with the help of generative AI. The group then investigated OpenAI models to find the most appropriate ones for each application.
Viswanathan said to VentureBeat that this is what motivated the company to choose many LLMs and then train them using their own proprietary models so that they would perform optimally in the aforementioned use scenarios. Our clients will be able to take use of these models when they develop automations, integrations, APIs, and application connectors on our platform since we have integrated them effortlessly.
In 2018, Workato launched RecipeIQ, an artificial intelligence (AI) tool that uses the company’s own unique ML algorithms to identify actions for customers to take inside their workflows. The business claims that the Copilot will improve upon this capability by allowing the builder to converse with it as the recipe is being constructed.
To eliminate the requirement for pre-built components, Viswanathan claims that the WorkbotGPT capabilities would enable real-time automation in corporate operations.
For teams using Slack or Teams, WorkbotGPT provides conversational automation. You may use natural language cues, and it will search for transcripts of recordings in Zoom, your email, and your CRM in real time, generating a summary of action items for you.
Safeguarding the Future of Automation
Workato claims that the platform’s incorporation of a strong governance architecture makes it possible to use AutomationHQ to manage federated workspaces across business units.
Customers get complete access to all of their assets, data, and logs with this organisation. Customers may decide who is allowed to use AI services thanks to the platform’s role-based access controls and granular permissions.
Customers may also choose the log storage time, have sensitive data masked, audit all changes made by users, stream logs for centralised monitoring, and more.
Customers that must adhere to stringent data residency and sovereignty standards may rely on the services of our data centres located in many regions. Our Copilots do not utilise information gleaned from these conversations with customers to train any model, as Viswanathan noted. Multi-layer encryption, hourly key rotation, bring-your-own-key (BYOK), and zero-trust rules serve as the backbone around which these features are constructed.
Where will Workato go from here?
Viswanathan said that at this time, the firm is training its models on data collected from user-created automations, integrations, and internal APIs. Through this education, the firm hopes to create further effective technologies like Copilot and WorkbotGPT.
He thinks businesses will become more open to share data with external LLMs as they realise AI’s potential.
That, he added, will lead to “a set of exciting possibilities,” some of which we can already imagine and others of which we won’t be able to conceive of until we completely comprehend the scope and depth of the accessible data. To address this issue, we’re developing new products and services that combine artificial intelligence (AI), automation, and integration into a unified framework.