Artera

Coaching is vital for cultivating a culture of ownership, empowerment and leadership within the team. By prompting discussions, asking questions and guiding individuals to find solutions independently, team members develop a mindset of proactivity and accountability.

I believe that leadership is a mindset and is accessible to every employee at the company, regardless of whether they're just starting out in their career or have been in the field for more than 20 years. Not only does this mindset lead to a positive team culture where employees feel empowered and motivated to take initiative, but it makes them happier and more likely to stay at the company, and they generally have more fun with their work.

Ultimately, encouraging a leadership mindset across all levels fosters a collaborative environment where decision-making is shared and ownership is embraced by everyone, resulting in increased productivity and employee satisfaction.

For engineering managers looking to instill healthy communication habits, embracing a coaching approach can be transformative. Take the time to learn more about coaching by reading, taking workshops or engaging with a coach, all of which can be incredibly beneficial for the whole team.

By prioritizing asking questions over providing immediate answers, managers can empower their teams to think critically, solve problems effectively and grow into effective leaders themselves. Creating a space for open dialogue and encouraging continuous learning can enhance team dynamics and drive success.

Through coaching, managers can guide their teams to tackle challenges systematically, prioritize tasks and foster a culture of continuous improvement, leading to reduced stress, clearer decision-making and enhanced team performance.

We are currently building fully autonomous AI agents for healthcare providers. These agents leverage the leading large language, text-to-speech, speech-to-text and generative voice models to automate complex patient communications, engaging patients in a dynamic and realistic manner.

Our agents are highly customizable, including voice and speed options, among others, and feature multi-language detection, ensuring they can seamlessly engage diverse patient populations. Our AI agents can handle a wide range of use cases, such as rescheduling, cancelling or confirming appointments.

By using AI for routine administrative tasks, providers can reduce staff burnout, optimize efficiency and lower costs.

Artera engineering has successfully integrated AI into its development processes, starting with Microsoft Copilot in our VSCode IDEs and expanding to include Claude tooling. We utilize Claude agents for various tasks, such as PR reviews, script and Jest test generation, code research, error investigation and "vibe coding," where agents build fundamental functions and classes based on precise prompts.

We're seeing positive results as we continue to refine our prompts and as LLMs improve. Additionally, we're using the agents to handle some of the more procedural work, like generating comprehensive pull request descriptions in a matter of seconds, which frees developers from an important but time-consuming task.

With all that said, it is crucial to note that a human review is required for all AI-generated content. AI may not always produce optimal code or understand the best approaches, so our team must review and iterate on its output.

At the moment, I'm particularly interested in using AI to create integration and unit tests for our React app. Crafting React component tests has always felt somewhat tedious, especially when it comes to mocking needed and identifying all the branching cases.

While generating component tests with AI can be challenging, I've found it proficient at recognizing branching logic, though the implementation often results in overly complex test code. To address this, we're developing clear guideline files — CLAUDE.

md files in the code base — that Claude will follow based on our prompts, so tests are built with a consistent set of rules. We're still learning and growing, but it's getting better at a pretty fast pace. Above all, AI serves as a powerful tool, particularly for managing repetitive tasks and enhancing code quality with innovative solutions I may not have considered.

While it's great for quick checks and speeding up my workflow, I don't rely on it entirely. Instead, I use it as an enhancement to my process. At the end of the day, it's just a tool to support my work, not replace it.

It starts with making sure that all teams are aligned from the top down regarding their goals and success metrics. When each team has a different understanding of what success looks like, teams will naturally start to diverge from the roadmap. Getting that initial alignment on what our near- and long-term goals are helps us create a roadmap that we're confident we can deliver on.

Data work is a great example of where this can become extremely challenging to manage. Often a horizontal that's impacted by other product teams upstream, it can be challenging to keep everyone aligned. To help keep the data teams abreast of specific needs, we do a quick biweekly meeting in a round-robin style with the design and product management teams to create general awareness of what everyone is working on.

Like many organizations, our teams participate in the quarterly planning process. A crucial exercise we've identified is having every team articulate priorities and dependencies to the entire innovation organization after planning. As a mostly remote team, this has helped us have one last check to identify gaps in dependencies or areas of concern.

As we progress throughout the quarters, we frequently connect with our teams through demos, regular meetings and one-on-ones to stay aligned with our goals and communicate any risks. For large cross-functional projects, we often use weekly or biweekly status reports to help create continued alignment across many different stakeholders and allow those stakeholders to engage with our product teams asynchronously.

Project needs can absolutely change during the development process. When this happens, we like to start with understanding the "why," breaking down the root cause of that change. For example, if a new technical dependency or blocker impacts the functional goal unexpectedly, we work closely with our design team to try and mitigate any negative impact on the user experience.

We're extremely fortunate to have an incredible design team that's both knowledgeable in foundational design principles and technically savvy. They're able to help our product team navigate the changes without sacrificing experience to achieve the original goal.

If the overall goal of a project needs to change, we try to overcommunicate. As product leaders, we prioritize sharing feedback from our executive teams and customers directly with our teams, but we also make time to ensure that our stakeholders are aware of any major changes as well.

Meeting monthly with a core group of informed leaders in the company helps us pressure-test the decisions we're making as product leaders and helps us keep the company informed on any critical roadmap changes.