GP Bullhound’s Post

Some thoughts

Discover how one major financial institution took their analytics factory from utterly unscalable to state of the art. Learn about their #AI and #analytics transformation and how they increased productivity by 1200% using HPE Ezmeral Software. https://hpe.to/6040e4pZ0 #HPESoftware

By 2027, almost all providers of software designed for finance organizations will incorporate #AI capabilities to reduce workloads, says analyst Robert Kugel. Read this perspective to learn more: https://bit.ly/4a4LIpC #LLMs #GenAI 

⭐️ Remember Deepdesk AIx? In a recent post on the principle "garbage in, garbage out", we touched upon the limitations of AI and its dependency on quality input data. But what if we told you there was an exception to this rule? That's right—AI has evolved to the point where it can sift through what appears to be disorganized, scattered data to find valuable information... Remember, Deepdesk AIˣ. 🧐 https://hubs.li/Q025zPTM0 #ai #agentassist #llm #technology #saas

“A manufacturer can be dealing with potentially hundreds of different systems, all of which are producing their own data,” says Avanade’s Brendan Mislin.    In the latest issue of Technology Record, Alice Chambers speaks with Mislin who reveals why it is difficult to track industrial data across multiple applications and how generative AI and analytics tools are helping to improve operations.    Read more 👇    #technology #manufacturing #AI 

"The build vs. buy dilemma in the age of AI models is indeed a complex one to navigate." - I don’t know if it’s so much complicated as it is uncertain. There’s a strong, justified dialogue that “every company will become an LLM company” in the way every company became a software company, but we’ve seen in the last 4-5 generations of technological leaps - personal computer, internet, cloud, smartphone/4-5G, big data /machine learning - that the vast majority of companies didn’t becomes {technology} companies. What we did see was: 💣 Disruptors displacing incumbents who were slow to recognize how the technology impacted their industry (see:Uber vs. Taxi's) 🧪 Localized teams inside of engineering/r&D/IT focused on leveraging that technology internally (see: Data Science / Data Analysts for big data/machine learning) --> having spoken to many leaders with in-house data science teams while at MadKudu, I can confirm that as many regret the decision to internalize as are leveraging it effectively to differentiate from competitors. ( Rodrigo Veríssimo maybe you want to weigh in here 😏 ) 🤝 Having internal resources working on such technology does not end up being incompatible with purchasing from vendors leveraging the same technology - IT teams today both build internal tooling and also evaluate third party vendors, with the prevailing assessment coming down to "is it strategic/expensive enough for us that we want to commit to maintaining this until the next innovation cycle?" What you build vs. what you buy often turns out to be very different than prevailing sentiment - we used to say "build what's key to your business, buy the rest" but payments are key to most businesses and we don't build that internally, as is reliability but we all rely on cloud providers. Given that, combined with the prevailing thinking in 2024 thinking that enterprise use cases around LLMs boil down to "give an LLM access to your internal knowledge base/customer data, we promise nothing will leak," it's no wonder that many organizations are investigating tools but slow to adopt. LLM-centric SaaS startups that aren't system of record incumbents leveraging the customer data you've already committed to their tool (see: Zendesk/Intercom) will need to demonstrate they can create value without sucking up confidential information, exposing it to all employees - with or without prompt injection security, companies are afraid of such a new technology. It goes without saying, but this is why we are focusing Waypoint AI on creating value on top of existing models without a prompt UI (except for super admins), without requiring fine-tuning on your internal data, and integrating with existing systems of record that already host the data.

Everywhere we turn, we hear more and more about #AI. It’s even making its way into #rental software – and the impact is only going to grow. From predictive maintenance scheduling to dynamic pricing, Anthony Tye discusses some of the ways it could shift the industry in the coming years as technologies continue to evolve: https://lnkd.in/eMpgRD_w

Why you’re thinking wrong when choosing between Custom AI-Driven Software vs. Off-the-Shelf Solutions. Frequently, when senior management opts to deploy a ready-to-use data-driven software that will power their business, they anticipate it will be more cost-effective, faster to deploy, and hustle-free, often overlooking the compromises involved. Here are a few cases when you should go with the custom solution: - When standard solutions lack requisite features and fail to leverage generated or imported data: -- Off-the-shelf ML tools, while suitable for specific applications, often demand extensive data preparation and still impose constraints on incorporating data from new customer touchpoints, product and business lines – thus hindering business agility  -- Any off-the-shelf solution still needs to be productized to fully harness its potential as a business - When existing solutions are incompatible with the scale or stage of your company, particularly in scenarios involving rapid growth or numerous API integrations - When an emerging market opportunity underscores the need for swift implementation of high-quality, data-centric operations: For instance, amidst the onset of the COVID-19 pandemic, many companies recognized an opportunity to penetrate markets where competitors were downsizing. However, to capitalize on this opportunity, they urgently needed to establish robust, data-driven processes - When off-the-shelf solutions fail to meet the industry compliance standards or demand an excessive amount of effort to achieve compliance. It is advisable to develop a meticulously documented infrastructure with transparency and integrated compliance mechanisms at every level of the product - When integration with legacy systems demands careful design to ensure compatibility, scalability, and minimal disruption to existing operations - In instances where you can’t use internal resources -- Your internal data team is taking care of day-to-day operations and planned product advancements -- When the team lacks the requisite expertise for AI or ML-driven software development -- When leveraging integration partners' technology stacks is necessary   If any of these cases resonate with your company's situation, we invite you to connect with us to explore the potential opportunities ahead. https://lnkd.in/dfC-eBfZ We excel at building long-term product strategies, helping with operating and maintaining products, focusing on business outcomes in a partnership-like collaboration, and making the business more cost-efficient, scalable, and sustainable. For further details, please visit: https://lnkd.in/d5wN5UjV #ai #productdevelopment #machinelearning