I’m watching Mad Men now. Yeah, I’m nearly 20 years late to the party but I’m glad I finally got here. If you’re not familiar, it’s the story of a 1960’s ad agency and its charismatic creative director, Don Draper. Like any creative director, Don is strongly opinionated and, like most creative agencies, all work has to pass through him and get sign-off before it can move on to the next step.

Now, in the 1960’s the work entailed paper, pen, artboards and other physical media that was created by hand. It took a lot of effort to create and edit. Approval of the work was a risk mitigation tactic. It was a deliberate speed bump designed to slow the team down because the cost of action and reaction was significant. We took that same approach into the world of digital product development. The risk here was similar. The cost of building the wrong thing was significant. This is no longer the case with AI which means that seeking “permission” is now the slowest bottleneck to forward progress. 

Approval is a false constraint in an AI world

AI made execution cheap. The cost of producing anything from a prototype to production quality code is extremely low. This reduces the risk of sending engineering teams off on a wild goose chase only to change direction on them as new information becomes available. That constant course correction was costly as was producing the artifacts necessary to get those engineering teams started. Sign-off, approval and permission were needed so that someone was accountable for sending a software development team down a specific path. This is no longer the case. 

Permission seeking slows product teams to uncompetitive speeds

The drawback of constant permission seeking from software development teams is speed. Ask any enterprise team what their biggest goal is and a top 3 answer every time is, “time to market.” Yet, what is it that plays a big part in slowing teams down? Seeking approvals and sign-offs. The bigger the organization, the more levels of sign-off need to happen. Before you know it, teams are spending more time on approval cycles than on the actual work itself. The market moves on in the meantime with customers left without the benefit of the improved functionality. This is expensive and risky. Whereas before seeking sign-off reduced risk, now it’s increasing it. 

Outcome-aligned guardrails replace permission loops

AI has made learning an order of magnitude faster. Product managers, ux designers, researchers and engineers can prototype an idea in minutes. In a marginally longer period of time they can get feedback on that prototype from customers. That evidence allows for better and faster iterations on the idea. Much faster than ever, teams can now build a compelling evidence-based mandate for a specific feature and design solution. If the business problem has been approved and clear outcome-based key results are in place, the team can move forward with confidence that their work is going to positively impact the customer and the company. 

The same is true for the software engineering teams are now creating. The cost of building an early test version of a validated solution is trivial in comparison to what it was just a year or two ago. The risk isn’t that the team won’t get another shot at this feature. It’s that the team won’t move the needle on their desired key results. These outcome-based guardrails – key results – serve as the guiding hand for product teams. If their prototypes show promise on moving the right customer behaviors forward, software engineers can start to produce solutions and put them into production. As feedback loops provide data on whether the work provided value to customers, the team can prepare the next iteration in a much faster and less expensive way. 

Let evidence—not hierarchy—do the aligning

AI is reshaping everything about how we work. Leadership approvals used to mitigate risk. Now they cause it because they happen too slowly. Aligning a team to customer-focused key results and giving them the AI-powered tools to help them learn, deliver and iterate as fast as possible takes the risk of building the wrong thing nearly out of the equation. Teams can quickly learn if their work is valuable through market-based evidence, not leadership approvals. That’s where alignment can and should come from. It’s also how we start to truly take advantage of the power of AI-based work. 

If you’re interested in a deeper understanding of human-focused Objectives and Key Results, I highly recommend our most recent book on the topic, Who does what by how much?(we also have a short video course on the same topic here).