.png)
Note: This article is an extended version of an op-ed published in French in Stratégies.
Far from being another technological innovation, Artificial intelligence is bound to deeply transform our organizations and, with them, a large portion of intellectual work.
In many professions, this transformation is already underway. Developers spend less time writing code and more time supervising AI agents. In ad creation, generative models produce images, text, and videos in minutes. In marketing, media buying and campaign optimization are increasingly automated. The question is no longer if AI will transform businesses, but how.
How do you move beyond proof-of-concepts and pilots? How do you select the right use cases? How do you reshape workflows in practice? And most importantly, how do you get hundreds or thousands of employees to change how they work? Clients turn to consulting firms and agencies for answers to these exact questions. Irony abounds: those advising others to transform are often far from embarking on their own journey.
Our industry will be among the first affected. Historically, a data consultancy sells services delivered by human consultants. Tomorrow, those services will be delivered by hybrid teams: humans working side by side with AI agents.
This shift challenges a deeply rooted model. If a task that used to take days can now be completed in minutes by an AI agent, its market value cannot remain unchanged. But treating AI merely as a cost-cutting tool misses the bigger picture. It also improves quality, standardizes expertise, and enables tasks that were previously too time-consuming or expensive to undertake.
The goal isn't simply replacing humans, but redesigning services around what humans and machines each do best. At fifty-five, we have been supporting brands through data and AI transformations for fifteen years. We decided to take our own medicine. That experience taught us one key lesson: an AI transformation doesn't start with AI.
It rests on three pillars: a business vision, a culture of measurement, and human change management.
The first mistake is asking, "What can we do with AI?"
We prefer a different question: "If we were starting this business today with AI at our disposal, how would we design it?" This forces a focus on business processes. Take a practical example at fifty-five: deploying web analytics tools like Google Analytics. It is a core offering with a proven methodology, from requirement scoping to ongoing maintenance.
We broke down our methodology step by step. For each step, experienced consultants and AI experts asked three simple questions: Where do we lose time? Where do errors occur? Where could a machine deliver better, faster, or at a fraction of the cost? Each potential use case was then evaluated against expected ROI and development cost. The result wasn't limited to adding a chatbot onto existing processes: instead, we established a brand new production model.
Following this logic, we built TrackSense, a suite of ten AI agents that assist consultants throughout the project lifecycle. The goal is to help consultants deliver faster, more accurately, and more efficiently. This distinction is vital, since too many companies simply overlay AI onto existing workflows. True transformation begins when you redesign the underlying process.
The second key lesson revolves around measurement.
I often meet AI leaders facing the same hurdle: they have launched dozens of pilots, users seem happy, demos are impressive, and positive anecdotes circulate. But when it's time to scale up, momentum stalls. Indeed, scaling requires significant capital, IT system integration, training thousands of employees, and ongoing maintenance. The CFO inevitably asks a fair question: "What is the return on investment?" But too often, nobody has a clear answer.
This is what I call pilot purgatory: a project successful enough to keep running, but lacking the hard metrics required to justify enterprise rollout. Measurement must be established before building the solution. Are we targeting a 30% reduction in production time? Lower error rates? Higher client satisfaction? Increased conversion rates? What is our baseline, and what target performance justifies the investment?
An AI pilot shouldn't simply prove that the technology works — it must prove that the business case works. At fifty-five, we enforce this rigor on ourselves just as we recommend it to our clients: every use case needs success metrics. This allows us to double down on what works and ruthlessly sunset what doesn't.
Then comes the third and arguably most critical pillar: the human factor.
You can build the best AI agents in the world, but if employees stick to old habits, transformation fails. AI can spark understandable anxiety; hearing constant predictions about job displacement leaves employees wondering about their future role. In the near term, however, the real dynamic is slightly different: we won't be replaced by AI as much as outcompeted by peers who master it.
AI raises the bar for quality. Subpar presentations, rough translations, superficial analysis, or manual repetitive tasks are far less acceptable when tools can produce higher-quality output in seconds. Standards are rising across the board, making AI adoption essential for everyone.
AI shouldn't be reserved for a handful of specialists; it must be an inclusive, company-wide initiative where employees actively drive their own workplace evolution. That's why fifty-five combines two complementary approaches.
The first is top-down: redesigning core methodologies and developing supporting AI agents (as detailed above).
The second is bottom-up: empowering consultants to identify daily pain points and invent solutions. In internal hackathons, consultants collaborate with AI engineers to address their own operational challenges. A recent outcome automated reporting slide creation—reducing repetitive work, lowering error rates, and freeing up time for high-value analysis and advisory work.
Adoption happens when AI moves from abstract strategy to solving concrete day-to-day pain points.
This shift will inevitably change consulting economics. Certain traditional services will lose significant market value, a trend already visible today.
A web analytics implementation audit used to require dedicated senior consultant time. Now an AI agent can execute much of that work at near-zero marginal cost. At fifty-five, such audits can be offered up front as a value-add rather than billed as a standalone deliverable. Far from destroying value, this shifts where value is created. And as execution becomes automated, human value shifts toward problem framing, critical judgment, solution architecture, strategic alignment, creativity, relationship building, and accountability. In short, toward the true essence of consulting.
The line between consulting firms and tech companies will continue to blur. Firms won't provide talent and advisory decks; they will combine human expertise, proprietary software, and specialized AI agents to deliver outcomes. Business models will evolve accordingly.
AI offers consulting a major opportunity: relieving consultants of repetitive tasks, scaling domain expertise, and delivering higher-quality, faster, and more cost-effective services to clients.
Yet it presents an uncomfortable question to our industry: can we convincingly advise clients to overhaul processes, measure ROI, upskill teams, and evolve business models if we aren't willing to do the same internally?
At fifty-five, our transformation is an ongoing journey, as technology continues to rapidly evolve. But we abide by a simple principle: never advise a transformation we aren't willing to undergo ourselves. In the era of AI, the best way to understand how to transform an enterprise is still to start with your own.
.png)
Discover all the latest news, articles, webinar replays and fifty-five events in our monthly newsletter, Tea O'Clock.