Cross-functional collaboration in the age of AI
Cross-functional collaboration couldn’t be more important today. Building, buying, and using AI raises critical questions that cut across disciplines: Is licensing this AI product worth it? Is our new AI product safe to use? How can we make this new AI tool useful for customers?
The only way organizations can answer these questions is to draw upon the combined expertise housed across its various functions — including engineering, data, design, product, legal, and more.
It’s tempting to imagine cross-functional collaboration as just another corporate buzzword, or a fancy term for a skill we all learned in grade school.
But as it turns out, organizations get cross-functional collaboration wrong more often than not.
Just a few common activities people mistake for cross-functional collaboration include:
Sharing round robin updates in cross-functional meetings and emails.
Including a representative from every function in every discussion or decision.
Ambushing cross-functional partners with requests for input without providing relevant context or sufficient time to respond.
Recognize any of these activities?
They can be more accurately described as cross-functional coordination and have the undesirable effect of increasing coordination overhead (a variation on the concept of Josh Kaufman’s communication overhead). I’ve certainly mistaken these and other activities for cross-functional collaboration in the past — and I can attest first hand that they actually increase inefficiency without driving meaningful impact.
So, what should organizations do instead?
Here are five tips for implementing cross-functional collaboration at scale.
Ring-fence initiatives with clear objectives to experiment with cross-functional collaboration approaches that make sense for your organization. The ring-fence can prevent ineffective past ways of working or unhelpful reporting lines from bogging down these efforts.
Empower teams to execute these ring-fenced initiatives, and assign clear accountability for execution. Empowerment must be meaningful, so factors like appropriate decisionmaking authority, resources, and organizational support are key.
Identify domain translators, i.e., people in your organization who are ‘fluent’ in more than one discipline. Plug them into cross-functional work to translate across teams and address common points of confusion. For example, they can help develop internal taxonomies to drive common understanding of terms with different meanings among different disciplines. And make sure to formally recognize this work — it is a real skill whose value should not be taken for granted. (Thanks to Jordan Famularo for the suggestion to make this explicit!)
Establish feedback loops about cross-functional work to identify useful collaboration practices. For example: Which team(s) needs to be involved and when? What context does each team need to contribute meaningfully? How can we remove these common bottlenecks? This feedback should be recorded and can inform the next cross-functional effort, and so on.
Design formal incentives for cross-functional collaboration, e.g., cross-functional KPIs, performance review goals. This may seem unnecessary (we all want to work together, right?), but it is critical for aligning priorities and resources like budget.
What do you think?
Would you add anything to this list? Would you change anything? I’d love to hear your thoughts!