Ep 328 – Christy Lake (Chief Administrative Officer, Twilio)

1. Stop trying to predict the entire future of AI

One of the hardest parts of leading through AI right now is that there isn’t a reliable playbook.

The capabilities are evolving too quickly. What feels cutting-edge today can be outdated tomorrow. And that makes it easy for leadership teams to get distracted by every provocative headline, prediction, or new use case.

Christy’s team found a practical way to manage that uncertainty: organize the work into three horizons.

Horizon 1: Things you’re confident you need to do now.

Horizon 2: Bets you’re placing based on where you believe things are headed over roughly the next 12 months.

Horizon 3: Interesting, provocative ideas you’re watching — but aren’t ready to act on yet.

“We kind of came up with this three horizon view… It gave us some comfort because mostly what we were hung up in is the stuff that was in Horizon Three that’s so futuristic.”

The lesson: you don’t need an answer for every possible AI future. Separate what requires action today from what simply deserves attention.

2. If you want to redesign work, start with tasks — not jobs

Most companies still think about workforce planning in terms of roles and headcount.

But AI increasingly forces leaders to go one level deeper.

A job is really a collection of tasks, and those tasks require different skills. If AI takes over some portion of those tasks, the question isn’t simply whether the entire job disappears.

The better questions are:

- What work is left?

- What new work is being created?

- Which tasks should move between teams?

- Where is human judgment still required?

That’s why Twilio invested early in a skills-based infrastructure. If work becomes increasingly dynamic, companies need visibility into the skills underneath jobs so they can reassemble work as technology changes it.

3. Don’t automate so quickly that you eliminate the human capabilities you still need

It’s easy to look at AI through one lens: What can we stop doing?

Christy argues leaders need to spend just as much time asking the opposite question:

What do we explicitly want humans to keep doing?

Some work can become faster, easier, or entirely automated. But companies still need to identify the areas where judgment, context, creativity, relationships, and human decision-making create value.

“It’s sexy to look at, you know, what do we not have to do anymore? But it’s just as important to think about what do you want your people to do.”

That becomes an org design principle.

Before automating a workflow, decide whether there are capabilities inside that workflow you still need employees to develop. Otherwise, efficiency gains today could accidentally weaken organizational capability tomorrow.

4. The future employee experience probably looks like one prompt box

Employees don’t want another tool. They don’t want better search across 15 disconnected systems either.

Christy believes the employee experience will increasingly move toward a simpler interface: one place where employees can ask what they need, access connected company information, and eventually have agents take actions on their behalf.

“People want one prompt box that has all the connectors to the data sitting behind it, that you can just go action.”

For People teams, that shifts the UX question.

Instead of asking, “How do we improve this HR application?” start asking, “Why does the employee need to enter this application at all?”

5. AI adoption is creating a permanent change loop

The employee challenge with AI isn’t only learning a new technology.

It’s that people rarely get to reach mastery before the technology changes again.

Employees learn one workflow. A better model arrives. They adapt. Another capability appears. Then the process starts over. And much of that learning is happening on top of people’s existing jobs.

“When you’re always being reset to learning a new thing, you never get to enjoy that period of… mastery.”

Christy shared the example of a leader responsible for getting a team of 120 people up to speed on new ways of working — while simultaneously trying to figure those new ways of working out herself.

That creates an important responsibility for People leaders: don’t treat AI upskilling as an individual side project. If employees are expected to fundamentally change how they work, companies need to create real time, structure, tools, and support for that learning.

6. Remote work succeeds when you design the company around it

Twilio didn’t stay remote-first because of ideology.

They looked at the data.

At the time they made the decision, only about 15% of teams were actually co-located. Even if everyone returned to an office, most employees would still be going into a building just to join Zoom calls with colleagues somewhere else.

So instead of asking whether offices were inherently good or bad, Twilio established design principles.

- Impact mattered more than physical presence.

- Employees needed to be able to work effectively from anywhere.

- Remote work could expand the available talent pool.

And because connection still mattered, the company redirected some of the money saved on real estate into travel and team gatherings.

“Trust is non-negotiable in an environment where you’re remote first.”

Remote-first doesn’t mean removing human connection. Twilio deliberately invested in off-sites, team building, shared experiences, and opportunities to build relationships because those interactions help create the trust distributed teams need.

And that may be one of the biggest lessons for the AI era too: technology can radically change how work gets done without eliminating the fundamental human need to feel seen, valued, recognized, and connected.

‍