Without a doubt, the launch of OpenAI’s ChatGPT was a watershed moment, capturing the fancy of almost everyone. Its ability to understand and process natural language, and at such a speed was something that amazed all of us at Scalefusion. The use cases of ChatGPT’s adoption in customer service and content creation/marketing were obvious (well, however, this article was written by a real person), but the Scalefusion team started thinking about it from an IT team’s and UEM perspective.
Scalefusion has constantly been at the forefront of adopting new technologies that empower our dear IT admins with the right tools that make managing devices a breeze. We also constantly focus to try and go beyond the generic UEM and think about the outcomes that IT teams want to achieve from a UEM solution.
Though we do see certain challenges with ChatGPT3.5 and the APIs that it offers, based on our diligent analysis, usage and testing, we believe that even in its current format, it can be of great help in complementing the IT teams’ day-to-day activities. It can, in all sincerity, become a trusted Wingman for the IT admins. With the public access of ChatGPT 4 around the corner, a tool that is designed correctly is going to be of great value and in all probability is a must.
Introducing AirThink AI
We are super stoked to present AirThink AI, a product that showcases Scalefusion’s wisdom in managing devices and powered by OpenAI’s ChatGPT APIs which we believe will be the IT admin’s superpower that they can invoke and interact with, at will, right from the Scalefusion dashboard
Defining the path for AirThink AI
When our team started off with developing AirThink, we were pretty clear that we are going to look at this from a ‘problem-AI-can-solve’ rather than ‘AI-created-problems-to-solve’ angle. We scrutinized the available APIs keeping in mind the core goals of AirThink AI which reverberated the following:
- Empower and not enforce, by providing limited and context-aware intelligence
- Facilitate interactions that would look and feel natural for IT Teams
- Retain the context of the conversation throughout the interactions
- Require absolutely zero setup
- Prevent out of context usage
We believe that we have been successful in achieving most of our core goals (well, almost 🙂) and think it is the right time to let our customers leverage what we have built so far.
Empowering with context-aware intelligence
The foremost thing to note is that with the introduction of AirThink AI, you would see that it is presented to you in contexts where an IT admin would most probably need hand-holding or guidance.
In our current version, you can use AirThink AI guidance to:
1. Generate Powershell, Shell, Bash or Python Scripts right from the dashboard with the Script with AirThink AI feature
2. Get a deeper understanding of self-developed/in-house scripts with the Validate with AirThink AI feature.
Making way for a human-first design
Next, our focus was to make sure that the interactions are easy, natural and free-flowing. A close look and you would observe that when you try our Script with the AirThink AI feature, we have not opted for overly-complicated design patterns but a familiar and easy-to-use chat interface.
This minimalist design allows the admins to:
- Ask or define their problem statement using natural and native language.
- The responses are presented with well-formatted script/code blocks, allowing them to maintain focus on the solution.
- These script/code blocks can be quickly moved over to a nifty editor right next to it, thereby leveraging the established design patterns of powerful IDEs.
Keeping up with context
Up next, the idea was to retain the context of the ongoing interaction with AirThink AI wherever applicable. IT admins do not need to provide the context of the conversation each time they interact with AirThink AI because the context is inherently maintained, paving the way for quick and efficient solutions. We also retain the historical interactions so that users can pick up from where they left off rather than starting over from scratch.
No-fuss, no-frills for setup
The final goal of phase #1 was a non-negotiable, and it was to make sure that using AirThink AI would require absolutely no additional setup. IT admins can get cracking right from the word go via their Scalefusion account. If your account has a feature that is powered by AirThink AI, your IT teams can use AirThink AI right away. And yes, there is no waitlist!
Limitations, Usage Advise and What’s Next
As for the roadmap and our future plans for AirThink AI, we are working on tackling the problem of safeguarding against unwanted usage. Though we have taken care in providing context-aware intelligence, the APIs are still limited in blocking the users from deviating from the actual conversation. Our machine and human interaction learning algorithms are being fine-tuned to mitigate this and overcome this challenge.
In addition to the above, like with any script or an app that needs to be rolled out in a phased manner, scripts or artifacts generated by AirThink AI need to be tested on a limited set of devices before they are rolled out en-masse.
We have always, and will continue to advise to make sure that deployments are done in a staged manner to avoid any business interruptions, and we are complementing it further by providing a fine-grained ACL/RBAC feature and auditing of all actions.
Notwithstanding some of the challenges of AI and its implementations, what is more important right now is to find the right set of problems using the available algorithms. With AirThink AI’s first version we believe we have found a natural fit for AI in the UEM space rather than a simulated need. What we are excited about is the areas where we see AirThink assisting IT teams and have quite a bit of aces up our sleeves.
Needless to say, we can’t wait for Scalefusion users to try out AirThink AI and transform their scripting capabilities right from the Scalefusion dashboard. While we brace ourselves for what’s next with AirThink AI, we would love to know what our users think of the AirThink AI experience.
We would love you to try AirThink AI and provide us with valuable insights. It would be great to get your ideas on what problem you would want to solve with AirThink AI or a place on Dashboard where it can help you with its wisdom, feel free to email your thoughts with me at product[at]scalefusion.com.
| “Please note that the usage of AirThink AI is at your discretion. AirThink AI uses external services (OpenAI ChatGPT APIs) but does not share any of your enterprise data or device information with the API. However, the questions you ask with AirThink AI may be passed to external APIs depending upon the question, so we advise you to exercise caution and not share any sensitive information.” |
