The Innovation Program That Actually Works

(And Why Your Current One Probably Doesn't)

Bonus Article 2 — 2weekAI Blog

Walk into the right call center and you'll see something that stops you cold.

Employee desks covered in small cards. Stickers. Printed certificates. Some desks have ten. Some have thirty. Some have over a hundred.

Each one represents an implemented idea.

Not a suggestion dropped in a box that nobody opens. Not a form submitted to a portal that routes to a committee that meets quarterly. Not a hackathon project that gets applauded on Friday and forgotten by Monday.

An idea that went from that person's head to a working improvement in the organization. Reviewed. Approved. Built. Deployed. Running in production.

I saw this firsthand at MBNA America. One of the most profitable banks in American history. And the desks that had a hundred implemented ideas weren't in the executive suite.

They were in the call center. Entry level positions. People answering phones.

That's not an accident. That's a system.

What MBNA Actually Built

The Masterpiece Program wasn't a suggestion box with better branding.

It was a structured innovation pipeline that treated every employee — regardless of title — as a source of competitive intelligence.

Here's how it worked.

Every idea submitted got a real review by the department that owned the problem. Not a generic acknowledgment. A real evaluation by the people closest to the work.

If the idea was accepted and implemented — $300. Per idea. To the person who submitted it.

If it was a team effort on a bigger idea — $5,000 split across the team.

If your idea was named Idea of the Year — a number significant enough that people remembered it. Significant enough that everyone in the building knew about it and wanted to be next.

Now do the math on a $300 bounty.

An entry level call center employee submitting one idea a week over a year could earn $15,000 in bonuses on top of their salary. On ideas that the company gladly paid because every accepted idea saved or generated more than it cost to reward.

The result was a call center floor where people were actively looking for problems to solve. Not because HR sent a memo about innovation culture. Because there was a real financial incentive attached to finding one and a real process for turning it into something.

100 implemented ideas per desk isn't a culture achievement. It's a system achievement.

Why Most Innovation Programs Fail

Before we build the upgraded version let's be honest about what passes for innovation programs in most organizations today.

The annual hackathon. Two days. Teams build things. Presentations happen. Winners get announced. Applause. Then everyone goes back to their regular jobs and the winning idea sits in a slide deck until someone finds it two years later during a reorganization.

The suggestion portal. Ideas submitted. Acknowledged automatically. Routed to a committee. Reviewed when the committee has capacity. Which is approximately never. The submitter hears nothing for four months and stops submitting.

The innovation task force. Senior leaders assigned to evaluate AI opportunities. They hire a consulting firm. The consulting firm produces a framework. The framework produces a roadmap. The roadmap produces a Phase 1. Phase 1 is still being scoped eighteen months later.

The mandatory AI training. Eight hours of modules. A certification badge in the learning management system. Zero application to actual work. The organizational equivalent of four years of high school Spanish that evaporates the moment the class ends.

Every one of these programs has something in common.

They treat innovation as an event instead of a system. And they treat employees as participants in the program instead of owners of the outcome.

The Upgraded Model: Crowdsourced AI Innovation

Here's what the MBNA model looks like rebuilt for 2025 with AI as the execution layer.

Step 1: Real Problems. Not Hypotheticals.

The program starts with a problem inventory. Not "submit your ideas for how AI could help the company." That's too abstract and produces ideas that sound interesting and go nowhere.

Real operational problems. Sourced from every department. Prioritized by frequency, cost, and frustration level. Published to the organization as an open list.

"Here are 47 real problems this company has right now. Pick one. Solve it."

That's a different starting point. That's a problem with an owner, a context, and a measurable outcome when it gets fixed.

Step 2: Access That Matches the Ambition.

This is where most programs fail before they start.

You cannot run a meaningful AI innovation program and give people Co-pilot with the guardrails locked to the point of uselessness.

Give them real tools. Claude. ChatGPT. Gemini. The tools that actually do the work. With appropriate guardrails for data sensitivity — yes. With the capability to genuinely explore and build — absolutely required.

If you're asking people to innovate and you're giving them the innovation equivalent of safety scissors — you're not running an innovation program. You're running a liability management exercise dressed up as one.

The tools are affordable. The enterprise agreements exist. The security frameworks are established. We covered all of this in Article 9.

Approve the tools. Then get out of the way.

Step 3: Individuals and Teams. Both.

Some problems get solved by one person with a clear idea and the right tool. Some require a cross-functional team that brings together perspectives that don't normally sit in the same room.

The program supports both.

Solo submissions for individual innovations. Team formations for bigger problems that need multiple domain experts. The system accommodates both and rewards both differently.

The cross-functional team formation is actually where some of the most valuable innovation happens. The developer who finally gets to sit with the operations manager. The legal coordinator who gets to work with product. The institutional knowledge that's been siloed by org chart finally gets to combine.

Step 4: Idea to MVP. Not Idea to Slide.

This is the part that changes everything.

The deliverable isn't a presentation. The deliverable is a working prototype.

With AI as the execution layer a motivated employee can go from problem identification to working MVP in days — not months. The business analyst who understands the process can now build the tool that fixes it. The operations manager who sees the inefficiency can now prototype the solution.

The governance team — the group that reviews submissions for implementation — evaluates working prototypes. Not decks. Not proposals. Things that actually run.

That changes the quality of the conversation completely. It's a lot harder to dismiss something you can click on.

Step 5: Recognition That Actually Means Something.

The sticker on the desk matters. Don't skip it because it sounds trivial.

Culture is built in the small visible moments. The physical artifact on the desk that says "I built something real and this organization acknowledged it" is worth more than any all-hands announcement. It's permanent. It's visible to every colleague who walks by. It starts conversations.

The MVP badge. The implemented idea marker. The visible signal that this person ships things. Make it real. Make it physical. Make it something people actually want on their desk.

Step 6: Compensation That Incentivizes Repetition.

This is where most programs get it wrong.

Flat fee bonuses communicate that the organization values the gesture of innovation more than the outcome of it. A $20 gift card tells someone their idea that saved the company $200,000 is worth a fast food combo meal. Maybe. If they skip the drink.

That's not incentivizing. That's insulting.

Tier it to the value created. Documented. Verified. Audited by finance.

$10,000 in annual savings → $500 bonus. That's 5%. Cost justified. Meaningful enough to notice. Not meaningful enough to create gaming behavior.

$100,000 in savings → $5,000. Life-changing for most employees. Rounding error for the organization.

$1,000,000 in savings or revenue → $50,000. Career-defining for the employee. Still the best ROI the company has ever seen on a single innovation investment.

$3,000,000 idea → $150,000. You just made someone's life materially different. And every person in the building knows what's possible now.

The person who gets $150,000 for a $3 million idea doesn't just do it again. They become the most powerful innovation advocate in the organization. Their colleagues start looking at problems differently. The culture shifts — not because of a program launch announcement but because of one real outcome that everyone witnessed.

That's how the MBNA call center got to a hundred ideas per desk. Not motivation. Math.

The AI Multiplier

Here's what MBNA couldn't have in 1995 that you have today.

Every employee with a real problem, real tools, and real incentive now has the execution capacity that used to require a development team.

The call center rep who identifies a broken workflow can now prototype the fix. The operations coordinator who sees the recurring defect can now build the detection system. The analyst who understands the reporting gap can now build the dashboard.

The idea-to-MVP cycle that used to take six months of backlog prioritization and sprint planning now takes days.

A $300 bounty that produced 100 ideas per desk in 1995 — with no AI tools, with paper forms, with ideas that had to be handed off to a development team to build — becomes something categorically different when the person with the idea can also build the solution.

The Masterpiece Program was powerful with 1995 tools.

With AI as the execution layer it's a compounding innovation engine that costs a fraction of what you're currently spending on innovation programs that produce slide decks.

What Leadership Has to Do

The program doesn't run itself. And it doesn't survive without leadership that actually means it.

Three things that make the difference between a real innovation program and another initiative that fades by Q3.

Mean the incentives. If the compensation structure says $50,000 for a million-dollar idea — pay it. Publicly. With recognition that makes everyone understand what just happened. The first payout is the program's most important marketing moment. Don't flinch at the number. You just got a million-dollar idea for $50,000. That's the best deal you'll ever make.

Protect the submitters. Nothing kills an innovation program faster than an employee who submits an idea that challenges an existing process and gets quietly managed out for it. The program requires psychological safety that is actively defended by leadership — not just mentioned in the launch email.

Close the loop. Every time. Every submission gets a real response. Accepted or not. With a reason. Within a defined timeframe. The black hole of the suggestion portal is what killed your last program. Don't recreate it with better branding.

The Question Behind the Program

Here's what the MBNA Masterpiece Program was really about.

It wasn't about cost savings. It wasn't about innovation metrics. It wasn't about the program at all.

It was about a fundamental organizational belief that the people closest to the work have the best information about how to improve it — and that belief is worth putting money behind.

Most organizations say they believe this. MBNA proved it with $300 checks and a hundred stickers on a call center desk.

The technology has changed. The belief required hasn't.

Your people are looking at broken processes every single day. They have ideas. Most of them have stopped sharing because nothing happened the last time.

Give them real tools. Give them a real process. Give them real money when they deliver real value.

Then get out of the way and watch what a call center full of motivated problem solvers can do when they have the execution capacity of a development team sitting in a browser tab.

The stickers on the desk are waiting.

*2weekAI deploys AI that actually works — in 2 weeks, at a fixed price. Dave has delivered AI solutions across Fortune 100 enterprises and growing businesses alike. No transformation program required. [Book a discovery call.]*