“The prediction [for 2026] is digital transformation in organizations is officially dead, replaced by AI native rewrites.” — Salim Ismail, Dec 2025

For the past few years, companies have leveraged AI for a new product feature, a marketing approach, a copilot, or an avenue for internal efficiency. They’ve looked at AI as something to sprinkle on top of the “real” systems.

That era is ending – at least that’s the prediction.

This is Part 1 of the Value-Led Enterprise AI series, defining the AI native rewrite and what it entails — so executive teams can weigh it alongside other AI strategy options.

The term ‘AI Native Rewrite’ is not yet a buzzword, but it will be. Among those talking about it, most are getting it wrong. The danger in the term ‘rewrite’ is that CEOs jump to the dread of the expense or see this as blowing everything up. I get it.

Instead, an AI Native Rewrite (empowered by the "AI native, Value-led Operating Model" (AVOM) introduced in part 2) creates enormous opportunity. This is about the latent customer needs that could never be served before.

Let’s use the carrot and stick analogy:
– Carrot: you can now deliver upon these unmet needs and grow your business, in a way you could never do before.
– Stick: Somebody else is going to do it, and the enterprise risks losing relevance, pricing power, and market share – and it's forced into a reactive game of catch-up.

For some businesses, continuing the add-on approach to AI will continue to make tremendous sense. For others, 2026 will mark the inflection point in AI for the next competitive phase of their industry. It will be won not by those building add-ons; it will be won by those who replace the tech foundation.

The Value-Led Enterprise AI series focuses on how business can build the conditions required to win in the next phase of applied AI and business operations.

When the Board of Directors ask senior executives, “What is our AI strategy?” there are several viable options on the table. One of them must be AI Native Rewrites. This article does not, per se, advocate for AI Native Rewrites, as every enterprise decision is context dependent. Rather, this article explains what an AI Native Rewrite is and what it entails, so the executive leadership team (ELT) can make the appropriate AI strategy decision for their organization.

AI Native Rewrites are a fresh start: a rebuild, or rewrite, of the tech system from the ground up. AI as the tech core and foundation where it now acts as the continuous orchestration and execution layer, not an add-on tool.

Some refer to this as AI-driven reengineering.

In plain terms, an AI Native Rewrite is not the answer to “how do we use AI inside our current systems?” It is “what would our systems look like if AI is the system?”

It reframes the tech stack and architecture question. To be clear, then . . .

– automating legacy workflows
– adding copilots
– wrapping LLMs around legacy – systems and data
– fine-tuning chatbots

Those AI features and automations have, and can continue to be, valuable paths for businesses. But they can also be a trap because a faster execution of yesterday’s processes can result in the death of your organization, as others retool and speed ahead into the next phase of your industry.

including several elements that are critical to success but rarely named:

1/ People stop doing the work; instead they direct the AI system
– We will need to learn how to articulate what we expect and how to manage.

2/ The agentic AI is now the orchestrating control plane
– Rather than an assistant, the AI sits at the center – routing and coordinating the work.

3/ Organizational knowledge is treated as one shared source, not data separated by department silos
– Documentation, tickets, and policies are machine-readable data accessible and actionable across the organization; sometimes referred to as a single semantic substrate.

4/ Eval-rigor, failure modes, confidence thresholds, and human overrides are first-class primitives
– Do not bolt these on later. AI native rewrites replace deterministic, human-driven pipelines with probabilistic, eval-driven systems. Once live, the systems no longer rely on human judgement. Therefore, teams must explicitly define good and bad output – encode it in the AI system and enforce it

5/ Explainability
– Build the system so humans can reasonably understand what the model did and why, enabling trust and informed decision-making.

6/ the Jumping off point for uncovering new services customers want that were impossible before
– How does AI make our old ways of doing things and our offerings irrelevant? Understand newly unmet needs and deliver products and services that can address them. As noted above in the ‘big picture’ section, this is not about efficiencies. This is about asking, “How can we grow our business by providing services customers want before our competitors do?” This is huge. However, AI Native Rewrite discussions focus on the engineering and technology, and this is where organizations risk becoming that cautionary tale. Therefore, enter the overarching AI native, Value-led Operating Model (AVOM) onto the stage – to turn the AI Native Rewrite investment into a business success.

Most companies do not yet know an AI Native Rewrite is a legitimate option. If they have heard of it, most companies do not how to do this safely. So let’s highlight the basics.

Building upon Salim Ismail’s aforementioned 2026 AI Native Rewrite prediction, he emphasizes the need to intentionally disrupt the organization from the edges of the organization.

I’ve seen this pattern work.

At Autodesk pre-LLM era, we did something similar. Leadership was concerned that nimble new entrants were chomping away at long-established market share in AEC & BIM (architecture, engineering, and construction & building information modeling) software. Many CxO’s are feeling the same today with AI native startups in their sectors.

To combat this at Autodesk, the executive team asked a handful of us to become a tiger team to re-think AutoCAD – a multi-product solution with approximately 7 million users across 160 countries. Apple, too, has famously invested in tiger teams for their special projects.

We were sectioned off from the existing daily churn, in a different building with a dedicated war room. We asked ourselves: if the market is shifting, as are customers’ needs and expectations – in the next generation of AutoCAD – where is value created? What might the future be? and How could we de-risk viability, value, usability, and feasibility in that future?

How do you start to conceive such a project internally? To begin, curate an innovation tiger team whose one mission is to disrupt your organization from the outside with AI as the foundation of the tech stack.

Keep it bounded. Select just 2-3 discrete use cases or groups of use cases with clean boundaries, with well-defined start and end points.

For example, I just finished building an AI-powered KYC* update engine. Other examples of well-bounded use cases include mortgage handling and NDAs. CRMs and Customer Service have had great success already. In manufacturing and transportation, self-contained candidates for foundational AI replacements handle equipment status and maintenance, reducing down time.

Create protective measures to ensure your AI Native Rewrite is done safely – refer to the numbered list above. And of course, clean data. This is not a call for enterprise-wide data hygiene programs. While many large organizations are, rightly, engaged in multi-year (read: 5-year long) data initiatives, that is not what is meant here. For tiger teams, avoid the immediate dread of “clean everything” thinking, and instead focus only on the data each use case needs to deliver value.

Give the tiger teams space to create and prove something real which includes pressure-testing – technically and with customers. Then, following change management best practices, bring the new technology into the core, conducting a full foundational system replacement of the legacy solution.

This wraps up the first – engineering – part of the Value-Led Enterprise AI series, a TL;DR to help senior executive teams answer the board’s question, “What is your AI strategy?”.

Making the appropriate AI strategy decision for the enterprise includes multiple factors including risk tolerance, capabilities, industry context, budget, and timing.

Now that you understand the engineering lens, beware however, the AI Native Rewrite alone fails without the operating model built to deliver value in the new context of ubiquitous AI.

Are you curious to see the operating model pairing that makes the AI Native Rewrite investment a business success?

If so, check out the upcoming part 2: "AI-Native Investment, Meet Your Operating Model". 


Don't miss part 2
→  Like or Drop the comment "REWRITE" in the LinkedIn post.
→  & Connect with me on LinkedIn 👍

Curious what this looks like in your organization? Contact me; let's talk.