Leading Through AI Disruption and Resource Constraints

The situation

It was 2023, and the house was burning.

PM leadership had decided the team needed to invest in a new Campaign Manager experience — AI-first. As the design leader, every partner came to me immediately. Do you have a solution ready? Are you going to prioritize our work over the other 30 PMs? The requests were pouring in, the pressure was acute, and the resources weren't there.

Here's what the numbers looked like: our designer-to-PM ratio was 4:1 — four product managers for every designer. Designer-to-engineer ratio was 26:1. In a single quarter, we received 94 project requests — 200% more than our team had capacity to deliver. We were two years out of the pandemic. Engineering and PM were still hiring. Design had a freeze.

Meanwhile, team morale had dropped to 66 out of 100. Managers were doing their best, but the reality was that the fastest way for people to demonstrate value was to deliver on what was being asked — and they were burning out doing it. Prioritization was reactive. People were stretched across too many things with no clear strategy owner.

The approach

The first step was to stop pretending the idealism was going to save us. Presenting designer-to-PM ratios to leadership and asking for parity wasn't going to work — it was grounded in an ideal org structure that didn't match reality. I needed a different approach.

I built what I now call the Observation → Action Cycle — a decision-making loop for navigating organizational chaos:

Step 1 — Name the undeniable truths.Three things were real whether we liked them or not: AI was a critical strategic opportunity and we couldn't ignore it. The team wasn't prepared for it. And any reorganization carried risk — with morale this low, we could lose even more people.

Step 2 — Make honest observations.Incentives were misaligned and political. PMs were optimizing for revenue, engineers for velocity, and design was stuck in the middle executing without strategic ownership. We were saying no to requests, but without clearly explaining why — and that was eroding our influence. There was no clear strategy owner on the design side.

Step 3 — Take deliberate action.I was transparent with my managers and ICs: "We see what you're feeling. We know the pain. We're here with you." That wasn't just a pep talk — it was a commitment to change the operating model. We advocated for healthy environments to do the best work. We set expectations upward. We started building the case with data, not emotion.

Step 4 — Arrive at a shared realization.Teams understood the constraints. Leaders aligned on reality. The key principle: you have to show the thing to change the thing. Leading with an informed point of view takes the emotion out of tough conversations because you're speaking from evidence.

From there, we restructured around themes. Under the four initiative areas I owned, teams were reorganized into working groups tied to strategic themes. Each group had specific KPIs, research tied to their work, prioritization criteria, and a narrative that others could join. This elevated our operational and research teams to have a stronger voice in the product development lifecycle.

Selecting the pilot team was deliberate. Four criteria: start with pioneers (people willing to try new things), prioritize growth mindset (adaptable talent comfortable with ambiguity), choose partners with highly collaborative cross-functional relationships, and look for systems thinkers — critical when building AI experiences that span multiple modalities across LinkedIn's business lines, including Flagship, Sales Navigator, and Talent Solutions.

We didn't boil the ocean. We started with a focused pilot, proved the model, and expanded.

The AI strategy was equally intentional. I distinguished between three things that were getting conflated in every conversation: AI readiness (how the organization is set up to adopt AI), AI strategy (thinking about AI experiences and building trust with users), and building with AI (the co-pilot tool side). There was no shortage of content about the third one — everyone was talking about vibe coding and new tools. But the first two were where design leadership could actually make a difference.

When we advised on LinkedIn's first AI-assisted campaign marketing tool, the design team's involvement shaped the product and engineering strategy. We championed transparency — as much as users want automation, they don't want it operating in a black box. Intervention and control needed to be available. UX copy needed to be clear and reassuring. AI capabilities should make users better at what they do — if they cause confusion, it's time to re-evaluate.

Then came the 2024 reorg. LinkedIn shifted from centralized to decentralized. We no longer had a VP of design. We no longer had a VP of product. Rather than panic, we reframed it as opportunity: we were now closer to the decision-making process. Design leadership could co-lead strategy alongside product partners — not just execute on product's direction. Design had a unique strength in storytelling — compelling narratives with artifacts that product couldn't replicate. And we had greater authority in defining quality, which was critical for AI experiences where trust is paramount.

Every level contributed to the new model. ICs — from senior designers through staff — built narratives that informed work being scoped earlier in the process. Managers at the lead through senior manager level brought clearer themes and drove prioritization conversations. Principal and senior principal designers led at the business level, not just the craft level. At my level, the focus was on building stronger asks with C-level and VP-level executives, keeping teams transparent, and creating the conditions for the team to show up in leadership discussions.

None of this was hypothetical. The shift from crisis to clarity showed up in the numbers.

The Impact

The turnaround was measurable:

• 31% morale increase from the 66/100 baseline
• 7 new hires secured through prioritization conversations and a clear resource narrative
• 2x velocity — concept-to-build timelines went from 6–9 months to within quarters
• 47% reduction in requests (from 94 down to ~50) through better prioritization and the ability to say no with a clear explanation
• 4 AI-driven experiences shipped — including AI-powered campaign creation, measurement & optimization, marketing overview insights, and an AI-driven campaign performance digest
• 34% adoption increase with SMB customers
• Design co-led strategy for the first time — no longer just executing on product's direction

Reflection

This period taught me that design leadership is at an inflection point. Some of the concepts the industry has relied on for the past decade — design maturity models, ideal org structures, the aspiration of "a seat at the table" — aren't working in today's organizations. The reality is messier. You're working with fewer resources, broader scope, and AI pressures that most teams aren't prepared for.

What worked wasn't a framework from a textbook. It was coming to terms with reality, creating clarity in the middle of chaos, and connecting every person's role to a purpose that led to outcomes. Shape systems, not just screens. Shape culture, not just projects. Shape tomorrow starting today.

One thing that still concerns me: too many design leaders are talking about vibe coding and not enough about how they're influencing strategy. That needs to change. We need more leaders willing to drive culture and shape products for the future — because if we don't, others will define it for us.

Where I'd push myself further: I should have started the theme-based restructuring sooner. The first year of the crisis, I spent too much time trying to make the existing model work under pressure — optimizing within a broken system instead of redesigning it. By the time I built the Observation → Action Cycle and restructured around themes, we'd already lost time and some trust. The framework works, but I learned that the instinct to observe before acting can also become a reason to delay when the situation demands urgency.