What would it take to make innovation inside a company as intelligent as the technology outside it?

That’s the question we’ve found ourselves returning to in conversations with leaders over the past year. Everyone is experimenting with AI tools, but few have figured out how to integrate AI into the way innovation happens in a consistent, scaled way.

In this piece, we offer a practical framework for thinking about what it means to embed AI into the infrastructure of innovation and how its role might evolve over time.

‍

From Tools to Systemic Advantage

We’ve seen firsthand how incremental uses of generative AI tools, like summarizing research or rewriting slides, can help at the margins. But the real breakthroughs come when AI is embedded into the broader systems that power innovation.

An AI-first system can deliver compounding returns, not just one-off boosts. For example, it  creates structural advantages in:
‍

• Speed. At Unilever , teams use digital twins and simulations to test product concepts virtually - cutting time-to-concept in by half and dramatically accelerating early stage innovation.
  ‍
• Precision. At Moderna, an AI-enabled R&D platform links over 200 scientists and synthesizes data across experiments, surfacing the most promising mRNA candidates weeks earlier than traditional methods.
  ‍
• De-risked bets. At Ocado Group, machine learning continuously analyzes overlapping innovation pilots and reallocates resources in real time - helping teams shut down duplicative work and double down on higher return experiments.
  ‍

These advantages aren’t just hypothetical. A growing number of companies are beginning to prove them out. Netflix’s algorithms now guide everything from content bets to creative briefs, illustrating how AI can become embedded across the entire innovation chain. Siemens, meanwhile, is using AI across its industrial platforms, from using machine learning to optimize turbine design to predictive maintenance across smart factories.

Yet these examples are still the exception. A recent McKinsey study found that while most leaders expect AI to reshape their businesses, only 1% of companies say they’ve reached AI maturity. The biggest obstacle: leadership inertia.

‍

The Innovation Intelligence Curve

As we recently introduced, we’ve developed the Innovation Intelligence Curve: a four-stage framework and scorecard for organizations to assess where they are today and chart a path forward.

The model outlines four levels that organizations move through as they evolve from experimenting with AI to embedding it into the core of their innovation system. We describe these levels in terms of the primary way AI is used:
‍

• Level 1 - Passive Assistant: AI is a reactive helper, used informally by individuals on innovation-related tasks.
  ‍
• Level 2 - Structured Partner: AI is embedded in guided workflows, helping standardize how ideas are developed and assessed.
  ‍
• Level 3 - Proactive Coordinator: AI surfaces patterns, flags issues, and recommends action across teams.
  ‍
• Level 4 - Autonomous Engine: AI takes the lead within defined guardrails, initiating action and optimizing in real-time.
  ‍
  
  ‍

To bring the model to life, we’ve highlighted a representative use case and the corresponding human role at each level of the curve. This helps clarify how organizations actually behave at each stage, and how the role of people shifts as AI becomes more engrained. Even at the most advanced level, human oversight remains essential, though its nature evolves.
‍

A Scorecard for Making Progress

Most organizations don’t live at just one level of maturity; they straddle them. One team might be experimenting with synthetic research and autonomous test cycles, while another is copy-pasting ChatGPT prompts into Word docs.

To help leaders evaluate where they stand and what next steps might make sense, we’ve developed the Innovation Intelligence Scorecard. Its eight dimensions encompass both foundational infrastructure for innovation and the integration of AI tools and workflows.

This scorecard is designed to help your organization assess its current state across eight critical dimensions of innovation maturity. It reflects both foundational practices (strategy alignment, governance, culture) and how AI is being integrated to enhance them. We use it as a diagnostic with leadership teams, a tool to guide innovation planning, or as a benchmark to track progress over time.


‍

What Progression Looks Like

So how does a company advance along the journey towards AI-powered innovation? It begins with clarity on desired outcomes: usually some combination of driving new revenues, improving efficiency, or reducing decision times. From there, the focus shifts to training AI systems on relevant internal data, equipping teams with intuitive tools, and setting clear guardrails for decision-making.

As an example, imagine a CPG company beginning in Level 1: a few product managers use ChatGPT to speed up idea generation, tidy up specs, and develop financial models. It’s helpful, but informal and fragmented.

As interest builds, leaders see the signal and invest in a structured tool that embeds AI into the early-stage idea funnel. They also set light-touch governance around how the tool is used and evaluated. That shift, from individual tinkering to shared workflows, moves them into Level 2.

From there, teams begin systematically tracking which concepts advance and why. They align the tool with internal data sources like sales performance and consumer feedback. They create feedback loops to learn from what works. Patterns emerge. Now AI is not just helping teams move faster, it’s helping them make better decisions. That’s Level 3.

Eventually, the company defines guardrails that allow AI to take the lead in certain low-risk decisions, such as testing packaging claims or triggering concept validation runs. As the team builds confidence, elements of Level 4 – the autonomous innovation engine - emerge.

You don’t need a moonshot to get started; but you do need a system. The real power of AI comes when it’s built into the workflows, data infrastructure, and decision-making guardrails that shape how innovation actually happens.

The future will belong to companies that move beyond scattered experiments and start building real organizational capability.

‍