Close Menu
FintechFetch
    FintechFetch
    • Home
    • Fintech
    • Financial Technology
    • Credit Cards
    • Finance
    • Stock Market
    • More
      • Business Startups
      • Blockchain
      • Bitcoin News
      • Cryptocurrency
    FintechFetch
    Home»Fintech»Getting Business Value from AI: Avoiding a (Gen) AI winter: By Alastair Gill
    Fintech

    Getting Business Value from AI: Avoiding a (Gen) AI winter: By Alastair Gill

    FintechFetchBy FintechFetchSeptember 4, 2025No Comments3 Mins Read
    Share Facebook Twitter Pinterest LinkedIn Tumblr Reddit Telegram Email
    Share
    Facebook Twitter LinkedIn Pinterest Email


    This may well be playing out more publicly because Gen AI technology is much more accessible than old style AI or ML and, because of this, many more people and organisations have been empowered to try it out for themselves. How therefore can those of us
    embarking on (Gen) AI projects build upon and learn from the experiences, successes and failures of AI and ML projects of the past? In this and future articles, I will attempt to bring out some of the key learnings vital for successful AI projects, but first
    let us be specific about what we mean by ‘AI’ and ‘Gen AI’ and, at a high level, how they work.

    In the case of traditional – or ‘old’ – AI approaches, such as machine learning (ML), algorithms work in a specific manner: with appropriate set up and direction, they can output an inference given a set of input data. Building a model in the first place
    is a task requiring skill, knowledge and creativity. However, perhaps the biggest challenge appears to be how to assemble the elements that can be solved using ML in such a way so that they can provide an overall solution to the business problem. Business
    problems don’t often fit an off-the-shelf ML solution. They are not set up like a Kaggle competition, whereby the lucky data scientist on the project has to optimise an algorithm given a neat csv of input data, with the output of 1s or 0s (or 0.2, 0.7, 0.9,
    etc.) generated by the algorithm magically solving the business problem. It is not that such inputs, outputs and intermediate models or algorithms cannot solve a business problem. Instead, it is just that the business use case needs to be correctly identified
    and structured in the first place, and that the resulting output needs to be wrapped in such a way that it can integrate into an overall solution, which then solves the business problem.

    Generative AI is similarly constrained by ensuring that the off-the-shelf solution (foundation model, such as a large language model) can be modified in such a way that the ‘prompt in / text out’ can behave as required for the business problem. Common approaches
    to adaptation include prompt engineering and retrieval augmented generation (RAG), in addition to the various checks and balances that might be required to promote predictability of behaviour and to decrease the chances of undesirable behaviour. Therefore,
    as with ‘old AI’, Generative AI similarly needs to be integrated into an overall solution.

    So, what is that overall solution, and how does it solve the business problem? Even before that, we need to understand what the business problem actually is and begin to get a sense of whether it can be usefully addressed using AI or whether other approaches
    might be more appropriate. All of this is impossible to know ahead of time, but there are things that we can do to get us to that point, with useful steps being to understand the business problem, speak to stakeholders and map the solution to the business
    logic. I will discuss each of these aspects, along with some final considerations for running AI projects, in more detail over the following series of articles.

    Finally, it is worth highlighting that this kind of process is key to the success of any kind of project, but is especially needed for AI projects, given the large amount of uncertainty that there can often be.



    Source link

    Share. Facebook Twitter Pinterest LinkedIn Tumblr Email
    Previous ArticleTrust Bank Marks Third Anniversary with POP MART Partnership, Car Giveaway
    Next Article Bitcoin Withdrawal Wave Points To Another Major Leg Up In The Bull Cycle, Analyst Says
    FintechFetch
    • Website

    Related Posts

    Fintech

    The Payments Infrastructure Powering Digital Competitions: By Paul Clarke

    October 17, 2025
    Fintech

    How Banks Deploy Digital Twins to Speed Property Lending: By Naina Rajgopalan

    October 17, 2025
    Fintech

    4 Surprising Ways Global Finance Has Been Remade Since the 2008 Crisis: By Stanley Epstein

    October 17, 2025
    Add A Comment
    Leave A Reply Cancel Reply

    Top Posts

    The Crucial Ripple (XRP) Metric That No One Is Talking About, and Why it Matters

    August 5, 2025

    Hyperliquid Delists $JELLY, Potentially Causing $900K in Losses. Here’s Why Best Wallet Token Can 100x

    March 30, 2025

    Bitcoin’s ‘KISS Of Death’? Arthur Hayes Warns Of Recession

    March 4, 2025

    Up just 8% in 5 years, what’s going on with the National Grid share price?

    February 28, 2025

    Singapore Banks to Hold or Block Transfers Over 50% of Account Balance From 15 Oct

    October 6, 2025
    Categories
    • Bitcoin News
    • Blockchain
    • Business Startups
    • Credit Cards
    • Cryptocurrency
    • Finance
    • Financial Technology
    • Fintech
    • Stock Market
    Most Popular

    Mastercard Rolls Out Payment Optimisation Platform to Boost Merchant Approval Rates

    October 14, 2025

    The UK Fintech Investment Dips to $7.2B in First Half

    August 11, 2025

    Pump.fun Tokens Drop to September 2024 Lows

    March 10, 2025
    Our Picks

    What Sora’s Martin Luther King Jr. problem revealed to OpenAI

    October 18, 2025

    Everything To Know About The OpenLoot Credit Airdrop This Weekend

    October 18, 2025

    Is Wave 5 Still Coming or a New Bull Trend Emerging?

    October 17, 2025
    Categories
    • Bitcoin News
    • Blockchain
    • Business Startups
    • Credit Cards
    • Cryptocurrency
    • Finance
    • Financial Technology
    • Fintech
    • Stock Market
    • Privacy Policy
    • Disclaimer
    • Terms and Conditions
    • About us
    • Contact us
    Copyright © 2024 Fintechfetch.comAll Rights Reserved.

    Type above and press Enter to search. Press Esc to cancel.