What's the expected opportunity for different sectors and different use cases of AI? A PwC article co-authored by friend of Broadmind, Colin Light, discusses the transformative impact of generative AI across various industries, highlighting its potential to significantly enhance productivity and value creation. Link: https://lnkd.in/e7ZTGAUu I think the section about use cases is helpful for anyone trying to decide which areas of their business they want to apply AI too. 1. Net-new creation: Generates entirely new content from prompts. 33% share of new gen ai value creation 2. Augmentation: Expands existing content, filling gaps or creating synthetic datasets. 21% 3. Transformation: Converts data into new formats or styles. 19% 4. Dialogue: Offers responsive guidance and information through questions. 14% 5. Deep retrieval: Searches specific information within documents. 12% 6. Summarization: Condenses documents or text into shorter forms. 2% While a lot of businesses are focused on developing chat bots and summarization, it's super helpful to see that PwC puts this at the bottom of the chain. That said work on these basics isn't a bad thing, the team there say that because a single AI model can be adapted and tuned for many specific tasks, applying a GenAI pattern to one use case can unlock pathways to similar use cases. The article also outlines a flywheel approach for businesses to leverage #GenAI effectively, which involves: creating a value hypothesis, prioritizing key use cases, identifying patterns to drive scale, selecting foundational GenAI tools, defining solutions to maximize value, assessing cost and carbon impact, and deploying with an emphasis on testing and learning. This methodology aims to help organizations harness GenAI for innovation, efficiency, and competitive advantage, while also considering responsible AI practices (PwC).
Piers Fawkes’ Post
More Relevant Posts
-
Have an idea for a Generative AI tool or feature? Check its viability with those 10 Cost-Benefit Questions first: In pre-COVID times, I developed a proof of concept for a call center analytics system. It recognized emotions and sentiments for all the conversations between agents and clients and provided insights on improvements for managers. Sounds like a good idea, doesn’t it? But in one case it didn’t work. It was a company that was selling cars. Expensive ones. Its call center consisted of a lot of agents and supervisors. At a first glance, if our insights could let them close at least a few deals more - this is 100% win-win. But during the conversations, they shared one straightforward argument: “We have so much at stake with each deal, it’s easier for us to have human supervisors listening to every agent-client conversation live instead of AI”. So, leaving aside markets and competition, check your idea first from pure cost-benefit and product point of view with those question: 1️⃣ Is the cost of AI for a single operation cheap enough to keep the entire idea profitable? (both in terms of $ and overall value) 2️⃣ What is the cost of AI error or inaccuracy? (think of healthcare cases) 3️⃣ Is the value from the correct responses still higher than potential harm from mistaken responses? 4️⃣ Is it possible to minimize chances of mistakes or inaccuracies with extra AI checks and filters? (prompt sequences work fine for most of cases) 5️⃣ If the cost of AI generation is high, is there a way to optimize it down the road? (i.e. by fine-tuning a cheaper model with the collected generations) 6️⃣ If your idea is about assisting in decision making - can’t other people do the same but better and faster? (especially for niche cases, like selling something expensive) 7️⃣ If you automate something with AI - should it really work live or could be done in the background mode? (just send users an email when the task is done) 8️⃣ If you automate something live - is the time to bring some results short enough to avoid user’s frustration? (usually <10s) 9️⃣ Do you need to use the most expensive models like ChatGPT 4 for basic reasoning, or just need to put words in correct order? (cheaper and faster model will do that much better) 🔟 Last but not least: Is it possible to convert your idea into an IP asset? (For example, if you own some unique data, fine-tuning an open-source model may give you unfair advantage) P.S. Half of those questions could be easily answered in weeks by building a proof of concept. Do you check your ideas often? Answer in comments. #AI #GenerativeAI #PoC
To view or add a comment, sign in
-
Powerful use cases outlined for organisations embarking on the generative AI journey with below takeways: - Transformative use cases, across sectors, from pharma to banking to retail, that offer practical benefits for jobs and the workplace already emerging. - Costs of pursuing generative AI vary widely, depending on the use case. Companies must take into account risk issues, and need to have strong governance and data foundation layer in place to ensure success - While there is merit to getting started fast, building a basic business case first will help companies better navigate their generative AI journeys. #genai #digitaladoption #customerservices https://lnkd.in/eFnRU83A
To view or add a comment, sign in
-
Use AI to know how to use AI! If you want to use AI for your business, but are not sure where to start, first identify your needs and what aspects of your business could benefit from AI. Need help with that? Here's an AI Needs Identification Prompt for you to try. Fill in the brackets with your details and paste it into a LLM like ChatGPT. If you get stuck, feel free to reach out. ** I had best results with ChatGPT 4 (paid version) and Claude (free) ** Prompt: "I'm looking to integrate generative AI into my business but need some guidance on where to start. Can you help me identify potential areas in my business that could benefit from AI? Here’s a brief overview of my business operations: [insert brief description of business operations, sectors involved, and any specific business processes]. Additionally, are there specific generative AI tools or technologies you would recommend for these areas?" If setting up AI seems daunting, jump on a free 15 minute consultation call with us, and we can discuss which AI solutions best suit your unique needs. Act Now! Slots are filling fast, and availability is limited. Click here to book your free session and start making smarter decisions with AI. https://lnkd.in/g2vqk7e9
To view or add a comment, sign in
-
Did you know ChatGPT reached 100 million users in just two months? The article below explores the unprecedented speed of generative #AI adoption and its transformative impact on business processes. Read more here ➡️ https://lnkd.in/geKPNJbv #analytics #dataplatform #datamanagement #graphdatabase #datastrategy #data #dataanalytics
Maximizing Business Value with Generative AI - DATAVERSITY
https://www.dataversity.net
To view or add a comment, sign in
-
AI: It's NOT artificial and it's NOT particularly intelligent either! I'm often asked to explain what AI means and how can it be used effectively to support a business. I believe that the term "Artificial Intelligence" is not a useful or particularly accurate definition of what's going on here. I believe it's time to redefine AI. Using the same initials, I suggest it might be better to call it "Augmented Integration". So let's start with the "artificial" part of AI. Information Technology will always need to be managed by people, even if it is outsourced to a technology provider. Yet surrendering too much of your technology decisions to a supplier is not a good idea. It's Unintelligent. Just as the word Virtual Reality morphed into the term Augmented Reality, I think it's time for the "Artificial" part of AI to morph into "Augmented" - demonstrating that the bots aren't going to take over completely - and that it's time for humans to take control of the whole migration and use the technology in creative ways to augment rather than capitulate to the doomsayers. And that brings us to the "intelligent" part of the two-letter acronym. The opportunity to use and benefit from AI technologies continues to be on the interfaces between processes, systems and datasets. After all, AI is just the next step in the trend seen over the past several decades to manage, manipulate and get more value out of data sets. And that value will always be on integrating various views and data sets within and across data domains - either within or outside of an organisation. Given that over 50% of the spend of most IT projects invariably goes into integration, it is INTEGRATION which is the key word here. By redefining what the letters mean, we can get clarity on where the opportunities lie, allay the fears of the bots taking over and focus our attention on integrating this amazing technology shift to make the world a better place. Or maybe you have an even better re-definition? If so, please suggest it below/
To view or add a comment, sign in
-
Did you know ChatGPT reached 100 million users in just two months? The article below explores the unprecedented speed of generative #AI adoption and its transformative impact on business processes. Read more here ➡️ https://lnkd.in/gmTdaTvN #analytics #dataplatform #datamanagement #graphdatabase #datastrategy #data #dataanalytics
Maximizing Business Value with Generative AI - DATAVERSITY
https://www.dataversity.net
To view or add a comment, sign in
-
Founder & CEO of KYield. Pioneer in Artificial Intelligence, Data Physics and Knowledge Engineering.
Excellent article by Sharon Goldman focused on generative AI through the eyes of Margo Argenti, CIO of Goldman Sachs. This provides a nice contrast to the hype we typically see and I think is more representative of the actual behavior in most large conservative companies. Most large companies are 1-2 years away from adopting GenAI, though many are performing tests like GS. Several common themes I often discuss in my EAI newsletter here on LinkedIn, including: 1) Inherent problems with LLM safety and accuracy. 2) Instant commoditization in LLM chatbots, meaning a competitive advantage must come from other types of AI functions, optimally in an EAI OS like our KOS. 3) Generative AI provides a new baseline for text-based work in the digital workplace, when combined with voice bots and other apps will change the nature of work for many, the type of services companies offer, and how they are delivered. Bottom line is the super majority of companies need safe and secure systems with strong governance. That requires either a great deal of thought and significant investment within large companies, adopting highly refined systems like our KOS, or a combination of the two. Very few outside the largest companies will be able to build internal custom EAI systems in the foreseeable future. #generativeai #ai #eai
Goldman Sachs CIO is 'anxious to see results’ from GenAI, but moving carefully | The AI Beat
https://venturebeat.com
To view or add a comment, sign in
-
Please see this article at VentureBeat by Sharon Goldman about #generativeai, and how Goldman Sachs is approaching it.
Founder & CEO of KYield. Pioneer in Artificial Intelligence, Data Physics and Knowledge Engineering.
Excellent article by Sharon Goldman focused on generative AI through the eyes of Margo Argenti, CIO of Goldman Sachs. This provides a nice contrast to the hype we typically see and I think is more representative of the actual behavior in most large conservative companies. Most large companies are 1-2 years away from adopting GenAI, though many are performing tests like GS. Several common themes I often discuss in my EAI newsletter here on LinkedIn, including: 1) Inherent problems with LLM safety and accuracy. 2) Instant commoditization in LLM chatbots, meaning a competitive advantage must come from other types of AI functions, optimally in an EAI OS like our KOS. 3) Generative AI provides a new baseline for text-based work in the digital workplace, when combined with voice bots and other apps will change the nature of work for many, the type of services companies offer, and how they are delivered. Bottom line is the super majority of companies need safe and secure systems with strong governance. That requires either a great deal of thought and significant investment within large companies, adopting highly refined systems like our KOS, or a combination of the two. Very few outside the largest companies will be able to build internal custom EAI systems in the foreseeable future. #generativeai #ai #eai
Goldman Sachs CIO is 'anxious to see results’ from GenAI, but moving carefully | The AI Beat
https://venturebeat.com
To view or add a comment, sign in