GAI Form Generator
Nintex
Focus
User Research, UI/UX Design, Interaction Design
Deliverables
High-Fidelity Mockups
The GAI Form generator uses AI to help form designers of all levels easily create a form from a single text prompt. Additional suggested fields can be added, and once the user is satisfied with the base form, they can take it into the full Form Designer experience for further customization.
Overview
AI is leading the way in innovation and is quickly being integrated into all SaaS companies. The GAI Form Generator marked Nintex’s first step in showcasing our active commitment to advancing AI initiatives.
The Challenge
How might we save users the headache and time of creating a form from scratch?
Turning a proof of concept into reality
Form Designer is one of the most widely used tools within Nintex Automation Cloud, offering an easy-to-use drag-and-drop interface. However, creating forms from scratch can often be cumbersome.
At the time, AI features were not on the roadmap, so our product lead decided to leverage the company Focus Week to test out it out.
Building on a POC that was created by one of our lead engineers, we spent about two weeks—exceeding the usual 1 week timeframe—developing this new feature.
Given the uncharted territory and new technology involved, we approached it as an experiment to gather feedback from customers.
I explored several iterations for integrating the AI tool into our UI, addressing challenges such as UI limitations, unexpected use cases, and concerns about AI speed.
Our generator could not apply real-time updates, it required users to make changes incrementally and click "apply changes" at the end.
We ended up launching the feature as a beta, with plans to release improvements in phases based on user feedback and identified issues.
Personally, I felt like this feature served more as a marketing tactic than a genuingly useful feature.
Post launch, I conducted a survey to understand the following:
Was the GAI form generator easy to use?
Does it add value to their day-today business processes?
Can they see themselves using it daily?
Did it behave as they expected?
What else is missing or could be improved?
Marketing tactic or solving a genuine user need?
While users found the form generator easy to use, they criticized the slow load times and the excessive steps needed for updates.
They appreciated it as a first step in AI but noted it added little value to their daily operations, particularly for complex forms with company-specific use cases.
This research provided a solid foundation for advancing our AI efforts.
AI gaining momentum in the workplace
As AI began gaining momentum at Nintex, an AI task force was created to standardize backend code and practices. Various products were also beginning to explore AI capabilities in their respective areas.
I was soon tasked with developing the next iteration, AI Co-pilot. The vision was for the GAI Form Generator to help users start a form, then AI Co-pilot to float along side the UI, assisting them in customizing the rest.
With this new focus on AI, we ensured thorough research was conducted across each domain. UX designers working on different products undertook AI research with the goal of reconvening to synthesize the data collectively.
Our primary focus was to understand our customers' current mental models around AI and to identify the best applications for it within each product.
AI Research: Understanding Current Customer Mental Models
We conducted moderated interviews which set out to learn the following:
Customer mental model/understanding of AI technology
What customers struggle with most in the current Form designing experience
What parts of the Form Designer they think would benefit the most from AI technology
We discovered that AI is still too new for most businesses. They are all still trying to figure out how to integrate it into their daily processes. Their primary concerns are around data accuracy and governance. As much as they would like to train the AI with their own data, this is also often challenging due to their data being unstructured and dispersed.
Most users see more value in AI automating repetitive tasks, because it saves time and offers a manageable bitesized approach to begin leveraging AI without risking data breaches.
A re-occuring theme also included assistance in visual styling and rule/variable setup. Styling is particularly hard for our technical form designers and rule/variable setup is not easy to learn for our non-technical designers.
They also liked the idea of this AI tool speeding up their workflow without forcing them to switch context and screens, it could allow them to manage most things from a single interface.
After the research synthesis, we conducted a Benefit/Effort analysis workshop to prioritize our focus areas, identifying key features such as:
Bulk editing controls (e.g. Making all Signature controls mandatory or all Submit buttons blue)
Using Co-pilot to quickly access Help documentation
Having Co-pilot suggest ways to improve form efficiency
AI Co-pilot exploration
Additionally, I designed the overarching framework for how users would interact with the AI Co-pilot. These sketches effectively fostered team discussions, allowing us to visualize the concept and raise important questions.
These sketches were also shared with other product teams and senior leadership to brainstorm AI co-pilot integration across all products at Nintex.
Our UX AI Guild also began gathering common AI interaction design practices to inform our UX strategy.
Solve genuine human needs
AI has quickly become a buzzword in the tech industry, yet many businesses are not ready to fully leverage its potential. Ironically, not having AI can raise concerns among businesses as well.
While AI offers great possibilities, it remains largely uncharted territory, and as UX designers we must make sure we help businesses focusing on solving genuine human needs to avoid building powerful system that address minor or nonexistent issues.
Given the apprehension around AI, it's essential to introduce it gradually, allowing customers to adjust, build trust, and eventually learn to work with it as with any new innovation.