Key takeaways:
- Spark is Torch’s AI agent that helps leaders practice difficult conversations and build capabilities in realistic simulations
- A built-in 360 shows leaders how they’re showing up, with feedback from managers, peers, and direct reports
- Use it alongside human coaches for complete transformation, or deploy it standalone to scale development across your organization
- Leaders can rehearse high-stakes moments, get real-time feedback, and share insights with their coaches to go deeper
- Spark adapts to your organization’s context, tying guidance to your specific goals and challenges
Intro
Most leadership training doesn’t work. People learn something in a workshop or coaching session, and within a week, 75% of it is gone. Not because leaders aren’t motivated, but because there’s no place to practice before it counts.
Spark is Torch’s always-on AI coaching experience. It gives leaders a place to work through challenges, prepare for tough conversations, and build leadership capacities that actually stick. And unlike generic AI tools, Spark learns your organization’s priorities, values, and leadership expectations, then builds on what it knows about each user over time.
Grounded in your context
Generic AI coaching tools treat every user the same regardless of role or goals.
Spark is different. It grounds every conversation in your organization’s leadership capacities, values, priorities, and expectations. And it builds a memory of each user: their role, their challenges, their goals, and patterns that emerge over time.
As leaders engage, Spark develops an understanding of how they show up in their role, what they’re navigating inside the organization, and where progress and friction tend to emerge. A first-time manager preparing for a performance conversation gets different support than a senior leader navigating a restructuring.
This is coaching grounded in your reality, not a library of generic prompts.
Practice for the moments that matter
Research shows only 10% of leadership training produces meaningful behavior change. You learn something in a session, then struggle to apply it when the pressure’s on.
Spark solves this by letting leaders practice before the stakes are real. You rehearse tough conversations with a realistic AI character who responds based on how the conversation unfolds. Spark coaches you on the side with real-time feedback. Preparing to give tough feedback to an underperforming team member? Practice it first. Navigating resistance to a transformation initiative? Work through the conversation multiple times, testing different approaches.
These aren’t scripted interactions. The AI character adapts in real time to how you engage with it; a manager who hesitates will experience different responses than one who takes a direct approach. The scenario shifts based on your choices, so the practice feels real.
Afterward, you get a clear breakdown of what landed, what didn’t, and what to try next.
See how you’re showing up
Spark can also include a 360 so leaders can hear directly from the people they work with. We partner with your organization to define the leadership capacities that are most aligned with your business goals and values, things like adaptive mindset, learning agility, and trust building. Then managers, peers, and direct reports weigh in on how that leader is showing up.
After the 360 is complete, the results feed back into Spark, so when a leader practices a tough conversation or works through a challenge, Spark already knows where they’re strong and where they’re still growing. And, admins get a separate report that surfaces patterns across the organization, creating powerful data.
Feedback that builds capacities
The feedback isn’t generic advice. Spark provides a personalized breakdown of what your responses signaled, how your behaviors aligned to leadership standards, and specific strategies for strengthening your impact.
This is fundamentally different from a chatbot. Spark guides you to your own conclusions through adaptive questioning and scenario-based learning. It helps you develop instincts that transfer to real leadership situations.
Connected to coaching (or powerful on its own)
Spark is designed to integrate with human coaching, but it doesn’t require it.
When paired with coaches, Spark creates a continuous feedback loop. As you engage in conversations or with AI role-plays, patterns surface about where you’re gaining confidence and where you’re still stuck. You can share those insights directly with your coach, so they can adjust their approach, go deeper on barriers, and reinforce what’s working.
Human coaches handle the complex, contextual work of transformation while Spark handles practice and strengthening skills that ladder up to big-picture leadership capacities. DDI’s 2025 Global Leadership Forecast found that organizations using five or more development approaches are 4.9 times more likely to report that their programs actually improve leadership capabilities. Spark adds a critical layer most programs are missing.
When used on its own, Spark gives every employee access to meaningful development conversations, anytime, anywhere. Gallup’s 2025 State of the Global Workplace reports that less than half of managers worldwide have received any management training, and Spark closes that gap. Leaders who don’t have dedicated coaches can still rehearse tough conversations, get personalized feedback, and build capabilities tied to organizational goals.
For a deeper look at how AI coaching and human coaching work together, see our P.O.V. on what it means to combine human and AI capabilities.
Who Spark is for
Spark works for organizations rolling out transformation initiatives and for individual leaders who need support between coaching sessions.
At the org level:
- Companies undergoing AI adoption, M&A, or restructuring
- L&D teams looking to scale development beyond those with dedicated coaches
- Leaders who need to extend growth opportunities to more team members
For individuals:
- Leaders at all levels preparing for high-stakes conversations
- Individual contributors navigating change
- Anyone between coaching sessions who needs practice or guidance
What makes Spark different
Most AI coaching tools are transactional. You ask a question, you get an answer, and nothing connects to what happens next. There’s no memory of your goals, no awareness of your organization’s priorities, and no way to tell whether anything actually changed.Spark works differently.
Your practice sessions build on each other, and feedback ties directly to your development areas and organizational context. Instead of measuring how often you logged in, Spark tracks whether your behaviors are shifting. Organizations that embed systematic coaching record 25% stronger business outcomes than their peers, and Spark brings that rigor to AI-powered development.
Privacy and security
Research shows that 56% of monitored employees report stress or tension at work. When people don’t trust how their data is being used, they hold back.
Spark is built differently. All coaching conversations are private and confidential. Individual-level data is never shared with employers or managers, and organizations don’t have access to raw conversations.
What organizations do see are themes and trends across the user base, not individual coaching data. This surfaces patterns that can inform leadership development strategy without compromising trust.
Spark does not use customer data to train our models or any third-party AI models. Data is used only to deliver the coaching experience, and Spark is designed to meet enterprise security and privacy expectations.
