Niki Kohari

COO at + biz ops consultant, leadership coach, happiness trainer, and mindfulness facilitator at BrightSpark Coaching

[Niki Kohari @nikibeth]

The most underrated capability in the AI era is knowing when to trust the output and when to push back!

We're spending a lot on helping workers produce with AI, yet we're spending almost none on helping them evaluate the output. Maybe this is where management will go...

[Niki Kohari @nikibeth]

Because AI is a new organizational shift, it falls into a trap.

Learning is occurring at the individual level, but they typically keep it to themselves. It's so early in the transition to AI that there's no system for sharing and learning together.

[Niki Kohari @nikibeth]

I've been thinking about how meaning at work has always been tied to contribution.

The sense that what you did required something from you specifically.

AI changes how the output is generated, and it's likely to impact how we think about and measure contribution. But, how?

[Niki Kohari @nikibeth]

Everyone seems to be talking about how to use AI tools. The more important question is, how do you evaluate the output? Who has the domain expertise to push back confidently when the AI is wrong? What context is missing from the output? What questions do we ask next?

[Niki Kohari @nikibeth]

Managing an AI-affected team requires a new set of skills, such as helping people re-anchor their sense of contribution, holding space for the learning curve, and building an environment where honest feedback on AI outputs is normal.

How are you supporting this transition?

[Niki Kohari @nikibeth]

I've been at companies where mistakes were readily shared from leadership on down. If your team is only sharing AI wins, you have a problem.

The most valuable signal in any AI-enabled team is when someone says, "Here's where it went wrong and here's how I caught it."

[Niki Kohari @nikibeth]

The most effective AI training isn't in your internal LMS.

It comes from one person on a team saying, "Let me show you what I figured out," and a culture where that kind of sharing was normal and valued!

Informal learning has always been the best form of knowledge transfer.

[Niki Kohari @nikibeth]

No one talks about the 90-day dip.

When knowledge workers first adopt AI tools, productivity often declines before it rises! New workflows and judgment calls are layered on top of an already full plate.

It's documented in every technology transition in organizational research.

[Niki Kohari @nikibeth]

AI rollouts fail because someone bought a tool, sent a Slack message, and called it change management.

They need to create psychological safety and reinforce a sense of identity, not just create an IT checklist.

[Niki Kohari @nikibeth]

I used to study identity back in grad school, and a recent experience had me reflecting.

A senior leader told me recently that she felt "deskilled" by AI.

She wasn't slower. Her manager loved the results.

But she felt like she was operating someone else's work, not her own.

[Niki Kohari @nikibeth]

AI shifts cognitive load; it does not remove it. You might stop drafting, but now you're editing, fact-checking, and deciding when to trust the output. For knowledge workers already at capacity, that trade isn't automatically an increase in perceived productivity.

[Niki Kohari @nikibeth]

Behavior change at work spreads peer-to-peer, not top-down. If you want your team to actually adopt AI and not just tolerate it, find the person who's quietly getting results and give them a way to experiment and share. It doesn't have to be a formal comms strategy.

[Niki Kohari @nikibeth]

Most AI anxiety at work isn't about job loss. It's about identity loss. When your expertise becomes a prompt, you need a new answer to "what do I actually bring?" We have to connect back with our purpose and what brings us meaning at work.

[Niki Kohari @nikibeth]

Your employees aren't learning AI from training decks. They're learning it from the person two desks over who figured something out and shared it. The smartest thing a team can do is make sharing systematic—not accidental.

[Niki Kohari @nikibeth]

AI doesn't automatically reduce cognitive load. For most knowledge workers right now, it adds new tools, new judgment calls, and new failure modes to watch for. The productivity gains come after the learning curve.

[Niki Kohari @nikibeth]

We’ve heard stories of other companies accidentally exposing private customer data to the internet via vibe-coded dashboards and organizations getting stuck between AI pilot programs and bringing tooling to everyone. We want to fix this and other AI-related problems!

[Niki Kohari @nikibeth]

At Ardent, we’re building a general-purpose AI agent for knowledge workers and coming up on private beta. I’m hoping to connect with those in my network, get their impressions of what we’re building, and hear about the struggles you’re having with scaling AI at your company.

[Niki Kohari @nikibeth]

Our approach is a little different, with a focus on:

- True customization using your unique internal data, tools, and workflows
- Security and privacy that you expect from enterprise software
- Collaborating with others in the app to share and scale AI knowledge across your team

[Niki Kohari @nikibeth]

Ardent learns what tools you use and can build its own extensions to connect itself to the systems you’re already using!

[Niki Kohari @nikibeth]

I’m not selling, just learning! Please come chat with us: https://t.co/uDC1qBxCaV.