OnBloom

landing page
employee database
onboarding initiation
slack bot personalization

Inspiration

We have all experienced onboarding at a new job. It's usually a 3-day or a week long event filled with meetings (aka information dumps) then you are left on your own to find information that’s actually relevant to your job.

You start reaching out to people for document and tool access, raising tickets and sending slack messages hoping they respond. By the time you gain access to the required documents and tools - it's been a week. And you still haven’t even connected with your assigned ‘buddy’.

Unfortunately onboarding is seen as just another box to tick for most organizations but the cost of not doing so is humongous - nearly half of new hires have second thoughts about their workplace during onboarding. Is that the kind of first impressions companies want to leave?

We felt that this was a great use-case to utilize Qloo’s taste API to deliver a hyper personalized onboarding experience to new employees aligned with their interests. Combine it with the power of LLMs - we have the perfect recipe to delight new hires and create a memorable impression.

What it does

🎁 Personalized Corporate Gifting: Treat each employee as an individual. Qloo's API analyzes interests and cultural preferences to recommend meaningful welcome gifts that actually resonate.

🤖 Intelligent Slack Bot: Automatically pings relevant team members when new hires join, offering culturally-aware conversation starters and connection suggestions based on shared interests.

📊 Visual Onboarding Canvas: Clear visual overview of exactly what new hires need: which people to meet, what processes to learn, and which documents to read.

📚 Notion Integration: Complete visibility into your company wiki. OnBloom's AI understands your documentation and provides new hires with the right information at the right time.

How we built it

Core Architecture

Completely serverless architecture for instant scaling and cost efficiency.
Data Sources: Notion serves as our primary knowledge base, providing structured company documentation and employee information.
Visual Workflows: React Flow powers our interactive onboarding canvas, giving users drag-and-drop simplicity for complex workflow management.
Qloo API: Cultural intelligence for personalized hangout spot recommendations and interest-based employee matching.

We use different LLMs for their specific strengths:

Claude 3.5 Sonnet: Relationship analysis between individuals, both direct connections and indirect cultural affinities.
Llama 3: Generates personalized messages on Slack.
Gemini 2.5 Flash: Summarizes Slack conversations for context-aware interactions.
Gemini 2.5 Pro: Creates role-specific onboarding training videos using built-in search capabilities.
Perplexity API: Powers shopping recommendations for personalized corporate gifting.

Technical Stack

Frontend: Next.js with shadcn/ui components and custom theming for consistent design language.
Slack Integration: Bolt framework for building responsive, interactive Slack bots.
Caching: Redis for fast conversation history retrieval and session management.

Challenges we overcame

Notion became our go-to source for docs and employee data, but API limits quickly became a roadblock. We had to get creative with data syncing, batching, and caching so new hires could always get the latest info without us spamming Notion or hitting hard stops.
Using Qloo’s taste API as a utility for our LLMs was new territory. It lets us suggest onboarding content, gifts, or intros that fit each new employee’s interests, not just “what everyone gets.” The technical wrinkles were worth it to move beyond generic onboarding.
Slack’s Bolt framework made building bot features fast, but deploying wasn’t smooth, especially with socket mode. It forced us off our pure serverless setup for part of the project, so we had to juggle flexibility versus operational simplicity.
Choosing the Right LLMs for Each Job: We strategically choose LLMs, Llama for cost-effective Slack messages, Claude for mapping relationships, Gemini for contextual summaries and video generation. This match-up kept costs down and quality high.

What we are proud of

We may be the first to use Qloo’s recommendation engine to connect people and not just recommend content. This opens entirely new market opportunities for cultural intelligence in HR tech.
We matched each LLM to its best use case, dialing up innovation to deliver a smarter onboarding experience.
Beyond building, we prioritized talking to real users and prospects, getting direct feedback before finishing the product.
We secured a verbal commitment for a pilot program from a startup founder.
Taking a crack at solving the really hard problem of employee experience and on-boarding

What we learned

Most organizations have simply "accepted" that onboarding is broken, it's seen as a necessary mess rather than a solvable problem
Team leads, not HR departments, carry the real burden of new hire integration
Documentation is consistently missing or scattered across multiple tools
IT leaders desperately want full automation to eliminate HR bottlenecks entirely

What Resonates with users

Personalized corporate gifting got enthusiastic responses from everyone we spoke to
Single source of truth was universally desired but rarely achieved
Integration with existing tools was absolutely non-negotiable—no one wants another standalone platform
Cultural connections were seen as genuinely valuable, not just a nice-to-have feature

Things to consider as we continue building OnBloom

Onboarding is discrete, not continuous; this makes pricing like traditional SaaS models challenging.
Corporate gifting trends are declining, but personalized experiences still hold value
Enterprises increasingly want to build internal AI tools rather than buy external solutions - as services + product approach is required
Most organizations lack dedicated HR until they reach 30-50 employees
The problem involves far more touchpoints, stakeholders and processes than we initially anticipated

What's Next for OnBloom 🚀

More prospect interviews with HR professionals, team leads, and recent hires to understand the problem (and solution) better
Run pilot programs with organization to test product effectiveness in real life scenario
Refine and strengthen core product workflows based on pilot feedback
Explore extending beyond onboarding into ongoing team building and employee engagement
Expand integrations with existing employee experience tools
Open API access for HRMS tools to increase adoption
Implement agentic workflows to fully automate the onboarding process