Inside the tornadoFrenetic is one way to describe the current zeitgeist in Silicon Valley as firms react to the disruptive potential of agentic AI. But there isn't really a word in English to describe the environment product builders find themselves in today. So we invented German one:Rauschverwirrung,a heady mix of exhilaration and confusion.According to Adobe Adobe’s Experience Platform SVP, Sundeep Parsa, tech firms have little choice. and itis as much an organisational culture issue as a technology one: “The way we are framing our questions is no longer about speeds and feeds of the technology, how fast it can go. We have solved this problem. Organisational agility is the next frontier. You’ve got to think about the people process… What’s your end-to-end intake process? How many approvals and steps are needed? Are you open to some simplification?”Parsa saysAdobe is reimagining its Experience Cloud “with an agentic AI-first mindset,” connecting data, content and journey orchestration so applications can reason about goals and context. Creativity is connected to productivity and customer experiences: coordinating agents that understand intent, tone and outcome rather than executing isolated tasks."Sam Allen, the CEO of cross-channel martech firm Iterable, describes in simple terms the kinds of changes marketers can expect as the stack agentifies. "In order to see what works, you need to spend a lot of time with analysts, looking client into the data, understanding cohort analysis. This is what's amazing about AI ... it be able to do all that hard work for you and then tell you how it did it."His firm has been rebuilding its proposition based on replacing the old model: campaigns that were efficient in the way assembly lines are efficient: one message, stamped out a million times, dispatched to audiences defined by blunt demographics. Even as segmentation and cohort models became more sophisticated, the rhythm stayed episodic: plan, build, blast, repeat. But now, as data analytics has become more sophisticated, and with the rise of generative and agentic AI, those old heuristics are breaking down."I do believe if you are stuck doing traditional campaign activity, you are going to get left behind."Rearchitecting requiredRearchitecting the way marketing operates, shifting from a campaign-led cadence to an always-on agentic-powered learning system, will require an overhaul of the whole stack, not just the cross-channel marketing slice that Allen and his team want to remake.There's a debate over over quickly this will happen. George Colony, the Forrester founder and CEO, and a tech analyst for over 40 years, is firmly in the "revolutionary" camp. He told Mi3 earlier this year he sees serious disruption ahead as agentic AI rewrites the economics of software.It's an argument about which he has elaborated on the Forrester blog site under a column headlined "The Seventh Wave"According to Colony, "Software systems, be they CRM or ERP, are structured databases — repositories for client records or financial records. Generative AI, coupled with agentic AI, holds out the promise of a new way to manage this data, opening the door to an enterprising generation of tech companies that will offer AI CRM, AI financials, AI database, AI logistics, etc. These systems offer the promise of being much more adaptable and learning-focused, as well as being easy to deploy and with a lower cost to deploy."He also argue that agentic AI will provide better functionality."AI-native systems will continually learn and flex, and adapt without millions of dollars of consulting and customisation. They hold the promise of being up to date and always ready to take on new business problems and challenges without rebuilds. When the business and process changes, the tech will learn and change."Colony advice to clients suggests he sees rapid change."If you are looking to buy a traditional SaaS platform in the next two years, wait,” he advises. “Or sign very short-term deals. Do not make big commitments."Scott Brinker, the martech sector's best known biographer, sees huge transformations coming – he talks about a world where billions and perhaps trillions of software agents blink in and out of existence as they are needed. But he is more of an evolutionist than Colony.One critical restraint is the capacity of organisations to absorb change, what Brinker calls Martec's Law."Technology changes exponentially. Organisations change logarithmically."It’s a formula for strain."When you put these two curves against each other, we’re in a world where technology is changing exponentially, but the pace of change within organisations is much slower. Those curves keep drifting further apart. I always picture it like having one foot on the dock and one foot on a ferry as it pulls away."It is, he says, the quintessential challenge of management in the 21st century.Rapid reskillingShifting to an agentic martech stack will not be a matter of rip and replace. Instead, it will require a fundamental rethink of skills and capabilities for both brands and their martech suppliers.Take the simple issue of prompting, a very important issue in an age where the prompt becomes the interface.According to Hightouch co-founder and co-CEO, Tejas Manohar, “Most of the prompts in the future are going to be computer-generated.”He sees a world where machines generate prompts automatically from data insights, turning marketers into reviewers rather than creators.“If, in the future, your content team is just going into a content AI tool and typing in prompts based on their ideas, you’re missing a lot of the value. You can now programmatically connect your data and insights layer to your content team, so the insights system can generate prompts for the content AI system to create more interesting variants.”While the headlines are full of stories of AI-influenced tech-layoffs, the reality is that Silicon Valley’s tech firms are investing heavily in preparing their staff for an agentic future.When we interviewed Wade Chambers, Amplitude’s Chief Engineering Officer, in late May, he was preparing to host the firm’s AI Week for all staff.It’s a commitment to capability that goes missing all too often in Australian companies.Chambers describes the company’s internal training regime as a deliberate attempt to “get everybody all the way to committed. There is no going back. It’s a horrible metaphor, but let’s burn the boats. We have to get on the other side of this.”To that end, Amplitude has instituted five days of immersion where theory gives way to practice. “Day one, it’s just everybody has the same experience… why AI first? How does all of this work?” he says. The following sessions move from classroom to code: “Have three of our most senior leaders live code something together, and maybe actually have different roles inside of it – the designer is now the engineer, the product manager is now the designer, the engineer is now the product manager.” The idea is to make every role interchangeable, at least temporarily, to force cross-functional understanding of how AI changes the creative and technical workflow.The week culminates not in exams but in performance. Teams present what they’ve built, describe “what we learned,” and share “where you had challenges and how you overcame it.”Chambers calls this “practice, practice, practice" approach an exercise in collective muscle memory. “Don’t think your way into a new way of acting. Act your way into a new way of thinking,” he says, quoting an old maxim. The goal is to normalise experimentation, make it “comfortable to talk about” mistakes, and embed fluency across disciplines before autonomy scales. In his words:“The more I can get people to practice... the quicker that can be discovered and folded back into the system, the faster we go.”