The challenge of fully representing—let alone empowering—women in technology has taken much longer. According to 2019 data from the National Center for Women & Information Technology (NCWIT), though 57% of all employed adults in the U.S. were women, they held just 26% of computing roles.

The technology industry clearly has far more work to do to attract women to STEM—science, technology, engineering, and math—careers. It also needs to do a better job of helping them succeed once they get there. At AWS, we believe the future of tech must include every color, gender, belief, origin, and community. We’ve been purposeful in working toward the creation of a business that reflects the diversity of its customers. Some of the ways we do that include recruiting from women’s colleges and organizations, maintaining women-focused affinity groups with global chapters to help strengthen a sense of community, and offering benefits like parental leave, mentorship, and sponsorship programs.

It’s also a personal passion. I want to create avenues for women leaders to grow in part because of how my own career unfolded. I don’t come from a traditional technology background; I started as a speech-language pathologist who came to see the potential for technology to improve health care and digitally transform an industry. I had great mentors and sponsors along the way who saw things in me that I didn’t always see in myself.

Here are some of the lessons I’ve learned along the way as we at AWS have worked to build a pipeline of female talent in technology.

Better connect education to jobs

The future of learning is career-aligned. People must be taught skills that are directly linked to in-demand jobs in a “stackable” manner that allow them to pick up new competencies and build a lasting career at the same time.

Partnerships are critical to achieving this. They bring together a diversity of views, capabilities, and resources to solve big challenges—challenges that partners could not solve on their own. By working with the private sector, government leaders can work backwards from the needs of employers to build new educational models that closely align curricula to the skills that firms identify as the most critical.

For example, AWS last year partnered with Louisiana Gov. John Bel Edwards, economic development agency GNO Inc., and the Louisiana Community and Technical College System (LCTCS) to create new opportunities in cloud computing across the state. LCTCS’s campuses are highly diverse: Black, Indigenous, and people of color (BIPOC) represent over 63% of the student body, and women represent over 50%. As part of the collaboration, each of LCTCS’s 12 campuses created an Associate of Applied Science in Cloud Computing degree, offering a skills-based program that will best position students for well-paying, in-demand jobs.

We need more public-private partnerships like this—among educators, private industry, and policymakers—that can accelerate the availability of opportunities for a workforce hungry for them. Such efforts can empower underrepresented people to obtain stackable credentials, give students more ownership of their education, and encourage lifelong learning.

Create structures that can scale

Change can’t be on just one person’s shoulders, and achieved with one-off ideas. Senior leaders should turn their experience and insights into programs and mechanisms that can scale.

We at AWS have tried to walk the talk. In 2015, we launched AWS Educate, a global initiative that grants students and educators access to our technology in order to accelerate cloud-related learning and help train a new IT workforce. In just five years, the program has grown from a collection of online courses to programs developed with established universities around the world. It’s now used in more than 200 countries and territories.

How did we get AWS Educate to scale so quickly? First, we didn’t try to stand up an entirely new educational program on our own, which almost certainly would be a long, arduous process that may not meet the differing needs of academic institutions around the world. We instead focused on helping those institutions adapt or modernize their existing certificate and degree programs, or create new ones by embedding cloud concepts and learning experiences. This approach allowed the institutions to move quickly, often standing up workforce development courses in a matter of weeks.

Second, we looked beyond the traditional degree. We knew we needed to offer modular offerings for working professionals and people looking to enter or reenter the workforce through community or technical colleges. For example, we collaborated with Pearson to develop the Business and Technology Education Council (BTEC) Higher National qualifications at levels 4 and 5 in cloud computing. The BTEC Higher Nationals are internationally recognized higher education qualifications and are delivered at colleges and universities in 50 countries around the world.

Finally, we didn’t go at it alone. Educators and corporations can work together to build academic and workforce programs, but the programs alone won’t get students jobs at scale—paving that path to a career requires engagement with employers, governments, and economic development organizations, too.

Monitor the data

When setting out to solve any problem, I always start by asking: What does the data tell us? The answer, at least as it pertains to the tech industry’s workforce: We still have a lot more work to do to engage women and other underrepresented communities. For the technology industry to match the diversity of the U.S., our companies would need workforces that are 50% women, 18.5% Hispanic or Latinx, 13.4% Black, and 1.3% Native American. The reality: The collective tech workforce is 36% female, 7% Black, and 8% Latinx, according to the U.S. Equal Employment Opportunity Commission.

To begin solving this problem, monitor the data within your own organization. The numbers tell the story—as long as you ask the right questions. At AWS, I’m lucky to have great human resources partners who help me understand the numbers within our team. We regularly audit workforce data to ensure that our employees are bringing a variety of backgrounds, ideas, and points of view to the table. We track the representation of women and underrepresented ethnicities because we know that diversity helps us build innovative teams that obsess over our customers’ needs. We have made year-over-year progress, but we still have work to do to achieve better representation across our various businesses.

Support mentorship and sponsorship

As a leader who has benefited so much from mentors and sponsors over my career, I am committed to following that example—and I owe it to everyone around me to encourage that throughout our business. If you’re a woman leader, make sure you’re pulling other women along and extending that ladder. Give mentees strong sponsorship, strong feedback, and roles and responsibilities that challenge them so they can demonstrate their ability to do more. Both mentorship and sponsorship are important. A mentor can help with seeing yourself clearly; a sponsor can support your career and aid in your promotion.

It is crucial that we continue building a pipeline of strong female leaders—especially right now. Women have been disproportionately affected by the coronavirus pandemic, and the percentage of women participating in the labor force dipped below 55% in April 2020 for the first time in 34 years. At Amazon, we’ve committed to providing employees as much flexibility as possible through this particularly difficult period, in addition to benefits like personal leave of absence, flexible and reduced time options, child and adult backup care, parent discounts for tutoring and child care, and more. It is clear that we must be proactive in reaching out to our female employees to make sure they have the support they need, and we must leverage technology to create opportunities in this new era of remote work. Waiting for them to ask is waiting too long.

Teresa Carlson is vice president of worldwide public sector and industries at Amazon Web Services.