How to think about applying new technologies like AI and ChatGPT
https://www.productfoundry.co/issues/issue-2-maximize-your-product-potential-with-ai-a-strategic-approach
Be strategic — How I think about applying new technologies like AI and ChatGPT to my products.
Most of the AI tech can actually be applied pretty easily through API. No longer do you need knowledge in ML to make an AI app
Even with image generation, you can easily implement that with Evoke (AI API tool i'm working on)
But finding a need is still important.
I am a developer, so I know at least roughly how AI works.
It is NOT intelligent. Basically just a huge set of conditions being executed.
However, after being shown the potential applications and trying it out myself, my perspective changed. I now see AI as a tool, similar to a hammer, that can be used for various tasks. While it is not capable of developing its own unique ideas, it can be utilized as an assistant by providing it with specific instructions and keywords to improve and refine the output. I personally have found success in using AI in this way, however, it is important to always fact-check and verify the information provided by AI as it is prone to errors.
AI assistants are undoubtedly part of the future. Today, the world's most successful and influential people are supported by an army of people helping them navigate the world - whether they are employees, coaches, or medical professionals.
The problem is these services are expensive and so most of us end up barely using them - but models like GPT-3 flip this on it's head. You can now build products that provide that kind of "personalized, high touch experience" for 1/1000th the cost.
...and if you haven't already guessed - that's exactly what I'm building with www.habitcoach.ai
I was listening to the My First Million podcast. One of those guys pays a nutritionist to call him daily to ask how his eating was yesterday and review what he ate. $600/mo. You might be onto something here.
Totally, this is a model that rich people are already using, we just need to make it more accessible!
I think we'll have a range of customized "assistants" for different use cases and then over time they'll merge into a single assistant.
This is the future of customer service. I firmly believe we will be seeing AI contact points, which are highly scalable, across the board. In-person calls with deep fake technologies, chatGPT like communication and high fidelity transcriptions are all the perfect building blocks for outcompeting any large customer service tool.
We are already seeing some being developed. But I think the right time to jump in is right now. You might just win the fight.
Zooming out there are AI bots which outcompete even artists. Imagine a chatGPT type interface and you just struck gold. I actually wrote an article which I would like to link on this phenomenon: https://medium.com/user-experience-design-1/digital-humans-are-here-they-sound-and-look-just-like-us-f1563c96e5f8
I couldn't agree more! The use of AI in customer service is definitely the future and it's exciting to see the possibilities that it brings to the table. The scalability and efficiency that it offers is unmatched by traditional methods. The advancements in deep fake technologies and chatGPT-like communication are definitely game changers in the customer service industry. You're right that now is the perfect time to jump in and potentially come out ahead of the competition. I'll definitely be keeping a close eye on this space and the companies that are leading the charge in this field. Thanks for sharing your thoughts on this!
No problem! I am excited to see if you have anything in store for us. Please share if you do!
I think you're correct about the impact on tech support. Load all your knowledge base, blog posts, and how-to's into a GPT model, and BOOM, you've got a fantastic contextually aware bot. You can even tune the voice it speaks in.
I even hadn't considered deep-fake spoken voice yet.
This goes far beyond my personal skill set but I consider it an ultimate b2b tool which could outcompete big customer service providers. The problem lies in responsibility and giving boundaries to the algorithm. How could anyone make sure that chatGPT is providing the truth and is not stepping on anyone's toes?
Thanks!
Great! See what we're up to with GitScrum + ChatGPT. Our AI system can generate descriptions, labels, and checklists based on the title of a task, helping you to reduce the amount of manual work involved in organizing and managing tasks.
Those all sound like solid use cases.
Thanks for sharing!
Pretty insightful stuff there!
Thanks Abdullah