Start with what you own and rebuild around your people.
Most companies are working to transform faster, and few have stopped to ask whether transformation is still the right goal. Transformation assumes the way a company operates today is worth keeping, just run better. The companies pulling ahead are setting that assumption aside entirely and building what replaces it, either alongside the business or in its place. The economics increasingly favor rebuilding, more than most executive teams assume. None of it works without the people who understand the old model best, since they are the reason the new one can move fast.
Picture a management team that spends a quarter building a three-year roadmap for AI-native competition. By the time the board approves it, a competitor a fraction of its size has already built what the team was planning for year two. That is becoming normal for any leadership team still planning in years while its sharpest rivals build in weeks.
The old rule held for decades: whatever a rival might eventually build would take years to reach the market, giving the bigger company time to see it coming and catch up. That rule depended on new capability costing the same time, money, and talent it always had, and AI has cut all three. In BCG’s January 2026 AI research, the fastest-moving companies, roughly 15% of the field, had put nearly three-quarters of their AI budget to work, while the rest committed about a quarter and waited to see if it paid off.
Legal and customer service show what that gap looks like in practice. Harvey, a legal AI company founded in 2022, is now valued at $15.5 billion and used by about 80% of the largest U.S. law firms, the incumbents it was built to challenge. Sierra, co-founded by Salesforce’s former co-CEO, reached a $15.8 billion valuation doing customer service work that used to run through the enterprise software he helped build. Neither company had an old model to protect, and that advantage is already showing up in their valuations.
Map the Choices From Adding AI to Starting Over
The decision in front of most leadership teams is whether the business becomes AI-native on a timeline it sets for itself or one the market sets for it later. New entrants do much of the setting. They compete for the same clients, and they are also setting a new standard for how individual functions, and whole businesses, can run. As that standard advances, more of what has always been done the traditional way looks outdated, and choosing transformation because it is the familiar route can set an organization back.
The choice is rarely one decision for the whole company. It is made part by part, and different parts of the business can reasonably be at different points along a range of options.
The first is to start with what you own. Most enterprise systems now come with several kinds of AI built in, and many organizations pay for more of it than they use, so putting that capability to work is the quickest and least expensive step available. Earlier in my career, inside a global professional services firm, I worked with colleagues to assess, build, and continuously improve our marketing technology stack. We kept its footprint as small as the work allowed, which stretched the budget and kept the tools simple enough that people used them. The same principle applies to AI, where each new tool adds cost and something else to learn. Starting with what you own also shows where the existing way of working limits what AI can do, which is often the clearest sign of which parts of the business are ready to rebuild.
Further along, a company builds something new beside the existing business, which lets it learn without disrupting what pays the bills, though the new model still answers to the old org chart and incentives. Further still, it separates the new model from the old so it starts free of the old one’s constraints. At the far end, it starts over from scratch, AI-native.
Starting over can apply to a single department or function as readily as to a whole business. It suits a department that has worked around the same problems for years and would gain more from a fresh design than from another round of fixes. It suits a function that has fallen behind and can use a fresh start to leap ahead. And it suits a function inside a large enterprise that has grown heavier than its work requires and could run leaner and faster.
For a large enterprise, the same choice can be made one function at a time or across several at once. Rebuilding one function at a time keeps each change manageable. Redesigning across functions together deserves equal consideration, especially where business units operate as separate silos, since a change confined to one unit leaves the handoffs between units as they were. In IBM’s May 2026 survey of 2,000 CEOs, organizations that had redesigned five core areas of the business (technology, finance, HR, operations, and collaboration across functions) were four times more likely to have delivered on their business objectives.
That is why starting from scratch, AI-native, belongs on the table alongside traditional transformation for every part of the business under review, priced side by side against the same goal. Every option short of a fresh start still carries the old model’s systems, habits, and assumptions, which have to be unwound before anything new can fully run. A fresh start skips that step, and that is a large part of why the companies furthest along have moved fastest.
“Choosing transformation because it is the familiar route can set an organization back.”
Learn From the Companies Already Rebuilding
Sabre spent years planning its rebuild before showing it in public. In March 2026, the travel technology company unveiled what it called a once-in-a-generation reconstruction of its technology and operations, built on a new platform named Mosaic. It also changed its brand identity, a step few companies take, signaling that the company underneath the logo was no longer the one that had carried the old architecture for decades. That is what pioneering looks like when a large, established enterprise commits to it. Not every rebuild works, which is a reason to go in with clear eyes and a plan for the risks.
Lemonade shows the same kind of model working from the start. The insurer was built AI-native from its founding, with AI running underwriting, pricing, and claims. In the second quarter of 2026, it spent 5% of premiums on handling claims, a record low for the company and roughly half the industry average of about 9%. Sabre chose to leave its old model behind and Lemonade never had one, and both show what an operating model built around AI can do.
Understand Why the Economics Favor Pioneering
The instinctive objection is cost. For decades, rebuilding an operating model was the expensive choice next to gradually improving what was already running. AI has changed that calculation, and there is a second cost most executive teams overlook: what it takes to run AI at scale, an expense that arrives every month after implementation. Layered on top of a multi-year transformation budget, that combination is what makes AI feel unaffordable to many CEOs weighing it. Pioneering removes the transformation half of that equation and leaves room to manage what AI costs to run, with one bill to carry.
Organizations that have been slower to adopt AI, or whose data is still scattered across systems, often assume they are furthest behind. Measured against a fresh start, they carry the least to unwind: fewer systems wired into the old way of working, fewer habits to retrain, and less money already spent on a model headed for replacement. Organizing the data is still part of the work, and it is often simpler to do from scratch than to untangle from years of earlier projects.
The platform and operating model I am building started from a blank page: what marketing and sales would look like for a mid-sized or enterprise services firm if none of its existing infrastructure existed, and what building that way costs compared with transforming what is already in place. For a mid-sized firm, that work points to millions of dollars in cost savings, with significantly less complexity and time than a traditional transformation. The organizations that see this first are setting the economics the rest of their markets will eventually have to match.
74% of AI's Value. 20% of Companies.
PwC surveyed 1,217 senior executives across 25 sectors. The companies capturing AI's value redesigned how work gets done around AI, rather than adding more AI on top of how they already worked.
Redesign Roles Around What Your People Know
Institutional knowledge is what makes pioneering fast. The people who understand the old operation’s real constraints, including what clients needed even when the process never asked for it, are the best placed to help design what replaces it, once they are given a role in doing so.
Treating the headcount tied to the old model as a cost to retire along with its systems can turn a sound decision into an outcome no one wanted, and Klarna offers a useful lesson. Its AI assistant took over a large share of customer service work, and the company let its workforce shrink, mainly through a hiring freeze and attrition. Its CEO later said cost had become too dominant a factor, and the company began recruiting people for customer service again so customers could always reach a person. A new model aimed at cost alone, with no plan for where people’s expertise goes, produces a result even the company making the change does not want to keep.
Culture is the other piece to design. Habits built over a decade stay in place when only the technology underneath changes. Pioneering is the chance to set new defaults from day one, an AI-first way of operating built into how the place runs, carried forward by the same good people.
In the strongest new models, operational expertise moves to where it adds the most. People who once ran processes AI now handles take ownership of the results those processes produce, decide where the next capability should be built, and step in on the judgment calls that need someone who knows the business. That is a more valuable use of talent already inside the company, and it is what separates pioneering that strengthens an organization’s people from pioneering that simply displaces them.
Decide Before the Choice Is Made for You
Every leadership team should know whether each part of its operating model is being kept because it is still the right one, or because it is the one everyone already knows how to run.
The practical step is to put a fresh start on the table every time a part of the business comes up for change. When a department, a function, or the whole company is due for a transformation plan, ask for the AI-native option built from scratch, priced beside it against the same goal. Making that comparison every time keeps the familiar route from winning by default.
Sabre decided its old playbook was the constraint and built past it while the choice was still its own to make. Harvey and Sierra never had a playbook to move past. Both versions of that advantage come down to the same thing: becoming AI-native on a timeline the company sets for itself, built around its own people, before competitors who got there first set it instead.
About John Fildes
I grow the top line by connecting marketing to business strategy. By leveraging powerful positioning, content marketing, and client insights, I help organizations drive qualitative and quantitative results at scale.
I've built an amazing network of incredibly talented people over the years. What I've appreciated most is those who have invested in me, mentored me, and helped me become the talented professional I am today. I pay it forward by doing the same for other high performing professionals and entrepreneurs.
Learn More: Growth Leader | Business Builder | Leadership Multiplier
All views are my own and not those of my current or prior employers.
