Leader Spotlight: Advising teams through rapid turnaround, with Ariel McNichol

---

Ariel McNichol is the founder of YesCraft.ai, a new kind of digital transformation agency. She began her career in product and design at DIRECTV and has since been a global creative director at PCCW and Lotus Interworks. Ariel has led digital product launches for clients like Disney and Apple and recently led a 120-person product team at CVS Health. She has founded multiple startups including mEgo Avatars and MotiSpark, which creates AI-driven personalized video nudges. Ariel started YesCraft.ai to answer the needs of multiple former employers seeking rapid, hands-on digital transformation and product turnarounds.

In our conversation, Ariel discusses her approach to product turnarounds and digital transformation and how she brings a fresh perspective to enact change. She shares insights on creating shifts in organizational mindsets and structures and how she encourages teams to leverage AI to increase efficiency and innovation.

---

Helping companies remain agile

To get started, could you talk a little about your background and how you got into product management?

I’ve been doing digital product development since the mid-’90s. I’ve been both at agencies and inside co-founding tech companies. I focused on consumer games, entertainment, and SaaS, and then zeroed in on digital health a few years ago with my startup, MotiSpark. After that, I joined CVS Health, but with the current shifting landscape, I’ve gone back to consulting and growing my startup. We help larger companies remain agile in the face of AI and new developments.

When you say helping companies remain agile, is it in the digital health space or a variety of verticals?

It’s a variety. I’m part of a consultancy that does this type of work for larger health-specific entities, but I also do more consumer-facing and nonprofit work. When I say agile, I should specify that we’re not just explaining how to use agile development processes — we’re helping companies create processes that enable large teams to stay innovative. Sometimes, it’s helpful to bring in an outsider who can help to change processes or mindsets.

You’ve been involved in advising product teams for many years. What aspect of the work do you enjoy the most?

I love becoming part of a team and releasing successful products. I love helping turn things around, launch new marketplaces, games, and features, and fix team morale. Unlike traditional advisory that’s mostly observation, I like to integrate within the team and help teams see a rapid turnaround.

I’ll wear multiple hats and encourage others as needed. It’s about goals, not roles. For example, when I joined CVS Health, I started as a UX and strategy lead and then moved to product leadership. Coming from my startup and digital-first companies, I was surprised at how such a large, matrixed organization approaches digital strategy, product, and tech leadership. Unfortunately, a lot of these legacy titans are not utilizing their internal talents to their maximum capacity, and this often creates resentment between design, UX, innovation, and business development. And, layoffs. I hope that our industry sees a rebirth in what it considers roles and responsibilities within digital.

Changing the structure to optimize teams

Can you provide an example of how you implemented organizational or cultural change at CVS Health?

Sure. CVS Health needs to plan headcount and budgets well in advance. Before I was hired, a new team for consumer chatbots was budgeted and staffed to be just like other digital teams — with UX/UI, content strategy, accessibility, frontend, and backend devs, plus the addition of a conversation designer, who was the only team member with experience building bots.

The team’s productivity and morale was low. After talking with everyone privately, with careful empathy, it became very clear that we needed to reconfigure the processes, get their software requests immediately fulfilled, and swap out some team members.

The entire incentive structure rated team members according to how they closed their stories, i.e., finished their planned work. If they couldn’t do their work but could point to another team member who they were waiting on, they were safe. Scrum masters were held accountable to make sure nothing caused a deviation from quarterly planning. It was about points and velocity seen on a report — separate from KPI success. Plus, each discipline reported to a different lead, so often, a person who didn’t have time to understand the team issues had to judge success by reports. The result was a culture that prized roles over goals.

To change this, we had to come up with a success story that leadership would buy into and that we could actually get done. I worked with some other leaders to quietly plan on enabling some team member movement, ensure that the needed software request was immediately fulfilled, and work with the scrum master and then their leader to enable this team to switch to Kanban mid-cycle.

With all this, I made some easy-to-digest collateral about the benefits of our changes, such as how the team was currently on track to fail to reach KPIs by large margins, but how our plan would enable them to catch up by the end of the quarter. We noted that if successful, this team configuration could serve as a playbook for future bot development and demonstrate our organization’s agility. Luckily, the other discipline leaders and senior leadership were smart cookies and immediately not only backed the shift but improved it further.

It sounds like you’re saying that an organization like CVS Health needs to remain flexible even after budgets are set to restructure digital teams to suit the particular work they are undertaking.

Exactly. But, considering how different disciplines are siloed and how budgets work, the solution isn’t easy. It’s like building a house of cards inside a moving car. In a highly matrixed organization, how do we plan headcount knowing we don’t know exactly what we’ll need? Doing layoffs and hiring contractors is too time-consuming.

My dream is that team members are eager to shift hats and learn and prioritize team goals over their roles. No one likes feeling useless, and I’ve seen genius designers step into QA and UAT with brilliance. Why not? In this era, it seems particularly helpful if leadership can convey the psychological safety that combats defensiveness by discipline and encourages “educating up.”

Using LLMs as a tool

You encourage your teams to learn to responsibly use GenAI to assist them in their product responsibilities. Are there other benefits beyond efficiency?

Yes, when GenAI is used properly, it can not only help team members ensure they’re questioning and reviewing their own work an extra time but enable them to do it in a shame-free psychologically safe environment.

We have to ensure everyone is using a safe instance and understands their limitations and dangers. We need everyone to always cross-check suggestions using multiple methods and flagging for human review. Essentially, we don’t want a product owner asking an LLM to write a feature spec filled with hallucinations. However, asking an LLM to review one’s user story to check for breaks in logic, lack of acceptance criteria, or areas that need clarification before handing it off to an engineer has been wildly helpful for people I’ve managed. They learn a lot, work faster, seem happier, and are more approachable.

For example, a PO hands off a story to their engineering partner asking them to double-check the UAT for Android after the LLM helped them write that. They’d say they had cross-checked the suggestions, but ultimately, the engineer would know best. It’s empowering, faster, but still respectful.

In one team, we saw a lot of complaints from engineering and UX about sub-par story writing from product owners. After hands-on training on how to use the LLM, this team experienced a wonderful turnaround and the PO went from feeling behind the times to cutting edge. It’s kind of funny to me — I would think everyone in digital has been playing with LLMs on the side and realizes they’re just tools.

How do you go about ensuring the information is correct, especially if you don’t have a technical background to confirm the AI’s accuracy?

Great question. The first rule is to assume there are hallucinations and get humans who do know these things in the loop. I love to show team members things like asking GPT-4 how many R’s are in the word “raspberry.” GPT-4 says the answer is two (but, in OpenAIs defense, o1 gets this right.)

Claude gets it wrong and then, when it gets it right, hallucinates about how it got it wrong to begin with. Keep prompting the LLMs to review themselves and train everyone to understand that these tools are just they’re eager-to-please generative bots. They say what they think you want, whether it’s true or not.

Further, ask for radical transparency about whenever you use an LLM. You should say, “Hey, I’m handing this over to you. I’m not totally sure about the user acceptance criteria for handsets. I asked Claude, but can you help validate that’s what you need?” I call them your improv and brainstorming partner who lies with total confidence. But, they’re great helpers. That sort of cultural tuning is new to the era we’re in.

Creating cultural change

Changing culture can be hard, and it’s also not a quick process. What approaches have you found to be successful?

I went over to the product side at CVS because there was a vacuum in leadership in our group. We couldn’t build the strategy and UX pilots that I was running because I couldn’t get the organizational needs fulfilled. So, I went over to see the product for all of these things. I found that people were very siloed and resentful. The first thing I did was say, “I need to take all of the nonsense from above me and not have the people below me feel that pressure. Instead, I’m going to define goals for and with them and let them off the hook.”

In general, if you feel like you’ve been underperforming, something is likely wrong. If you’re trying your hardest, then we need to reorganize and look at what your goals are and what you need to achieve them. We don’t want people to feel defensive — we want them to shift to a growth mindset. What are we doing? I always love asking, “If you had a magic wand, what would you change?” That can often release amazing ideas.

Specifically regarding consulting work, it’s really interesting when you privately ask team members what they would change. I posed it to a startup that was struggling. I built a form for employees to share feedback. I told them I wouldn’t share their comments — I just wanted to understand. It was amazing. At the end of the forum, I asked, “Would you like me to share this with other people?” And they all said, “Yes. I wish people would listen to me.” That uncovered the root of a lot of their issues. They felt unheard.

Do you find that teams have trouble figuring out how to track their progress, and if so, how do you help a team determine which metrics to look at?

First, they need software to organize and track their work and use it properly, of course. Jira, Rally, Asana, or ClickUp — it doesn’t matter much which one. But, what does matter is the integrity of the work inside of those tools. Performance needs to be based on data, not tracking internal progress alone. We need all team members to see the metrics that matter in their workflows.

I like to set up Slack or Teams to show feeds of key metrics. For example, I started getting calls-to-care data and email feedback into our scrum CVS teams channels. And, of course, the general analytics data that’s relevant for each scrum team. It’s an easy and powerful way to be transparent with everyone. It can be so empowering.

Regarding helping teams determine which metrics to look at, I was added to a team that had a section of their website dedicated to updating profile information. It was poorly designed and it was obvious when I went to the site for the first time. I was put in charge of that team, and I said, “Yay, we’re going to fix this!.”

Somehow, the UX research team looked at the data, saw long times on a page and multiple clicks, and interpreted this as users liking the way it was and successful jobs done. We lacked proper tagging to quantify key data, but it was clear they had failed to dog-food the product. I forced them to revisit the call center data and email feedback that made it clear that people couldn’t update their profile information.

All this to say, always try to use your product and understand what it feels like on a gut level. Was it easy or frustrating? If you look at data wrong, it can completely mislead you. You’ve got to make sure you look at the big picture and keep asking questions. Try to break your assumptions.

Giving leaders a new perspective

We talked about leadership briefly. In your role, how do you encourage trust between the C-suite and their teams, and why is that important?

This is where the product leader or advisor’s role of managing up and down comes in. I’ll start by saying that usually, a founder is so in love with their product that, if it’s not working well, they’ll say something in their team is not right. It’s a lot easier to blame downward. It’s easier to do that than open up the can of worms of what is actually going wrong.

For example, a CEO tells me his CTO and product leaders are lazy. Then I find that they’re not, but they are resentful and confused. They put tons of love into previous features, and suddenly, plans changed without any understanding as to why. They don’t feel respected. As an outsider who comes in and sees this, I help the CEO understand and rebuild that trust. It’s kind of funny how it feels like family therapy when all you’re doing is better new communication channels.

Do you feel like leaders are often receptive to your findings? How do you view the relationship between the advisor and the executive team?

It can vary and requires a benevolent kind of psych ops. I came into a situation where a founder was spending 80 percent of their budget on discovery in that area that had nothing to do with what they sold the investors or the core technology. Sadly, the effort wasn’t paying off and the choice seemed delusional. The founder was focused on proving that he hadn’t wasted money on this discovery. How do you help the founder? It’s hard.

In this case, I first had the founder watch me use the experiences over Zoom, and it was horrible. It was very clear by the end of that call that the experiment that had been developed thus far wasn’t working. It was painful for the founder to realize and I empathetically pre-empted shame by saying how frustrated he must have been because it was a good idea, etc.

I said to him privately “You can take this with a grain of salt. You don’t know me well, I’m just coming in from afar. But if I were your investor, I’d tell you to stop going down this hole because even if you got it working, you don’t know if your users even want this and meanwhile your core tech isn’t ready.” Calling them out respectfully and in private is important. Also, to be an outsider means I don’t have anything to lose by saying this. That’s the beauty of the advisor role. Your intentions are clear when you’re actively trying to help.