We are firmly in the era of AI becoming a permanent fixture on the most personal computing device in our lives: the humble smartphone in our pockets. Every operating system out there is racing to integrate a frontier AI model to interact with apps and get work done. The shift is huge, but the average smartphone user is still locked into a repetitive cycle where each task requires its own detailed prompt, often followed by a back-and-forth to get the job done. That’s where AI agents come into the picture.
But just how far can agentic AI go? Digital Trends exclusively worked with the team over at TECNO to develop a custom version of EllaClaw around the specific friction points in editors’ and content creators’ workflows. That exercise offered a much clearer glimpse of what agentic AI on a phone could become — the more meaningful promise of mobile agentic AI lies beyond a fixed menu of pre-set functions.
It should be flexible enough for each user to turn their own pain points into personalized intelligent solutions. In that sense, the custom EllaClaw did more than add a few tailored features: it pushed the boundaries of what the system could do and showed how powerful deeply customizable agentic AI can be. It also made the possibilities ahead feel far more tangible. The results wowed me.
Interestingly, not many smartphone labels out there have executed it yet. TECNO is taking the lead with the integration of agentic EllaClaw on its flagship, which was announced earlier in June this year. Even though it’s still a beta project, it can already handle multi-step tasks in one go, interact with apps in the background, and even perform hardware optimizations to make your device run fluidly.
I’ve used the most advanced AI models that an average person can now access — after paying $200 per month, of course — but the agentic implementation on the TECNO EllaClaw genuinely won me over with sheer simplicity and focused productivity.
Research
As a managing editor, one of my biggest responsibilities is running a newsroom. And when it comes to news, what matters most is trust and thorough research. On an ordinary day, if I am researching a set of events or a topic, it would usually involve manual research, spread across a dozen or two Chrome tabs, switching back and forth between different sources, and then taking notes before I can distill it all into my own draft.
It takes a lot of time. Of course, I can get it done by simply asking an AI chatbot to handle the research, but here’s the problem. Even the most advanced AI models have a habit of hallucinating and just spewing imaginary information. Second, they often pull information from sources that are not really trustworthy. That’s when I turned to TECNO EllaClaw and its custom implementation for my exact workflow.
Instead of having it generalize information, Tecno and I created a custom research agent that only pulled information from a select bunch of reliable resources, such as BBC, The New York Times, The Washington Post, and the Associated Press, to name a few. Second, I named this AI agent based on the purpose it serves, which is news Research.
Now, every time I have to perform deep background research on any topic or event, I simply write “News Research” followed by what I need, and TECNO EllaClaw handles the tasks by only extracting knowledge from the trustworthy sources that I have specified, ensuring that the information is correct and there is no risk of hallucinated data creeping into the information that is served to me. There is no back-and-forth involved, and the entire task is handled in one go. There’s even the flexibility to create scheduled tasks.
I absolutely love this feature because it also comes in handy for me when I am not at my desk, but still want to collect information on events and subjects that I plan to write about later on. While sitting in a cab or just walking to the local grocery store, I simply task TECNO EllaClaw to handle the research for me, and with a single tap, all of that information is also saved to the local notes app without ever leaving the Ella AI assistant.
What I appreciate the most is the flow of knowledge. Instead of giving a summarized answer, TECNO EllaClaw can also break down the chain of events so that you get a background recap, while also pulling opinions from authoritative sources. There’s a balance between history, current events, and expert commentary that offers a comprehensive recap of any topic or event one is looking for. As a journalist, you can only ask the right questions if you’re armed with the appropriate knowledge, and this is where the News Research capability comes in handy for me on a day-to-day basis.
Maintaining my voice
Lately, I’ve leaned more on AI and voice tools instead of typing everything out. When an idea strikes, I record a quick voice note and later use the transcript as raw material for an article. The problem is that those spontaneous notes still need a fair amount of cleanup — grammar, transitions, and basic structure — before they resemble a finished draft.
I could hand that job to a generic AI chatbot, but I’m wary of the distinct AI tone it can introduce. These tools are good at fixing typos and grammar, yet they often over-edit and smooth away the quirks that make the writing sound like me. For someone who writes for a living, that’s a deal-breaker. What’s more, conventional AI chatbots cannot carry out actions across multiple apps.
Thanks to TECNO EllaClaw, I no longer have to go through that process. To handle the aforementioned tasks, I created a language profile that is trained on my own articles. The whole process is pretty straightforward and intuitive. I can simply copy-paste my drafts, or pull the URL of a published article using the share button and feed it to TECNO EllaClaw with just a few taps.
Essentially, TECNO EllaClaw captures the essence of how I write and what my unique style is. Now, all I have to do is copy-paste the transcript of my voice notes, feed it to TECNO EllaClaw, and it tidies up those ramblings into a finished article that reads exactly like my own distinct written material.
This way, I can save a lot of time on doing basic language correction, while also ensuring that all the thoughts are my own, without any addition or filtering by a large language model. The whole system works like a charm, and I have saved plenty of hours using this specific personalized writing skill that TECNO created for Digital Trends.
A savior for scripting
EllaClaw’s personal writing skill also comes in handy every time I have to transform one of my own articles and hands-on impressions of a piece of hardware or software into a script. Now, writing an article and writing a video script are two entirely different things. The only shared trait between these two is your unique voice and flair.
But compared to writing an article, a script is a lot more technical and granular. This is where EllaClaw’s personal writing skill goes hand-in-hand with its scripting capabilities. Now, every time I have to write a script for one of my own reviews, all I have to do is invoke the shortcut, which is actually nothing special, but a natural language command.
I simply feed the URL of my own article, or a copy of the text that I have written so far, and TECNO EllaClaw turns it into a step-by-step script, complete with a breakdown of the duration for each segment, the view in front of the camera, and what exactly I am supposed to say. On an ordinary day, brainstorming a full script usually takes me hours. But thanks to EllaClaw, I can not only get it done within a few minutes, but also retain my own original voice that is carried over from my articles to my own videos.
There’s another side to the exercise that I often miss. While writing a script for videos, we usually rely on our own firsthand testing and what we have already documented as notes or drafts. However, that may not necessarily paint the full picture of the product or topic at hand. When using Tecno EllaClaw for scripting Word, the AI assistant can also perform online research and look through community forums to gauge what the world thinks about a certain topic.
Accordingly, it can make suggestions and additions to the script so that the research is thorough and there are no blind spots. These additional inputs are, of course, presented alongside the source from which the information was extracted, so that if I need any additional context, I can visit those sources myself and get a more comprehensive picture before finalizing the script.
Getting ready for social media
Social media may seem like a straightforward journey where you simply open the camera app, say what’s on your mind, and share it with the whole world. Technically, that’s how it should work, but for a content creator, it’s a whole maze. One slip of the tongue, and your video can get shadow-banned, demonetized, or simply blocked from the algorithmic dispersion. Additionally, an overtly marketing-friendly tone or just rehashed opinions are also algorithmically punished.
Needless to say, you need a lot of prep work, but it’s hard to remember all those unsaid rules while writing the script for videos. To save myself the hassle and post-publish regret, I turned to another custom skill that readies my script for social media and warns me if there are any problematic segments that can be punitive for the video’s reach.
All I have to do is simply paste the script and ask EllaClaw to scan it for any risk words. In just a minute or two, the custom AI agent gives me a long breakdown of any risky words that can suppress or attract an algorithmic penalty. It also takes a look at the overall language and flow of the script and makes appropriate suggestions to ensure that the content is interesting and ready to grab some eyeballs on social media platforms.
A leap forward for the agentic smartphone dream
When generative AI first landed on the scene, and to this day, getting the best out of it has required a certain learning curve. You have to write an extremely detailed prompt, or go back and forth to make corrections, and specify certain instructions before you get the desired outcome.
That should not be the case. In my opinion, AI should actually reduce the manual effort and eliminate the steps that require repetitive work. But most importantly, it should do so in a naturally interactive fashion without any code or custom prompting to act as the magic ingredient. It should be an enabler for everyone, and that’s exactly what TECNO EllaClaw has accomplished.
It automated tasks for me that save hours each day, without requiring me to remember any special commands. Moreover, it seamlessly interacts with the other apps on my phone if I need the data from my AI conversations to be saved elsewhere. For example, as soon as TECNO EllaClaw is done writing a script or naturalizing my voice notes into an article that sounds like me, with a simple word, I can get it saved in the Notes app, without ever switching back and forth between two applications.
The overarching goal is to present an agentic AI system on a phone that is not only autonomous when it comes to performing multi-step tasks, but also happens to be deeply customizable to each person. A truly useful AI is one that is flexible, adapts itself to different scenarios, fixes the usual pain points of human-machine interaction, and does not limit users or ask them to switch between multiple tools to get work done on a day-to-day basis.
In my opinion, that’s how AI should work. I can’t wait to see the deep interplay between system-wide AI and the applications that we use on a day-to-day basis. TECNO EllaClaw offers a glimpse of that convenient future, and I hope to see it ease my daily smartphone chores. It has already eased my usual back-and-forth workflow across different apps and reduced the number of steps it takes to perform a certain task. I can’t wait to see more customizable automation routines tied to Tecno’s AI framework and turn my smartphone into a true AI-powered computing companion.