The image on the right is an AI-generated image created with Midjourney by Molly-Anna MaQuirl
In a recent conversation with Giancarlo Erra, a seasoned creative tech entrepreneur working within the world of AI, we gained valuable insights into integrating artificial intelligence tools within business operations.
With over two decades of experience spanning various industries, including e-commerce, environmental conservation, music production, and tech development, the entrepreneur sheds light on the intricacies of incorporating AI effectively.
Hi Giancarlo! Firstly, can you tell us about yourself and your career so far?
“I’d consider myself a creative tech entrepreneur; I think that is the best way of putting it. I started out working in a variety of different sectors, from environmental science to e-commerce to music. Over time, I transitioned into creative and tech roles, eventually founding my own ventures. Over the last few years, I’ve been particularly focused on start-ups in the world of Artificial Intelligence. For me, the allure of AI lies in its fusion of creativity and technology. It offers boundless opportunities for innovation.”
Do you think that companies should embrace AI in their business strategy?
“I think it’s like any technology; it depends on the company and the service they provide. At the moment, there is a very polarized view of AI. My approach isn't necessarily pro-AI or anti-AI; it's about pragmatism. AI is a tool, not a solution in itself. The challenge lies in dispelling misconceptions and understanding AI's potential within a business context. Many businesses misconstrue AI as a one-size-fits-all solution, overlooking its complexities. It’s putting technology before the problem. If you’re thinking about using AI, consider first what problems you actually need solving – and if AI could help second.
Many people use AI without being very literate about what it can actually do. If you want to implement AI into your business, I recommend using it as an internal tool first. This is a safer way to start and a safer way to get to know it. Start by simplifying an internal task, before venturing into customer-facing solutions. This will allow you to get to know the technology safely without the risk of brand defamation. A lot of people just want to cut costs with AI, but the focus should be on improving productivity. There is still cost savings; just a bit more indirectly.”
What do you see as the greatest benefits of businesses embracing AI?
“The benefits – like the risks - are limitless, so it’s hard to define. The best way I see companies use AI successfully is to fulfil repetitive tasks. Automating repetitive tasks with AI is great as they often take the most time and are the least enjoyable to do. Also, AI generally works better at repetitive tasks. This means saving time, and brain power to spend on something more important - or even more enjoyable!
This should come first before trying to save money. It’s a brilliant drafting tool. It could be anything—emails, articles, code, case studies. It’s a great tool for improving productivity, especially internally. Using it this way is not necessarily going to cut costs directly, but it's probably going to improve your productivity significantly. So, there is still a cost saving, it's just a bit more indirect.”
What are the risks of using AI in business, and what solutions do you recommend to mitigate them?
“AI can give you a great foundation to work from; however, this should always be proofread or reviewed before going to your customer. Human input is important. It’s tempting to automate processes completely, but it’s risky. You must always remember that you are responsible and accountable for any output your company generates, even if it’s created with AI.
As an example, AI Chatbots have been the subject of much news recently. DPD recently disabled its AI chatbot after a customer, through clever prompting, tricked it into swearing and writing a poem about how unreliable DPD's service is, calling it ‘useless’.
The challenge is that people don’t understand fully how the technology works before they implement it. You must always remember the technology is unreliable, and so you should be careful when using it for customer-facing services. If you want to use tools in this way, make sure you test and practise thoroughly. You must always remember that you can't trust it 100%.”
What advice would you give to companies who want to start integrating AI into their business operations?
“I think I’ve said it before, and I will say it again: Don’t put the technology first and the problem second. AI is amazing, but it’s not a one-size-fits-all solution. Companies should be pragmatic and make sure they have a thorough understanding of AI’s capabilities and limitations. First, think about what problem you're trying to solve, and then maybe talk with a consultant and see if that is really a problem that can be solved with AI.
Think about the most frustratingly manual process for you and start from there. If you’re doing something at work and you think, “Why can’t this be automated? Or simplified?’. That might be a good place to start with integrating AI.
It’s also worth remembering that current law states that you can’t assign yourself the copyright of something generated by a machine. To copyright something, it must have been generated by a human.”
What regulatory issues should business owners concern themselves with?
“It’s important to remember that there are additional challenges with bringing AI into regulated markets, such as finance or health. As I said before, if you integrate AI into your system, its output becomes your responsibility. It's a black box, so it's not accountable. It’s very hard to go back and see why an AI gives an output it does. Every time you ask an AI something, it will often give you a slightly different answer. Even if you develop your own LLM as an example, you still can’t guarantee consistent results. It’s one of the biggest usage challenges of AI, especially in these markets. If you let an AI decide something critical, even something it has been trained in many times, you still can’t guarantee it won’t suddenly be wrong, and you’re responsible for the result of that.”
Are there any AI tools that you would personally recommend?
“Most companies that use AI are using some kind of OpenAI API, or ChatGPT itself. There’s a reason for this, in my experience, OpenAI is way ahead of other tools. However, if you do, it’s worth experimenting. It’s worth investing in prompt engineering training and experience. If you know how to use it, it is incredibly powerful. If you don’t, you will create very generic text that’s easy to spot. It uses a very specific dialect. Alternatively, Claude AI is a good tool – its language style is less easy to spot than a tool like ChatGPT. In terms of image generation, I always use Midjourney over DallE, but that’s mostly personal preference.
If you only use AI as a drafting tool, it isn’t as relevant which tool you use, as you can always edit in that human element. If you are more advanced as a business, you should consider using custom solutions, such as deploying your own LLM. I’d never suggest people start with that, though – it requires understanding and experience. Of course, you need developers to host your own LLM, and that has a cost – but you can personalise it more to your own requirements. So, you know, that is just a different sort of product. There’s no point in making that investment unless you really need it; it’s always worth testing open tools first.”
What AI business trends do you foresee for the future?
“It’s always hard to predict! GPT 5 will likely be big this year, as ChatGPT updates always are. It will be interesting to see how the tool continues to develop and improve – I think it will have a big impact.
Otherwise, I think Sora will be very significant. It might be because I’m creative, but I think AI Video Generation in general will explode over the next year. There are currently teething problems – much like there were when AI Image Generation first came out and was giving everyone 6 fingers! Audio and Music Generation will likely see huge improvements too, especially in terms and background music such as jingles or loops. Whether we will see AI tools writing tunes as well as, say, Adele, is another story. I don’t think they have solutions for things like that yet – but who knows what’s to come.
In broader terms, I think we could see steps in the development of AGI (Artificial General Intelligence) in the next year, though I think we’re far off a real result. Recent enhancements, such as memory and context integration in ChatGPT models, enable better understanding and interaction with diverse information sources. These advancements pave the way for more comprehensive AI systems that can grasp larger contexts and offer more meaningful responses.
Looking ahead, it's all about integration. It’s about small, significant steps rather than major breakthroughs. Microsoft's Copilot, for instance, is bringing GPT into everyday tools like Windows, making AI more accessible and user-friendly. So, while we might not see any earth-shattering breakthroughs in the coming months, I think the future of AI seems to be all about fine-tuning and seamlessly integrating it into our lives.”
Thank you so much for coming in and talking to us today about your experience and thoughts, Giancarlo; it means so much. We appreciate your insights and look forward to seeing where you continue to go next on your AI journey!
“Thank you for inviting me in. I look forward to seeing how AI Atlas continues to grow!”
___
Giancarlo is the founder of Promethean Box, which has developed multiple AI-focused services, including Words.Tel, AudioShapes.Ai, and Copy Forge AI. You can learn more about his work at giancarloerra.co.
Stay tuned for more industry expert insights across the world of Artificial Intelligence.
The image on the right is an AI-generated image created with Midjourney by Molly-Anna MaQuirl
In a recent conversation with Giancarlo Erra, a seasoned creative tech entrepreneur working within the world of AI, we gained valuable insights into integrating artificial intelligence tools within business operations.
With over two decades of experience spanning various industries, including e-commerce, environmental conservation, music production, and tech development, the entrepreneur sheds light on the intricacies of incorporating AI effectively.
Hi Giancarlo! Firstly, can you tell us about yourself and your career so far?
“I’d consider myself a creative tech entrepreneur; I think that is the best way of putting it. I started out working in a variety of different sectors, from environmental science to e-commerce to music. Over time, I transitioned into creative and tech roles, eventually founding my own ventures. Over the last few years, I’ve been particularly focused on start-ups in the world of Artificial Intelligence. For me, the allure of AI lies in its fusion of creativity and technology. It offers boundless opportunities for innovation.”
Do you think that companies should embrace AI in their business strategy?
“I think it’s like any technology; it depends on the company and the service they provide. At the moment, there is a very polarized view of AI. My approach isn't necessarily pro-AI or anti-AI; it's about pragmatism. AI is a tool, not a solution in itself. The challenge lies in dispelling misconceptions and understanding AI's potential within a business context. Many businesses misconstrue AI as a one-size-fits-all solution, overlooking its complexities. It’s putting technology before the problem. If you’re thinking about using AI, consider first what problems you actually need solving – and if AI could help second.
Many people use AI without being very literate about what it can actually do. If you want to implement AI into your business, I recommend using it as an internal tool first. This is a safer way to start and a safer way to get to know it. Start by simplifying an internal task, before venturing into customer-facing solutions. This will allow you to get to know the technology safely without the risk of brand defamation. A lot of people just want to cut costs with AI, but the focus should be on improving productivity. There is still cost savings; just a bit more indirectly.”
What do you see as the greatest benefits of businesses embracing AI?
“The benefits – like the risks - are limitless, so it’s hard to define. The best way I see companies use AI successfully is to fulfil repetitive tasks. Automating repetitive tasks with AI is great as they often take the most time and are the least enjoyable to do. Also, AI generally works better at repetitive tasks. This means saving time, and brain power to spend on something more important - or even more enjoyable!
This should come first before trying to save money. It’s a brilliant drafting tool. It could be anything—emails, articles, code, case studies. It’s a great tool for improving productivity, especially internally. Using it this way is not necessarily going to cut costs directly, but it's probably going to improve your productivity significantly. So, there is still a cost saving, it's just a bit more indirect.”
What are the risks of using AI in business, and what solutions do you recommend to mitigate them?
“AI can give you a great foundation to work from; however, this should always be proofread or reviewed before going to your customer. Human input is important. It’s tempting to automate processes completely, but it’s risky. You must always remember that you are responsible and accountable for any output your company generates, even if it’s created with AI.
As an example, AI Chatbots have been the subject of much news recently. DPD recently disabled its AI chatbot after a customer, through clever prompting, tricked it into swearing and writing a poem about how unreliable DPD's service is, calling it ‘useless’.
The challenge is that people don’t understand fully how the technology works before they implement it. You must always remember the technology is unreliable, and so you should be careful when using it for customer-facing services. If you want to use tools in this way, make sure you test and practise thoroughly. You must always remember that you can't trust it 100%.”
What advice would you give to companies who want to start integrating AI into their business operations?
“I think I’ve said it before, and I will say it again: Don’t put the technology first and the problem second. AI is amazing, but it’s not a one-size-fits-all solution. Companies should be pragmatic and make sure they have a thorough understanding of AI’s capabilities and limitations. First, think about what problem you're trying to solve, and then maybe talk with a consultant and see if that is really a problem that can be solved with AI.
Think about the most frustratingly manual process for you and start from there. If you’re doing something at work and you think, “Why can’t this be automated? Or simplified?’. That might be a good place to start with integrating AI.
It’s also worth remembering that current law states that you can’t assign yourself the copyright of something generated by a machine. To copyright something, it must have been generated by a human.”
What regulatory issues should business owners concern themselves with?
“It’s important to remember that there are additional challenges with bringing AI into regulated markets, such as finance or health. As I said before, if you integrate AI into your system, its output becomes your responsibility. It's a black box, so it's not accountable. It’s very hard to go back and see why an AI gives an output it does. Every time you ask an AI something, it will often give you a slightly different answer. Even if you develop your own LLM as an example, you still can’t guarantee consistent results. It’s one of the biggest usage challenges of AI, especially in these markets. If you let an AI decide something critical, even something it has been trained in many times, you still can’t guarantee it won’t suddenly be wrong, and you’re responsible for the result of that.”
Are there any AI tools that you would personally recommend?
“Most companies that use AI are using some kind of OpenAI API, or ChatGPT itself. There’s a reason for this, in my experience, OpenAI is way ahead of other tools. However, if you do, it’s worth experimenting. It’s worth investing in prompt engineering training and experience. If you know how to use it, it is incredibly powerful. If you don’t, you will create very generic text that’s easy to spot. It uses a very specific dialect. Alternatively, Claude AI is a good tool – its language style is less easy to spot than a tool like ChatGPT. In terms of image generation, I always use Midjourney over DallE, but that’s mostly personal preference.
If you only use AI as a drafting tool, it isn’t as relevant which tool you use, as you can always edit in that human element. If you are more advanced as a business, you should consider using custom solutions, such as deploying your own LLM. I’d never suggest people start with that, though – it requires understanding and experience. Of course, you need developers to host your own LLM, and that has a cost – but you can personalise it more to your own requirements. So, you know, that is just a different sort of product. There’s no point in making that investment unless you really need it; it’s always worth testing open tools first.”
What AI business trends do you foresee for the future?
“It’s always hard to predict! GPT 5 will likely be big this year, as ChatGPT updates always are. It will be interesting to see how the tool continues to develop and improve – I think it will have a big impact.
Otherwise, I think Sora will be very significant. It might be because I’m creative, but I think AI Video Generation in general will explode over the next year. There are currently teething problems – much like there were when AI Image Generation first came out and was giving everyone 6 fingers! Audio and Music Generation will likely see huge improvements too, especially in terms and background music such as jingles or loops. Whether we will see AI tools writing tunes as well as, say, Adele, is another story. I don’t think they have solutions for things like that yet – but who knows what’s to come.
In broader terms, I think we could see steps in the development of AGI (Artificial General Intelligence) in the next year, though I think we’re far off a real result. Recent enhancements, such as memory and context integration in ChatGPT models, enable better understanding and interaction with diverse information sources. These advancements pave the way for more comprehensive AI systems that can grasp larger contexts and offer more meaningful responses.
Looking ahead, it's all about integration. It’s about small, significant steps rather than major breakthroughs. Microsoft's Copilot, for instance, is bringing GPT into everyday tools like Windows, making AI more accessible and user-friendly. So, while we might not see any earth-shattering breakthroughs in the coming months, I think the future of AI seems to be all about fine-tuning and seamlessly integrating it into our lives.”
Thank you so much for coming in and talking to us today about your experience and thoughts, Giancarlo; it means so much. We appreciate your insights and look forward to seeing where you continue to go next on your AI journey!
“Thank you for inviting me in. I look forward to seeing how AI Atlas continues to grow!”
___
Giancarlo is the founder of Promethean Box, which has developed multiple AI-focused services, including Words.Tel, AudioShapes.Ai, and Copy Forge AI. You can learn more about his work at giancarloerra.co.
Stay tuned for more industry expert insights across the world of Artificial Intelligence.
Navigating the World of Artificial Intelligence
Copyright 2024 AI Atlas. All Rights Reserved. Registered address: Union Building, Rose Lane Business Centre, 51-59 Rose Ln, Norwich NR1 1BY. UK.
Search
Menu