How Generative AI is Changing Search and Website Experiences
5 février 2025 • 2 Minute Read • Elizabeth Spranzani, Chef de la direction technologique
See All Our Insights
Verndale x AI Agents: Building the Future at Our 2025 Hackathon
21 mai 2025 • 5 Minute Read • Elizabeth Spranzani, Chef de la direction technologique
We've always kept up with new tech, experimenting and exploring how to turn new capabilities into real value for our teams and clients. We've tracked AI closely as it evolved from a futuristic concept into a practical toolset.
As always, we’re testing, implementing, pushing its limits, and having fun along the way, especially during our hackathons.
One of the most exciting breakthroughs in AI is the rise of agentic AI, a technology that goes beyond answering questions and actually takes action.
Hackathons have always fueled innovation at Verndale. This year, we raised the bar, focusing entirely on AI Agents and inviting teams across disciplines to imagine, prototype, and pitch agents that would allow us to work smarter and deliver more value.
The results were nothing short of inspiring, filled with creative ideas, practical solutions, and that unmistakable feeling of "oh wow, the future's already here."
Here's a behind-the-scenes look at Verndale's Hackathon 2025.
What is Agentic AI and Why Now?
Agentic AI refers to a new class of AI applications that don't just sit around waiting to be prompted. These agents combine large language models (LLMs) with real business context and tools that can autonomously make decisions, take steps, and follow through.
Think of these AI agents as the overachievers of the AI world. They're capable, autonomous, and always taking initiative.
What sets agentic AI apart from traditional automation or even AI-assisted workflows is its ability to operate with higher autonomy, adaptability, and decision-making complexity.
Unlike traditional AI or chat-based assistants, agentic systems can perform tasks like:
- Evaluate project scopes
- Generate business specifications
- Run an analysis on websites
- Gather insights from unstructured documentation
- Organize your strategic recommendations
And as tools like Microsoft Copilot Studio, LangChain, OpenAI’s Agent Builder, and Salesforce Agentforce mature, the barrier to building these experiences is lower than ever.
But new tools don’t deliver transformation on their own—people do. That’s why Verndale launched the hackathon—to put innovation directly into the hands of the curious, the creative, and the collaborative.
Hackathon Purpose: A Culture of Learning and Action
We designed the AI Agent Hackathon with four key goals:
- Inspire: Spark fresh thinking about how AI agents could reshape internal operations, client services, and web experiences.
- Learn: Provide hands-on experience with emerging AI tools, platforms, and frameworks.
- Collaborate: Break down silos by encouraging cross-functional team formation and knowledge sharing.
- Ideate: Surface meaningful, practical ideas that could evolve into tools Verndale actually implements or pitches to clients.
Participants could select from three categories:
- Website Experience
- Verndale Process Improvement
- Client Process Improvement
The rules were simple: come up with an idea, form a team, and submit a short written proposal and a five-minute video explaining the concept. A working proof-of-concept wasn’t required, but it would earn bonus points with the judges.
By the Numbers: Participation and Scope
The response was incredible.
- 20 teams submitted complete entries
- 60+ Verndalians participated across departments, including engineering, experience design, marketing, project management, operations, sales, and client strategy
- Participants spanned the US, Canada, Ecuador, and Brazil
- More than 15 distinct platforms were used across submissions
While many teams built agents focused on internal operational improvements, others tackled external-facing experiences, from self-service ecommerce chat to dynamic CTA personalization. A number of entries blurred the lines between categories, with agents that not only streamlined workflows but also directly elevated the user experience.
How It Worked: The Hackathon Framework
The hackathon kicked off in early April, with a four-week sprint to submission. The process was structured but flexible, designed to give participants room to experiment while guiding them through essential steps:
- Team Formation: Employees formed teams of 2 to 4 people, with bonus points awarded for cross-department collaboration.
- Inspiration & Research: Participants explored examples of agentic AI in the wild and attended internal Q&A sessions.
- Planning: Teams selected their category and documented the agent’s purpose, value, solution, and technical approach.
- Build (Optional): Teams could build prototypes, demos, or full-fledged agent flows in sandbox environments.
- Submission: Each team submitted a proposal and a 5-minute video explaining the concept. Optional demos were encouraged but not required.
We encouraged teams to use whatever platforms best suited their use cases, including the following technology highlights:
- Enterprise AI builders: Microsoft Copilot Studio, Salesforce Agentforce, Sitecore Stream, Optimizely Opal, and Atlassian Rovo
- Custom LLM workflows: OpenAI, Gemini, LangChain, Mistral, Pinecone, and Flowise
- Low-code/no-code tools: n8n.io, Streamlit, and Replit
- Front-end deployment platforms: Vercel, Netlify, Google Cloud, and Microsoft Azure
- Collaboration: Slack integrations, Microsoft SharePoint sync, Jira/Confluence indexing
The only requirement was that there be no production environments. This was a space for sandboxed innovation—safe, experimental, and employee-led.
What They Built: Themes and Technology
While the entries varied widely in function and scope, several clear themes emerged:
1. Documentation Automation
Multiple teams tackled the persistent challenge of documentation, whether in Confluence, GitHub, or SharePoint. These agents helped surface relevant content, generate specs, summarize functional details, and reduce time spent on "digital archaeology."
2. Recruiting and Talent Optimization
Some submissions reimagined the talent acquisition process with agents that sourced candidates from LinkedIn, analyzed CVs, conducted automated interviews, and even generated structured feedback for hiring managers.
3. Content Generation & Migration
From AI-powered case study authors to agents that streamline CMS migration content tagging, teams built tools that reduce the manual burden of content-heavy workflows.
4. Ecommerce & Lead Engagement
On the website experience front, agents were developed to support everything from interactive event booth kiosks to conversational ecommerce interfaces that guided users through product discovery and cart management.
5. Strategic Decision Support
Some of the most advanced agents were virtual strategists—tools that combined Salesforce data, historical performance, and business logic to generate idea-starters for account plans, prioritize roadmaps, or propose new tools for adoption.
The Judging: Scoring Creativity, Feasibility, and Impact
All submissions were reviewed by a panel of internal judges, who scored each project against six weighted criteria:
The top 5 scoring teams presented their entries live in mid-May for a final round of judging, with cash prizes awarded for:
-
Top Overall Agent
-
Best Operational Idea
-
Best Website Experience Innovation
All participants with valid entries will also receive a significant number of points in our employee rewards program as recognition for their contribution.
What It Means for Verndale: Building with Momentum
AI isn't a passing trend—and at Verndale, we recognize that. The hackathon goes deeper than a showcase of technical creativity; it's a reflection of what makes Verndale, Verndale:
- A company where exploration is encouraged
- A team that thrives on collaboration across roles and borders
- A culture that balances client delivery with internal innovation
As agentic AI continues to shape the industry, we're focused on building the muscle to use the tools with intent. That means training our teams and experimenting with frameworks. While vendors race to release AI features, Verndale is investigating how it can actually help.
What's Next: We're Just Getting Started
Some of the ideas from the hackathon are already being tested, adapted, or piloted with clients. Others are becoming prototypes, pitch decks, or springboards for what we'll try next.
AI is moving fast, and the companies that thrive will be the ones that help shape what comes next. This hackathon showed that Verndale is more than ready to do just that.
Thank you to everyone who built, supported, or joined the ride. You're helping us build what's next.
If you're interested in learning more or piloting one of these programs, get in touch with your Verndale account manager or one of our experts.