#30DaysAIPost Challenge on LinkedIn (Batch 2)

Augmented Reality + Audience Targeting | Day #6

My Thoughts –
I am sure you must have heard and experienced Augmented Reality in your day to day life on social media, ads, websites, games etc. I feel its potential is yet to be explored to the fullest extent. As a marketer I have seen the immersive non-intrusive experience AR can provide to the consumer using just a mobile device. And for advertising the most important use case I feel that works is VIRTUAL TRY ONs.

I have seen that VIRTUAL TRY ONs create a sense of inclusion among the consumer. Take for example, it lets the consumer try a shoe virtually beofer they buy. Or try a new hair color or the newest sunglasses from Ray Ban. Experience new furniture or appliance in their home setting to make better layout decisions before a purchase. Such experiences make it convenient, fast, flexible and gives consumers a variety of options at their fingertips. Indirectly, It also saves wastage of valuable man hours and avoid resources limitations for the brand. 


Such brand engagements create a strong brand recall and users are able to interact with the product in realtime. And with advancement in the technology it’s getting accurate by the day. Wearable AR Devices, accessibility to 5G and high speed internet, Cross platform application systems and integration of AI together is adding to hyper-realistic and smoother AR experiences.



Adding a layer of audience data will refine the consumers experience manifolds. Imagine an ecommerce platform showcasing the user AR Experience basis their past purchase history or an Apparel brand advertising to the travel enthusiast audience suggestions basis the travel intent for example beachwear, high fashion, party wear, business casuals etc. I am surely keeping an eye out on this stream for product development and some market first innovations. Will keep you guys posted.

No Code + Generative AI | Day #7

My Thoughts –
A few years back when I was talking to a friend she mentioned that CODING skills was part of her kids curriculum. It surprised me as her kid was in the 5th Standard and getting exposure to coding at that age meant that the kid would have a bright future. Or so I thought. As time went by and AI took major leaps in the years it became more and more clear that CODING / Programming as we know it was about to change drastically. With the Assistance of AI the world is moving towards a NO CODE era.

What is No Code? So No code is a stream of Product development that lets users with basic or no knowledge of any programming language, create softwares / apps using NO CODE Tools. These No code tools can be of different types. But their basic idea is to make the software development process as simple as possible without complex coding knowledge requirement. 

I have see 2 major categories in this.

The first one is GUI Based, in this category you will find tools like Bubble.io which lets users design interface of the app using Drag and Drop features and readymade components and use connectors to connect it to databases / apis and other relevant services needed to run the App. The tools usually provide a SaaS model where you pay for the usage on a monthly basis and then charge you basis usage. 

The second one is Prompt based and it’s what I like personally. These tools are usually Full stack SaaS applications that lets you build anything just with basic prompts. You will find tools like Bolt.new in this category. These tools basically handle all operations right from design, development to even deployment in some cases. You can prompt to customise design, make changes to functionalities, add new features etc. They again work on AI models hence charge you a fixed monthly fee and also additional as per usage. 



Now circling back to the story I shared in the start, and its connection to this. I always felt that the schooling and education system was the base of your start to actually build things. But I feel things are changing. It’s still great for a kid to start learning coding early on in his life but then again, it’s more important to understand how things work. Both in the 1st and the 2nd category of No Code tools that I mentioned, it’s more important to know what you want to build and how it functions than to actually code. 

The entire focus is shifted towards using either the flow of the app (for User interface) or breaking down the problem to smaller chunks (for prompting each section of the app) than the actual programming. And here the more important function that will help the kids would be better communication, logical thinking and reasoning. So kids need to now be more aware of the holistic growth than just learning a programming language as languages keep evolving and the logics remain the same.

As Jensen Huang (Nvidia) rightly said – It is our job to create computing technology such that nobody has to program, and that the programming language is human. We are entering a world of automation where we will be like Aladdin instructing our No Code Genies to build things for us. And the Genies will do as per our wish. The only difference is that there wont be a limitation to the number of wishes you have. So in such a scenario, what would you wish for? 



LoRA + Generative AI | Day #8

My Thoughts –
Let me introduce you today to the fascinating world of LoRA. LoRA, or Low-Rank Adaptation, is a machine learning technique that helps adapt large models to new tasks. It’s a lightweight approach that can quickly fine-tune models without retraining the entire model. 
Think of it as a specialised system. Generative AI models generate images pixel by pixel. The pixel data is trained when generating the models using tagging systems. That way it understands what the pixel is about to build. But imagine, if you want the AI Model to generate images of something it’s not trained on? How would the model know? Thats where LoRA’s come into play. All you have to do is train it with specific data. For example your photos. This helps in letting the Model know how you look and what your features are.

LoRAs can be trained on specific data so that the Generative Model can identify and use that information to build something new with the unique identifier. Why this is important? It because it changes the game entirely. It makes the entire experience customisable. We can training LoRAs to generate images of Products, Brand Ambassadors and a whole lot of other things. The possibilities are endless.

In advertising this gives creative teams and the brand freedom to explore, experiment and visualise things that would otherwise be too tough. For example, once a LoRA is trained on a brand ambassador, the model would easily be able to render images of the Brand Ambassador in various scenarios, various outfits and a lot more. This speeds up the production process and also makes it flexible to work for pitches or decks as well. 
The LoRA training is not just limited to Images, it can be trained for a wide range of datasets. I would not get into the technicalities but the point here is that such methods can also be used for inspiration. For example, a LoRA trained on a car model can help the car designers visualise unique variations of the car. Even futuristic concept models can be visualised in minutes instead of days.

This is the power of LoRA and Generative AI. Maybe one day I would not have to post on social media, maybe It would be my LoRA trained model automatically posting on my behalf? I will leave you with that thought! See you tomorrow.

LLMs + Research + Education | Day #9

My Thoughts –
It was a normal day until I checked my mail and found this gem. An X.com user HudZah https://x.com/hud_zah built a neutron-producing nuclear fusor in his kitchen using Claude over a 36-hour livestream.

My mind was blown! Did I read it right? I went to the thread and started reading in detail and of course the news was true. 
Why was a I shocked? This isn’t new. There were examples of people making DIY innovations just by googling and reading about it but what was so different about this one?

The fact was that he used Claude AI. A ChatGPT alternative that I have been using for a while and really find very helpful. 
This just makes me realise that in the current age, knowledge is easily accessible. It’s only the earning to mould it into actions thats needed. Also AI and LLMs role in the field of education will be really huge. As it breaks away from a cluttered search model to a highly precise data driven approach.

I have been using these tools to self educate on prompting, writing python code, experiment with plugins/extensions for chrome / after effects / figma etc. And every time I build something it gives me the push to build one more.


As Abraham Lincolns quote – Give me six hours to chop down a tree and I will spend the first four sharpening the axe. In the current context of my life If I have a job at hand, I would spend 4 hours researching on the LLMs to find better, efficient solutions. So that I can build a faster / better and smarter output in the next 2 hours.

So what do you plan to learn and build next?

UGC + Generative AI | Day #10

My Thoughts –
I have a habit of researching online for any product before I buy it and I see tons of video reviews in the process. And I feel it’s a part of the advertising journey too as User-generated content (UGC) comes from people, not companies.
User-generated content (UGC) has become a pivotal element in modern marketing strategies, particularly in enhancing word-of-mouth advertising (WOM). UGC encompasses any content created by consumers about a brand, including reviews, testimonials, photos, and videos. Its impact on WOM advertising is profound due to its authenticity and relatability, which resonate more with potential customers compared to traditional brand-generated content.

If you look at stats, approximately 79% of people report that UGC significantly influences their purchasing decisions. This statistic highlights how consumer-created content can drive sales more effectively than traditional advertising (Source Nielsen) Also from a marketer’s perspective, 93% of marketers agree consumers trust content created by customers. Source: Hootsuite

This shows the power of UGC and imagine what happens when AI enters the game. I have, of late, seen a lot of tools that help you generate UGC video reviews / content / ads using AI Avatars. Some of the tools are https://www.makeugc.ai/?ref=abhilash, https://www.captions.ai/tools/ai-ad-generator and https://try.arcads.ai/ai-ugc. I haven’t had time to test it but I would be soon trying these. But the concept itself amazes me.

The amount of flexibility it provides is immense. you can choose your UGC character, his / her ethnicity, his / her voice and language and even make them hold your product. Yes, it’s possible. With AI you can make an AI Character hold your brand product and review it. The fact that such an innovation is possible boggles my mind.
The fact that such UGC can bring down the marketing production cost drastically and help in a faster GTM strategy is worth appreciating. It’s great for scaling operations by creating variations of the same content in various, languages and perspectives with just a click of a button especially for SMEs and startups.

Although It does raise some ethical questions. What happens if the product is bad and people believe the UGC content and purchase the product. What about misusing such ads for scams and frauds. Yes, like any great power comes great responsibility. I hope in the coming days there are few regulations set on these UGC ads to standardise and identify them. Till then let’s be amazed and see the route the AI powered UGC Content charts.