A spectre is haunting AI – the spectre of hype. AI alludes to human capabilities. But most “AI solutions” don’t get near having those. Nor do they answer key questions: what value can AI add to a business, how and what are the costs and risks?
DigitalBiz’s new videos help owners bypass the hype and explore a low cost approach to clearly demonstrating the potential of AI in a business. This combines AI with human experience to create an AI solution fine tuned to a specific use case.
To view the videos, visit https://digitalbiz.ai
Major breakthroughs announced daily suggest the “AI train” has left the station and businesses must scramble to get aboard or be left behind. Whilst doing nothing is not an option, in reality, AI is an umbrella term which oversells itself and can mislead wary managers into high deployment failure rates.
Most practical use cases of AI improve existing business processes in straightforward ways. The term “AI” suggests an end to end process but that is not how it works.
AI is an array of focused technologies, each with advantages and uses. They have a high level of precision, remarkable speed, the ability to learn from mistakes and a reliance on humans to learn from.
Until an AI goes down the human experience learning curve, it will make mistakes. An AI needs time to soak up human experience in a business until it is fine tuned for a specific use case.
The problem for the business owner is making the decision to get involved, identifying the use case and selecting which AI to use. Ignore AI and the business will get left behind. Select the wrong AI and the impact could be expensive. Get it right, and productivity gains of 40% plus could result.
Start with the use case. Identify mundane business tasks which require data input, are repetitive or involve prediction. AI will likely do such tasks better, faster and cheaper.
Select a task and tackle it with a quick, simple AI pilot test. Snags will be encountered. If the AI can’t cope, slot in a human to fill the gap. Compare the before/after performance to calculate the gain. Then repeat.
All businesses are different. But most will have separate, distinct, data driven, repetitive and prediction tasks as part of a process. Use a variety of applied and generative AI’s to implement the tasks. Use humans to help the AI’s up the learning curve and to link gaps in the process. AI’s gradually learn from human edits and the level of process automation will increase.
Practically, of the 3 types of AI – Machine learning, applied and generative AI – most businesses use the last two.
Machine Learning algorithms learn from exotic data produced, say, by nuclear fusion or dark matter analysis, is hugely expensive and very risky with a high failure rate.
Applied AI are solutions pre-trained on tasks like marketing analysis or supply chain processing. On deployment, they’ll tackle general tasks but must then learn from staff corrections to fine tune performance in that particular business.
Generative AI (like ChatGPT), creates new content which didn’t exist before. Initially, it throws out data in the form of word, image or sound combinations but doesn’t know if it is accurate, relevant or a waste. Gradually, it learns from staff edits to maximize accuracy and relevancy, until it is fine tuned to that particular use case.
The lesson for all adopters is AI is not yet ready for prime time. Applied AI is brilliant at executing data driven, repetitive or predictive tasks. Generative AI creates original content. But both need time to learn from human experience and fine tune performance.
For the next few decades at least, the future is combining the best of AI capabilities with the best of human experience. The following example combines Applied AI, Generative AI and human insight to automate common content planning and operations processes.
Tasks include content planning, coordination, analysis of hot topics, content performance prediction, creation, publishing and distribution. Except for content creation, all tasks are data driven, repetitive and predictive. An Applied AI will do them – helped along by human corrections so it can fine tune itself to business needs.
ChatGPT creates the content but lacks emotional understanding, It learns from each human edit until it is finely tuned to produce content which genuinely connects with people and maximizes engagement.
In summary, the applied AI – with human corrections – creates a plan of hot topics to write about. Hot Topics feed ChatGPT which generates unique content then soaks up human edits until it is able to mimic the writer’s tone, style and emotional understanding. It is now fine tuned to produce future content. The Applied AI – with human help – processes the content, publishes to multiple platforms, and consumers engage.
This is not an “end to end” process as the term AI implies. Invariably, humans must join up individual tasks to create the process. But productivity gains of 40% plus and significant resource savings are not uncommon.
Adrian McKeon, DigitalBiz CEO states ” Don’t go overboard or make a huge financial commitment. Pick a job that needs data entry, is repetitive or predicts the future. AI will likely do it more proficiently and cost-effectively. Do a small pilot test. Compare the before and after performance – this is the use case. Move to the next task. Don’t forget about people – the more AIs learn to mimic staff the greater the business value.”
Using DigitalBiz, smaller companies can take advantage of the advanced AI technologies and services typically only utilized by large organizations. They’ll get their hands on content production, organization, and dissemination capabilities without paying through the nose or investing a huge amount of time. It’s an efficient and affordable way to start leveraging AI for business success.
Interested parties can access the newly launched videos at https://digitalbiz.ai
Initial Business Centre
207 Regent Street
Disclaimer: The views, suggestions, and opinions expressed here are the sole responsibility of the experts. No Sahyadri Times journalist was involved in the writing and production of this article.