Culturally Informed Recommendations On Rialto
How We Developed Rialto to Personalize Discovery for Our Users
Data is the new oil.
But just like crude oil, its true power is unlocked only when refined with care and intention.
Recommendation systems drive what we buy, watch, and read. But when it comes to business development, especially for underrepresented founders and business communities, recommendations can’t be generic.
A one-size-fits-all system misses the point.
It ignores people. It ignores context. It ignores real needs.
Our Mission Is Different
At MINWO, we’re not just building another recommendation engine.
We’re building a discovery platform that listens. One that understands where ESOs (Entrepreneur Support Organizations) and ENTs (Entrepreneurs) are starting from and where they want to go.
It’s not about what’s trending.
It’s about what’s relevant.
Why Culturally Informed Recommendations Matter
Most systems today are trained on:
Clicks
Ratings
Purchase history
That works for products. But for people building businesses, it’s not enough.
At MINWO, our system is trained to understand:
Representation: Who is being served?
Relevance: Is this the right resource at the right time?
Resonance: Does this reflect the user’s reality and identity?
We ask deeper questions:
What stage is the business in?
What pain points are they facing right now?
Do they need funding, mentorship, or program support?
What organizations do they draw support from?
These answers shape every recommendation, because being culturally informed means embedding empathy, not just logic, and being personalized not just through behavior, but through lived experiences and identity.
How Our Recommendation System Works
1. Structural Relevance
We don’t match based on surface-level tags.
Our engine uses real data:
Business stage: is your business in growth mode or are you just hiring your first team member? Maybe you’ve already launched and now you’re figuring out how to scale into new markets.
Pain points: are you struggling to secure that first grant or investment? Are you trying to build systems, manage burnout, or find culturally aligned mentorship? Maybe you just need clarity on what step comes next.
Resource-use patterns across similar identities or Neighborhoods: What kind of content are people sharing in their neighborhood? Are they mostly posting about funding opportunities, business templates, or local events? By analyzing the types of resources actually circulating, we gain insight into what truly resonates, what people find useful, relevant, and worth passing along.
This means we prioritize what matters to the growth of your business, not just what’s popular.
2. Integrated Data from Diverse Sources
We pull in data from:
Resource libraries: this helps us know what’s available across ecosystems and can match you with what fits your needs right now.
Feedback and ratings: your experience matters, what you say was helpful or not helps the recommendation engine learn and recommend smarter next time.
Business profile submissions: these help us understand where you are in your journey, what your goals are, and what kind of support would move the needle for you.
Neighborhood and support history: to tailor personalized recommendations.
Each new cycle makes the engine more responsive, intuitive, and tailored.
3. Contextual Intelligence
Our natural language models understand:
Pain points directly from written responses
Growth stages and shifting needs
Cultural context and Neighborhood identity
We use sentence transformers and similarity scoring to recommend resources that align with what users need. These tools help the recommendation engine understand the meaning behind what a user says, not just based on tags or keywords. The engine can surface resources that actually make sense in context. Whether a user's pain point is “I’m stuck with marketing” or “I need help growing visibility” the recommendation engine recognizes the intent and connects them to relevant support resources like “digital strategy toolkits” or “branding workshops”.
Feedback Makes It Smarter
We believe feedback isn’t just data, it’s a dialogue.
We learn from:
What resources are clicked and how they’re rated
What’s marked helpful vs. irrelevant
What new patterns or needs emerge over time
Tools like Amazon Data Firehose help us capture data in real time, so the system gets better automatically. We’re not just waiting for quarterly reports or manual updates, our recommendations evolve as users engage. While most systems rely on static data snapshots, we’re continuously learning from actual behavior, which makes our recommendations more relevant, responsive, and reflective of what users need right now; helping ESOs scale impact seamlessly.
The Road Ahead
Our system is built to honor the journeys of underrepresented founders who face the greatest hurdles as they navigate entrepreneurial ecosystems.
It listens to words, not just data.
It learns and evolves through every interaction.
It respects that support should be timely, personal, and rooted in community.
This is just the beginning.
We’re building a smarter, more empathetic future, one where discovery isn’t dictated by algorithms, but shaped by the people they serve.
Let’s shape the future of business discovery on your own terms.
Until next time, stay curious.
Data Scientist
 | A guest post by
|