Reading Time: 3 minutes

Business leaders can’t stop talking about generative AI. The noise is loud, the pace is quick, and the choices can feel endless. Each week in “The Prompt”, we share what we’re hearing from customers and partners, alongside the latest AI updates at Google. In this edition, Philip Moyer, Global VP, AI & Business Solutions at Google Cloud, and Carrie Tharp, VP, Strategic Industries at Google Cloud, explain why starting with simple use cases often delivers the fastest, clearest wins.

From Google I/O last month to our Executive Forum last week, plus many events in between, we’ve had constant conversations with customers as our generative AI rollout gathers pace.

Over the past month, those conversations have changed. Earlier in the year, leaders were still wide-eyed at what the tech could do. Now, they want practical answers, mainly which use cases can move into production quickly.

Unless a business already has deep machine learning and data science roots, most leaders are aiming for straightforward projects like these:

  • Building chatbots and digital experiences for self-serve customer support and product discovery
  • Generating marketing content
  • Organising internal knowledge, from next-generation enterprise search to workflows for drafting and reviewing routine documents, including RFIs and RFPs
  • Creating personalised content for account-based marketing and other B2B work

This list isn’t exhaustive. Still, it shows a clear pattern. Many teams want high-value, lower-risk solutions they can ship soon, rather than chasing the most talked-about future visions.

That approach makes sense.

Big bets matter. AI has inspired more ambitious ideas than ever, and every large organisation should keep a few on the roadmap. For some start-ups, the big bet is the whole plan.

Even so, anyone following AI online has seen the loud claims: “Generative AI will 10x productivity, don’t get left behind!” That hype can push teams into targets they can’t measure or meet. In contrast, practical use cases don’t need to reinvent everything to deliver real progress.

Take near-term applications like code completion or text-to-code. For most organisations, efficiency gains closer to 10 percent than 10x are a better starting point. Larger gains may come with time because the technology is still young. Yet every shift in tech follows a familiar rhythm, people need room to learn before they sprint. Starting, testing, and measuring helps. Chasing headlines doesn’t.

It’s encouraging that these common starting points line up with the guidance we shared earlier in this series in the second instalment. Put simply, focus on helping customers (for example, self-service chatbots), making internal data more useful (for example, combining enterprise search with generative AI), and reducing repetitive work (for example, producing multiple versions of a marketing message, or speeding up RFP creation and processing).

In addition, it often pays to begin internally before going customer-facing. For example, when an organisation gets confident managing generative foundation models for enterprise search scenarios, it also builds the skills needed for public apps and websites.

Because organisations vary in maturity and requirements, they need different onramps to generative AI. That thinking shapes how we’ve built our platform. Products like Google Workspace bring AI into everyday workflows. Generative AI App Builder reduces the effort of building common generative apps such as search and chatbots, so teams can move faster. On the other hand, generative AI support on Vertex AI provides advanced tools to access, customise, and manage foundation models in production for bespoke generative applications. Alongside that, a range of infrastructure options and partners helps tailor solutions to different needs. As we shared last week, these capabilities form a framework that supports early projects while keeping options open as needs change.

Starting simple is often the quickest route to real value. Even modest use cases can deliver measurable results, and they also lay the groundwork for more inventive work in the months and years ahead.