Conversational AI Success in HR: Key Takeaways from Our Webinar

On November 13, 2025, ChartHop hosted a dynamic conversation, "How to Embed Conversational AI into Your Work," featuring two of our valued customers: Cerys Cook, Chief People & Transformation Officer at Swift Medical , and Will Chung, Director, Global Compensation & Operations at GumGum.

The discussion, moderated by ChartHop’s Head of Product, Justin Garrity , explored the real-world application of conversational AI in HR—from overcoming the fear of the "blank screen" to leveraging AI for strategic talent management and organizational change.

Here are the key takeaways and insights shared by our guests on their journeys with AI adoption.

‍Cerys Cook’s Journey: From the Kitchen to the Cutting Edge

Cerys Cook shared a truly unconventional and inspirational path into HR. She began her career as a chef , describing the professional kitchen 30 years ago as a "really, really tough environment" marked by long hours, volatile conditions, and rampant sexism.

Feeling frustrated and realizing that "workplaces shouldn't and didn't have to feel like that" , Cerys eventually transitioned to an office manager role where she "naturally started doing the people stuff". This evolved into a dedicated HR career, which she describes as the first time she felt like she was "meant to be" at work.

Adopting AI: The Blank Screen Moment

Cerys’s approach to AI adoption has been one of early enthusiasm followed by a thoughtful, disciplined reset. She was an early adopter, playing with ChatGPT the week it came out because it "felt like magic".

However, she realized she had gone "too far" when she found herself "sitting in front of a blank screen just going, 'I don't even know where to start without using AI'". This moment spurred a critical change in her workflow: She resolved never to "outsource my thinking or my voice".

‍

Her new, disciplined approach:

1. Write First, Critique Second: She now drafts everything herself first, using a blank sheet of paper, and then lets AI critique the work.
2. Secure Environment for Sensitive Data: She relies on contained tools like Ask ChartHop and her internal tool, Swift DAI , because they operate on their own data and company context, ensuring "no risk of our company in- information really leaking out into the internet". She warns that if you don't have contained AI tools, your information is "everywhere".
3. Prompting for Dissent: When using conversational AI, she treats it like a "teammate" and actively prompts it to challenge her assumptions and check for blind spots. She asks questions like, "What might not land here? What's unclear? What am I missing?" and "Is there any bias in here that I'm not noticing?". She is also "really relentless now about sources," demanding citations for any stats.

‍

Will Chung: A Compensation Expert Embraces Agentic AI

Will Chung’s career began in a junior sales role, but he pivoted into HR after a recruiter introduced him to a job in Talent Acquisition (TA). After years in recruiting, he found his passion in Compensation , a field he loves for its depth (executive comp, equity, incentives) and its broad, company-wide impact.

GumGum, an advertising technology company , embraced AI early, quickly securing enterprise licenses to encourage widespread adoption while protecting data. Will’s personal journey began with playful experimentation (using AI as his own travel agent for a trip to Tulum) and evolved into using it as a tool for accelerated learning.

AI-Proofing Talent Assessment

A highlight of Will’s AI journey was using it to revolutionize his hiring process for an analyst role. He was nervous that candidates would simply be "loading this into an AI tool, having it spell it exactly, be 100% right".

To combat this, he did something ingenious: He used AI to assess how he wanted to assess the candidates, effectively "AI-proofing" the screening process.

The AI suggested creating "questions and prompts that don't always have a right answer". This forces candidates to make judgment calls, allowing Will to "engage with, like, how someone's thinking about something, and not just, like, 1 plus 1 equals 2".

Using ChartHop AI for Onboarding:

After the hire, Will leveraged ChartHop AI to onboard and train the new analyst. Because the tool is a "closed environment" with protected people data , it can help the analyst learn GumGum's data set and ChartHop systems, showing them "where to click and different toggles". This frees up Will’s time to focus on strategic relationship-building.

‍

The Need for Execution Guarantees

Will also shared his experience building custom projects (Gemini’s "gems") to streamline data auditing , focusing on the "garbage in, garbage out" concept.

He uses "execution guarantees" written into the prompt to ensure a consistent, structured output. He specifies conditions like, “don't consider this query done until XYZ is met... until you produce a downloadable Excel file with three tabs that include X, Y, and Z on each tab". This is essential because the AI’s "confidence level in the tone is always so high" even when the answer is wrong.

ChartHop on Product Philosophy: Consistency & Control

Justin Garrity weighed in on how ChartHop addresses the tension between human judgment and machine output, and questions about inconsistent outcomes with any conversational AI tool. Justin highlighted that one of the most significant impacts of AI has been reducing the complexity of customization. ChartHop has always aimed to be the most customizable and configurable HR platform, but prior to AI, achieving this required many options across different parts of the interface, which could be "challenging or overwhelming".

• Take the Edge off Complexity: ChartHop AI has taken the edge off that complexity to configuration by allowing users to request configuration updates or changes through plain language, which ChartHop then interprets and translates into its query language.
• Solving Inconsistency: ChartHop’s conversational AI, Ask ChartHop not only provides a natural language response in the chat window but also redirects and manipulates the core UI to create the results. For instance, if a user asks for a chart or a table of employee data, the UI will generate that result with "very little variance, if any" , even if the chat analysis might change from time to time. This provides the user with both the innovation of the AI chat and the consistency of the UI response.‍
• The AI-First Interface: Justin noted that AI has removed the complexity of customization. Increasingly, new users are making the AI interface their "primary and first interface to our platform", using plain language requests that translate into configuration.

The Future: A New Org Chart

The conversation closed with the philosophical future of HR and AI.

• Augmentation over Replacement: Will believes the partnership between human and machine is crucial , seeing AI as a tool that "serves up the data" and shortens the time spent pulling data, thus creating "more time for the person to assess, make judgment calls, that sort of thing".

• The Fear of Internal Tools: Cerys noted the slow adoption of their internal AI tool (Swift DAI) is not fear of replacement, but resistance rooted in privacy concerns. Employees are hesitant to use the internal system because they wonder, "Can they see the questions I'm asking it?" or "Can they see that I don't know this thing?". This highlights the deep need for a change management approach that addresses trust and vulnerability.

• AI as a "Buddy": Will’s product wish was for an AI assistant that truly "learns over time" and acts as a "buddy" that gets "smarter with you over time in your whole career there". This would help solve the issue of lost organizational and individual memory as people change roles or leave a company.

Cerys wrapped up with a profound thought-provoking question for the audience:

"In a world where like, we truly have agentic AI embedded in how we work, like, what might even an org chart look like that captures that organizational capacity?". "But you know, 50 years from now, that one box could represent a million agentic AI support".

This shifted the conversation to the role of the future HR leader, who will be "engineering those conditions for human potential and machine intelligence to coexist" through "human systems design in a really profound way".

Go Deeper: Access the Full Webinar Conversation

Want to Explore Conversational AI for Your HR Workflow?

Cerys and Will are successfully using ChartHop AI to transform their teams, from protecting sensitive data to building a more efficient and strategic HR function.

Would you like to speak with a ChartHop expert about incorporating secure, conversational AI into your organization’s people operations platform? Reach out here.