The Playbook: Driving AI Innovation from Within

This is the companion to The Case for Change, which covers the why — the urgency of the exponential economy, the psychology of resistance, and the principles behind successful adoption. This article is the how.

It’s a field manual for intrapreneurs: people who see opportunity and want to make change happen from within. No theory. Just tools, recipes, and practical steps drawn from SRI International’s Five Disciplines of Innovation.

Step 1: Find Your First Painkiller

In risk-averse environments, “nice to have” initiatives don’t survive. Focus on painkillers, not vitamins (The Case for Change explains why this distinction matters so much).

Before proposing any AI solution, answer these questions:

What measurable problem exists?
Who feels it?
What does it cost in time, money, risk, or reputation?

Recipe

Identify a workflow consuming disproportionate time or causing errors.
Quantify the baseline — hours, error rates, costs.
Confirm someone owns the problem and cares about fixing it.
Define what improvement would actually matter.

Done when you can state:

“We reduce X from A to B within Y weeks.”

A practical example: imagine opening your inbox to 30 emails that need to be prioritised, filtered, responded to, and converted into action items. Done manually, that’s easily an hour. With an AI-powered system, it takes under a minute. Is it perfect? Probably not. But the time saved is real, measurable, and immediately felt.

Marginal improvements rarely justify change. Aim for 2 to 10 times better than the status quo. That’s the threshold where institutional inertia starts to lose.

Step 2: Create a Shared Language — NABC

One of the biggest obstacles in any large organisation is the Tower of Babel problem. Engineering talks feasibility. Finance talks ROI. Operations talks disruption. Legal talks compliance. They’re all discussing the same initiative but might as well be speaking different languages.

The NABC Value Proposition cuts through this:

Need — What important, measurable unmet need exists? Not “we need to be more innovative” but “our document review takes 200 person-hours per month with a 12% error rate.”

Approach — What’s the proposed solution? Describe it simply enough that a busy person gets it quickly. This is one of the toughest challenges — making something complex feel clear and accessible.

Benefits per Costs — What are the measurable benefits and what will it cost? Be explicit about the Value Factor — the ratio of benefits to costs.

Competition — How does this compare with alternatives? This includes doing nothing. The status quo is always a competitor.

Recipe: Write a 6-Line NABC

Need: “Our manual review takes 200 hours/month with 12% rework.” Approach: “AI-assisted triage with human validation.” Benefits per Costs: “Reduce time by 70%, errors by 50%, minimal infrastructure cost.” Competition: “Status quo; outsourcing; commercial SaaS.”

If your proposal isn’t clearly better than the alternatives — including doing nothing — it won’t move forward.

The power of NABC is that it forces rigour without requiring a 300-page report. In Silicon Valley, where billions are invested annually, typical business plans are ten to fifty slides — because quality, not quantity, is what matters.

Step 3: Diagnose Resistance — The FUD Translation Table

Resistance is normal. As The Case for Change explores in depth, FUD (Fear, Uncertainty, and Doubt) isn’t the enemy — it’s diagnostic information.

The practical skill is reframing: turning each complaint into a design requirement.

What you hear	What it means	What to do
”I’m too busy"	"Show me this is worth the effort”	Demonstrate time savings in the first week
”This won’t work here"	"I can’t visualise it”	Run a proof-of-concept with real data and workflows
”AI makes mistakes"	"I need to trust it”	Design a human-review step and clear error-handling
”This will replace our jobs"	"Will I still be valued?”	Define how roles evolve, not disappear
”We tried this before"	"What’s different now?”	Identify what changed since the last attempt
”Management will never approve"	"We lack sponsorship”	Build executive support with a compelling NABC

Build this table with your team. Keep it visible. Every complaint, reframed this way, becomes a specific problem to solve — and solving problems is something your organisation already knows how to do.

Step 4: Practise Nemawashi

There’s a Japanese practice called Nemawashi that is essential for driving change in any consensus-oriented environment. The word literally means “going around the roots” — from the art of growing bonsai, where more time is spent developing the root system than pruning the branches above.

In practice, nemawashi means talking to stakeholders individually, one by one, before you ever make a formal proposal. Share your idea informally. Gather feedback. Understand concerns. Adjust your approach — all before the idea enters any official process.

This is the opposite of how most change initiatives fail: someone presents a polished proposal in a large meeting, catches people off guard, and triggers a defensive reaction. By the time the meeting is over, the proposal is dead and the political damage is done.

Recipe

Identify the 5–10 people whose support or opposition will most influence the outcome.
Meet each one individually.
Share your NABC. Ask: “What am I missing?”
Listen carefully. Incorporate their feedback.
Then — and only then — formalise the proposal.

By the time it’s presented formally, most people in the room should already know about it, understand it, and feel their input helped shape it.

Step 5: Help People Vault the Void

The Case for Change explains why AI adoption is fundamentally an identity challenge, not a skills challenge. Here’s how to address it practically.

Recipe

Identify the person’s core strength. What are they known for? What do they take pride in?
Show how that strength becomes more valuable with AI. The lawyer’s legal judgment evaluates AI output. The developer’s expertise spots subtle code errors. The analyst’s domain knowledge turns raw analysis into insight.
Create a low-risk trial. Don’t ask for commitment. Ask for an experiment. “Try it on one task this week and see what happens.”
Celebrate early wins. Publicly. Specifically. Name what they did and what it achieved.

When people see continuity between who they are and who they can become, adoption accelerates.

Step 6: Assign a Champion

Every successful innovation has a champion. No champion, no project, no exception.

A champion is not a project manager. A champion:

Takes personal responsibility for the initiative’s success
Anticipates obstacles before they materialise
Builds alliances across departments
Refines the value proposition continuously
Models the behaviours they want to see — visibly using AI themselves

If you’re reading this playbook and thinking “this is my initiative,” then you’re the champion. Own it fully.

Step 7: Remove Waste Before Adding Tools

There’s a Japanese concept — Muda — meaning non-value-adding activities. Endless meetings with no action items. Reports nobody reads. Manual processes that exist because “we’ve always done it that way.”

Before introducing AI, take an honest inventory:

Which activities actually create value?
Which are waste?

Removing obvious waste before asking for new investment builds credibility. And when you can show that AI itself helps eliminate remaining waste — automating manual reviews, accelerating approvals, reducing rework — the proposal becomes much harder to refuse.

The wonderful thing about muda is that once your team internalises the concept, they start identifying waste everywhere. When someone can point at a process and say “that’s muda” and everyone nods, you know you’re making progress.

Step 8: Call Out Red Herrings

Every organisation has beliefs that function as barriers to change but are, on examination, untrue. These are Red Herrings — assumptions that take the project off track.

Common ones around AI:

“Our data isn’t clean enough.” Often an excuse rather than a genuine barrier. Start with a subset.
“We don’t have the budget.” When the ROI would easily justify the investment.
“Legal won’t allow it.” When no one has actually asked legal, or the regulations are less restrictive than assumed.
“Our industry is different.” When competitors in the same industry are already adopting AI.

Red herrings need to be tested against facts. Some barriers are genuinely real and need real solutions. The key is to distinguish between the two, name them clearly, and make them visible so they can be addressed rather than left to fester as vague objections.

Step 9: Build Brick by Brick

Change rarely happens through grand pronouncements. It happens brick by brick.

Start with a small, high-value pilot. Show results. Let early adopters demonstrate success. Use that credibility to expand.

Expect roughly this distribution:

~20% will engage early
~70-75% will watch for proof
~5-10% may never come around

Focus energy on the 20%. Help them create demonstrable success. Then leverage those wins to engage the wider group. Over time, the flywheel begins to spin. The cautious middle shifts from “wait and see” to “how do I get involved?”

Momentum beats persuasion.

Use Quantitative Rigour

In a risk-averse culture, the intrapreneur is most effective when positioned not as a risk-taker but as a risk-reducer.

Good venture capitalists are often perceived as risk-takers, but the reality is the opposite — they work hard to reduce risk through research, analysis, and iterative validation before committing resources.

Use Value Factor Analysis for each AI proposal: assess the importance of each dimension to the stakeholder and how well the approach satisfies it. Compare total benefits divided by total costs against the current approach.

This quantitative rigour serves two purposes. First, it gives decision-makers evidence to justify the change to their stakeholders. Second, it demonstrates that this isn’t a technology enthusiast’s pet project but a carefully analysed opportunity.

Your Starting Sequence

If you’re driving AI innovation from within, start here:

Find one measurable painkiller. Quantify it.
Draft a sharp NABC. One page. Maybe less.
Begin nemawashi. Quiet, one-on-one stakeholder conversations.
Run a small pilot. Real data, real workflow, real results.
Show results. Let the numbers speak.
Expand deliberately. Brick by brick.

Innovation inside established organisations is not theatre. It’s disciplined value creation, applied consistently.

In the exponential economy, waiting for perfect certainty before acting may be the riskiest choice of all.

Lay the first brick today.

This playbook draws on principles from SRI International’s Five Disciplines of Innovation (Curtis R. Carlson & William W. Wilmot). Interpretations and applications are my own.