Intrare

Across Latin America, over 100 million diverse individuals live not only in poverty, but also face systemic barriers to employment. This includes refugees, migrants, marginalized women, the LGBTIQ+ community, the silver economy, indigenous, and others.

To date, no scalable solution is available for them. 

In building our solution, we were particularly interested in the factors that perpetuate and even exacerbate this problem:

For most of them, job platforms are inaccessible. They require higher digital literacy, data access and sufficiently well-equipped phones that most people do not have. These barriers also make them seem off-putting to people: as our user research has shown, they often believe they could never access or don’t deserve a better-paying, formal job.

Even if they can use such platforms, most algorithms - based on keywords, rankings, or Large Language Models (LLMs) - perpetuate or even increase bias against them. 

Recruiters additionally apply conscious and unconscious bias against them. Diverse groups are more likely to have little experience with formal interviews or lack documents, creating an additional barrier. 

Employers have a strong interest in hiring more diverse but face barriers:

1. Human resources (HR) lacks the know-how and processes for hiring diverse candidates.

2. Non inclusive environments lead to discrimination, dissatisfaction, and high turnover. 

In Spanish-speaking Latin America, companies lack effective solutions for these issues.

As AI in recruitment advances, there's a risk that predictive models insufficiently calibrated for fairness and bias become prevalent. Commonly, LLMs like Open AI are embedded in Chatbot - but they have been proven to be biased, for example against groups with darker skin colors or specific ethnicities. In other cases, AI models employ deep learning, but learn from biased data and lack adequate training and adjustment. 

This is likely to further increase economic inequality and poverty. 

Addressing these challenges can help millions gain economic stability and resilience.

To make economic empowerment attractive for employers, we offer them an AI SaaS tool that eliminates bias and includes marginalized talent in their hiring, while saving them significant time and cost. This allows us to give marginalized individuals access to high-quality opportunities through our platform Talento Intrare, empowering them to build economic resilience. The engine behind this are proprietary AI models that we build for unbiased, fair, and optimized hiring.

1. The engine: our AI.

We build AI models for unbiased, fair, and optimized hiring, using a novel approach that avoids common bias-reinforcing patterns. 

We use machine- and soon deep learning with high-quality, unbiased data gathered by our team. We implement human feedback loops and mathematical fairness assurance techniques. Our models allow us to explain predictions clearly to companies and candidates.

Our Chief Data Scientist is an AI fairness PhD at Worcester Polytechnic Institute (WPI) working with experts at MIT, Harvard, and Stanford.

2. Bringing diverse, marginalized groups in poverty into fair jobs. 

We reach out to users via TikTok, Meta, and partners, inviting them to join our platform through a simple sign up process they can complete on their phone. This experience requires minimal digital literacy and data access, making it accessible to all.  

If a user has to abandon their registration mid-way, we seamlessly transition communication to WhatsApp and guide users to completing it. Once onboarded, users receive support, TikTok-style employability training, and tailored job recommendations curated by our AI. Once they select the opportunities they are interested in, their profiles are transmitted to prospective employers. Throughout the hiring process, we provide reminders and support, ensuring candidates can attend interviews and effectively engage with employers.

3. Our solution for companies is designed to deliver maximum business value to make it easy to decide using it. 

We built a simple yet powerful SaaS tool. Companies can upload job openings and process them using our AI. Our AI matches candidates from Intrare’s talent pool and additional ones from the companies’ databases with the openings, providing precise and unbiased matches. It shows match scores and explanations of each candidate’s fit with a job. 

Companies can then select preferred candidates and manage the entire hiring process. 

Our AI and user journey is so efficient that is can save companies an average of 1,000 USD per vacancy. 

Our solution can be used directly or integrated with existing Applicant Tracking Systems. 

By providing such a high-value solution, we foster interest among companies in using our solution and facilitate fair job opportunities for marginalized individuals. 

4. Supporting companies in creating an inclusive environment.

The economic empowerment of marginalized groups not only requires creating access to the right opportunities, but also fostering welcoming work environments where they can grow.

To support this, we offer webinars and training for HR departments and leaders. This covers two crucial aspects: developing inclusive practices for diverse hiring, and cultivating inclusive work environments.  

Demo for companies: https://vimeo.com/937981913?share=copy  

1. Our target population 

We target individuals in poverty and informality, especially diverse groups - refugees, migrants, marginalized women, LGBTIQ+, silver economy, indigenous, and others. Across Latin America, we estimate that over 100 million individuals in diverse groups are economically active but in poverty. 

2. Why this population is underserved

There are thousands of recruitment solutions, but they either create barriers for this population, or are not able to serve it at scale. 

Most platforms (OCC, Indeed) require much higher digital literacy and data access than most have. Even if users apply to a job on such a platform, their algorithms tend to be based on rankings, keywords, or LLMs and produce bias against diverse and marginalized candidates. These pages then send large volumes of candidates to companies, where human recruiters filter them, applying own biases. 

There are several solutions for specific marginalized groups (refugees, LGBTIQ+ talent) in the region. But they cannot offer scalable solutions to employers, and don't integrate with companies' current systems. 

This leaves a significant gap: companies have no inclusive solutions they can use across their whole hiring process, and marginalized populations lack solutions that can allow millions of them build economic resilience. 

3. How we address their needs 

a. Accessibility 

Our solution for users is built for accessibility. They can register in a few clicks, with very little data access. Our user experience is built with users, making sure it is usable even for individuals who are very unaccustomed to digital experiences. After registering, all interaction happens via WhatsApp, which is the most used app in Latin America. This makes it easy for users to receive training, get support, interact with job recommendations we provide them, and follow our support through hiring processes.

b. Attractiveness for companies 

Companies in the region want solutions that offer them business value first, impact second. When we piloted our first ideas, we got a lot of interest from companies. But they only wanted to hire a few diverse individuals at a time and we struggled convincing them to transform their hiring and workplace practices incorporating fairness, equity, inclusion, and diversity.

Our SaaS tool offers the perfect mix. The efficiency and savings it offers are very attractive. The bias-free hiring and access to our diverse talent pool are the cherry on top. 

We believe that will allow us to incentivize companies to become inclusive and empower marginalized individuals in poverty.

c. AI that eliminates bias  

In filtering and pre-selecting candidates, recruiters narrow down from 100-200 candidates to just 3-7. Innumerable conscious and unconscious biases influence it, leaving marginalized individuals out. 

That is why we decided to build an AI to make this pre-selection without bias and fostering fairness, while offering humans clear information on its suggestions. 

Our AI ensures all diverse groups have comparable chances. And it brings results: with our AI in beta, over 37% of all matches we make are invited to interviews, and over 11% of all matches are hired. These metrics are already above industry standards.

We believe that building with and for the communities we serve requires three key ingredients: 

1. Hiring people that understand and connect with our users through their own lived experience. 

2. Building with the communities. 

3. Being experts in building for marginalized communities. 

We are ourselves a diverse team, from a single mom to LGBTIQ+ community members and a refugee. The two co-founders, Hannah and Marcela, are women. Hannah’s trajectory has been entirely focused on economic development and inclusion, and she has worked with diverse communities from Germany to Ghana and Mexico. Since founding Intrare, she’s been at the forefront of building the team and our practices around working with the communities we serve in building solutions for them. Marcela, in charge of building Intrare’s AI, ensures this focus by incorporating user and team feedback in every step, and through working with renowned AI fairness experts from MIT, Harvard and Stanford. 

Our way of working focuses on building together with the communities we serve. That has been the case since the first pilot project that a few of us conducted in our free time in 2018. 

When we started building Intrare as a digital solution in 2023, we made designing with communities part of our process. That includes:

- Consultations with previous users 

- In-depth user research, interviews and conversations, including all teams (from user acquisition to engineering)

- Focus groups to pilot new solutions and adjustments 

- User behavior analysis 

Finally, we focus building for our communities. That sounds easier than it is: few startups are truly experts in inclusive products and user experience (UX). We built this expertise by: 

- Building internal product processes focused on building with users.

- Ensuring everyone on the team, especially engineers and operations, are continuously in touch with users and analyze their behavior to better understand barriers and needs. 

- Exchanging with product managers and CEOs of other successful digital solutions for marginalized groups. 

- Working with metrics that reflect whether communities like our product, especially user engagement in key steps.

May 2018 to May 2023: We conducted pilots and provided manual matching and support to users and companies. 

2022: We started building our AI. 

2023: Developed WhatsApp-based service platform for users, launched with the AI’s first beta in May 2023. Since then, we have made numerous changes and improvements to optimize it for our communities. Developed the first version of our platform for companies, launched in September 2023. 

2024: In February, based on the very positive results of our AI, we decided to change our platform for companies from talent marketplace to SaaS (as described above) to provide a much more attractive offer. Internal launch and testing of SaaS during April. 

Results to date: 

- Over 5,000 diverse, marginalized users

- Over 5,000 users received support 

- Over 1,500 have taken training

- Over 3,000 matches with jobs made. All users that hold necessary documents have received at least 1 job match  

- 373 individuals hired 

- 297 hired. An estimated 70% were in poverty before and broke out with Intrare job

- Hired individuals earn 1.7 times more than average workers in zone

- 47 active employers, over 800 staff trained in inclusive practices

We believe that Solve is precisely the solution and network we need to grow at our current stage. 

It is especially important to us to connect with legal  and software resources that can help us scale our AI and tech from Mexico to the rest of Latin America. We also believe this support can help us achieve partnerships with other crucial actors in the system. 

Secondly, we are very interested in connecting with other leaders in AI and tech for good, to offer each other support and inspiration. In the current environment, in which the large majority of solutions in the field are at risk of increasing inequality and poverty, we believe that those that push for AI for good have to join forces to make a change. 

Finally, we believe that Solve can provide us crucial advice in taking the next step in building an even better team. While we now have a strong team complementing each other in key areas, we want to strengthen both our product and engineering to get ready for scale.

There are thousands of recruitment solutions, but all exclude marginalized communities involuntarily. 

There are also hundreds of NGOs and impact solutions, but they lack the scalability and are not sufficiently attractive for companies to use them widely. 

We built our solution to change how companies give job opportunities to people: 

Our AI allows us to eliminate bias in the most inefficient and biased hiring steps, filtering and pre selection. 

It also offers high efficiency for companies: we can reduce vacancy periods in industry from 15 to 10 days, saving companies over 1,000 USD per opening. 

This puts us on the map for finance, supply chain, and operations, who need to optimize their metrics; as well as for HR, who need better solutions. In addition, we integrate with the key solutions companies already use (Applicant Tracking Systems). 

That makes our solution a no brainer for companies, combining savings, efficiency, and Diversity, Equity and Inclusion. 

Reaching marginalized users at scale also requires innovation. These groups tend to have low digitization and digital literacy. That is why we built a user experience for very low digital literacy and data access, making it as frictionless and easy as possible to sign up, receive support, take simple employability training, and receive job opportunities. Once they are invited to an interview we follow up with reminders and support, considering most individuals do not have organization tools. This ensures a very high user engagement and success rate from sign up to interviewing and hiring. It also allows users to feel empowered in their experience and gain confidence in interacting with these types of solutions.

The innovation described in the previous question allows us to unlock companies’ interest, train them in more inclusive practices, and unlock a large number of job openings. Companies can offer our online micro learnings on DEI practices to their leaders and staff, ensuring a continuous improvement of conditions in the company. 

This, in turn, enables us to match marginalized individuals with fair job opportunities where they can improve their income, and gain stability and economic resilience.

We estimate that over the next 5 years, we will reach 1 million marginalized individuals and that over 6,000 companies will join in. 

This also creates a systemic impact. We influence how companies hire, and even if they stop hiring with us, practices and policies persist. Within the private sector, leaders in these companies influence and push each other to improve practices, creating a new standard. We also start observing that other recruitment providers are interested in seeing that our cutting cutting-edge, bias-eliminating AI works, and seek solution to improve their own practices and offers. These dynamics open the job market for marginalized groups, much beyond our direct services.

5-year goals:

- Offer 1M individuals access to support, employability training, and fair employment opportunities 

- At least 50% of users are in poverty when signing up, and at least 50% are not previously in formal employment 

- All users that are hired through Intrare are afterwards above the poverty threshold and in formal employment 

- Our matches offer salary equity to all different diverse and marginalized groups according to fairness standards 

- Over 6,000 companies engaged, with approximately 3,000 internal champions 

We measure additional indicators to understand the impact on individuals' perception of their economic stability, social and economic integration, and of companies’ perceptions. These help us guide our work into the right direction.

We build proprietary AI models, using machine learning based on high-quality, de-biased data. They are built with human feedback loops and mathematical measures allowing us to optimize for fairness. 

Predictions of job quality matches are clearly explained to users and companies. 

Co-founder and Chief Science Officer Marcela Vasconcellos at Worcester Polytechnic Institute is in charge of the models, working with experts at her university as well as MIT, Harvard, and Stanford. 

Our digital solution is built by our engineering team, headed by our CTO experienced in HR tech. 

Our technology and AI models are constantly tested and results are measured by the AI team and advisors outlined above, the engineering team, and the team in charge of attending users and companies. In addition, an internal task force continuously ensures optimizing our user experience for digital literacy and accessibility.

Our tech and AI, launched in 2023, have been proven to work with over 4,000 diverse users and the job openings of over 16 companies.

12 full time

1 part time

Start of piloting: 6 years ago 

First full-time employees, working on a previous solution: 4 years ago 

Start of AI development: 2.5 years ago 

Start of tech development (digital solution for companies and users): 2023

Launch of AI embedded in digital solution for users and companies: 2023 

Our solution is and will be free for diverse, marginalized users. We are starting to charge companies since April 2024. The product we offer is described above. Below, we describe how we are charging companies, and who we target. 

Our model is a Software as a Service (SaaS) subscription model with 4 tiers according to the number of hires a company makes per month: small, medium, big, and enterprise.

1. Our market (companies): 

Our initial market are mid-market companies in Mexico that are in urgent need of reducing inefficiencies and cost in recruitment, and have faster sales cycles. As explained below, we focus on direct sales via LinkedIn, mailing and industry events, targeting operations, finance and HR VPs and directors most desperate for the efficiency and savings our product offers.

Market size

The initial potential clients in Mexico make up a 88M market. Once we have established stable sales and proven our product, we will move on to big and multinational customers. Those tend to buy products regionally, allowing us to expand throughout Latam and the U.S. and grow tickets significantly. 

Our serviceable market is the HR tech market (59B) and the capturable segment of optimizable recruitment spending (+15B) which is nearly totally untapped right now, in total a 74B market in the Americas (U.S. and Latam).

Our total market is the recruitment market (277B) plus the total optimizable recruitment staff spending (75B) in the Americas.

Reaching first paid clients

We are conducting outbound lead generation, targeting customers that we have identified as potentially interested. Focusing on industries and segments outlined above is key, as those are in most urgent need of efficiency and savings.

We target leads through LinkedIn, mailing, and industry events where we position our brand and offer. Direct sales allow us to access a faster sales cycle while receiving continuous feedback from prospects.  

To feed the midterm funnel, we conduct inbound through webinars with partners such as Betterfly, and LinkedIn content. 

Traction 

We have a +100 pipeline of companies that we have been in touch with and that expressed interest. We have conducted discovery calls with over 30, and 13 have requested a quotation. 2 already bought our product. 

With our AI in beta mode, 38% of the matches we make are invited to interviews and 11% hired. That shows the amazing quality of our results, well above industry standards, even though this is still an initial version of AI. Our tests show that we can increase these numbers much more drastically.