5 Lessons from Women in AI on the hello, Human Podcast

For this season of the hello, Human Podcast we celebrated Women in AI with a special series of eight episodes highlighting women across different industries including high tech, healthcare, and government. Throughout the series, the executives shared their stories, lessons learned, and advice for the next generation of women in tech. We rounded up some of our favorite pieces of advice — here are the takeaways:

1. Explore the Unknown 

  • Start with a mentor. It’s hard to figure out where to go and where to begin- especially when many women in tech don’t even start in tech. Sherika Ekpo started her career in human resources while recruiting talent for a government agency. She transitioned from government to tech and now works as the Global Diversity and Inclusion Lead for Artificial Intelligence at Google.
  • “Identify a career field that’s evolving, and identify mentors or sponsors who can help you get into that space.” Sherika didn’t start in tech, so she leaned on strong mentors to help her learn and advance in her career. It’s not about where you start but the resources you find. 
  • In some roles, it’s important to know the technical skills- whether that’s code or understanding the ins and outs of automation. But, it’s also important to keep advancing those skills through different certification courses, webinars, and conferences/events. 
  • Preeti Adhikary encourages young women to “be open to learning, and have that learner’s mindset.” Technical skills can be taught, but emotional intelligence can’t. Being a woman in tech, especially an executive, means knowing how to communicate with your team and the rest of the company. 

2. Battle Bias

3. Address the Gap

  • Katica Roy, the founder of Pipeline Equity, shared the “2021 Retrospective,” which shares the story of gender equity from the past decade in ten trends. “Since 1970, women have added $2 trillion to the US economy through their increased labor force participation. We were set back so far that we’ve lost over $1 trillion out of our economy just due to women leaving the workforce.
  • There are not enough women in tech and not even in the workforce. Employers need to understand that women are needed. There is a statistical gap, but also a need to address it- immediately. Businesses and organizations must continue establishing various inclusion programs for women, POC, and other groups. Creating an equal space and ensuring initiatives encourage the future workforce.
  • Females are powerful. Tripti Sethi was the first data science hire at Avanade. She built a team of 400. Yael was the first female doctor in her town. Growing up in a highly religious community, she would have never imagined being a doctor and now running her own business- Embryonics. Embryonics is developing and applying data-driven solutions to improve the journey and the success rates of fertility treatments. She’s been able to help thousands of women with their IVF journeys. To think, if these women had listened to those around them, their paths could have been very different.

4. Don’t Take No for an Answer

  • Zahra Timsah, the lead for AI Governance at MassMutual, encourages women to “be proactive, not just reactive.” Career is a risk, but life is an even bigger risk. Go for the goal and don’t turn back. If you believe in it, keep pushing for it. If there is no one to champion for you, champion for yourself. And if you can, always support other women. 
  • Katica Roy emphasized the importance of confidence and believing not only in yourself but other women. “Be brave, jump, and the ledge will appear.” It’s bravery that will push you to take that jump and leap of faith. Without taking that risk, you never know what the other side will hold- especially if you listen to others and not yourself first. 

5. Support Other Women (especially other Women in Tech)

You can tune into the Women in AI series here, and if you like what you hear, leave us a review. Ready to share your advice for the next generation of Women in AI? Be sure to tag us on Twitter and LinkedIn using the hashtag #hellohuman. 

The Low-Code Movement Will Spur Process Intelligence

Interest in low-code development is skyrocketing. Annual market growth is predicted to exceed 25%, growing from $13B in 2020 to $65B in 2027, according to research from Brandessence Market Research. For organizations looking to win in the digital-first agile world, low-code is quickly becoming a critical component of a modern enterprise technology stack. If you’ve been around technology, you have probably heard the phrase, “Faster, better, cheaper — pick two.” But low-code gives you all three and adds “flexible” to the mix.

With their roots in the Rapid Application Development (RAD) tools of the 1990s, low-code platforms are an application development environment that use graphical user interfaces and configuration instead of traditional hand-coded computer programming. Formal software engineering skills are not required to create applications since a visual user interface in combination with model-driven logic is used. This opens the door to a wider range of people who can build apps for the business. With a little training, employees can rapidly create and deploy secure scalable software. One note to remember – low-code is not interchangeable with no-code which is a subset of technologies aimed at business users primarily working on enhancing their individual productivity.

A New Player in the Enterprise Technology Stack 

It’s impossible to argue against the need and demand for low-code in the enterprise. According to Gartner, more than 65% of application development in 2024 will be performed by low code platforms. That’s a remarkable shift for a software category that did not exist a decade ago. The number of digital applications and services being built is exploding as well. Between 2018 and 2023, more than 500 million apps will be created according to IDC. To put that massive number into perspective, that’s more than the previous 40 years combined. With low-code, companies can rapidly produce applications within a shorter time span and at a fraction of the cost. Schneider Electric launched 60 apps in 20 months, delivering most in just 10 weeks, and Ricoh replaced critical legacy systems 3X faster with a positive ROI in 7 months using Outsystems. Some skeptics may point out the lack of available IT resources, but business users can learn low-code development methodologies quickly, typically in less than one month. Clearly the old way of building apps cannot keep pace with today’s digital marketplace. 

Process Intelligence Jump Starts Low-Code Journey 

To help accelerate utilization of low-code and scale it across the enterprise, process intelligence is a key enabler. You cannot improve how you operate tomorrow if you don’t fully understand how you work today. And most companies truly don’t understand how they operate on a daily basis, especially at a granular user activity level required to automate a process or streamline a workflow. They have limited process understanding. They don’t know how their applications and data interact, and they don’t really understand what their customers expect.

The impact of this gap in process data is well documented. The 70% failure rate of transformation programs is widely reported. McKinsey pegs the cost at nearly $1 trillion annually and noted on 14% of companies have seen a sustained and material improvement in their business. Another study from Gartner noted only 1% of companies have sufficient understanding of their processes to take full advantage of the technology solutions. 

Before embarking on a major initiative, a company must map its processes, its systems and its experiences. Today, that necessary level of operational intelligence just does not generally exist in most companies and on top of that it is very difficult to obtain without process intelligence.

Low-Code and Process Intelligence – Better Together Than RPA

Process intelligence was a similar catalyst for Robotic Process Automation (RPA), but the opportunity with Low-Code is even greater. RPA programs enjoyed massive early uptake, but the challenge was how to scale the initiatives. Companies still struggle getting more than 50 bots deployed. Once any obvious low-hanging fruit is automated, it becomes difficult to identify what to tackle next and how to tackle it. Process intelligence answers those questions to help scale RPA. 

With Low-Code, Process Intelligence accomplishes that and more. The combination will overtake RPA as the gateway for AI and automation. Low-Code is more efficient and scalable than traditional RPA development because it does not operate at the user interface (UI) layer, making the applications more resilient. Additionally, where RPA is traditionally limited to task activities, Low-Code is much more capable of handling sub-process and process level activities making it much more valuable to the enterprise. Coupling Process Intelligence with Low-Code helps steer an enterprise toward a next-generation operation model that is faster, cheaper, better and flexible. It may also fully realize the promise of Citizen Developers that RPA struggles to achieve.

VentureBeat: Why Process Mining is Seeing Triple-Digit Growth

ISO 9000 is a set of standards for quality management developed in the late 80s. It was based on procurement standards used by the U.S. Department of Defense. But, in a long-ago training session, the entire premise was summed up by the instructor as, “Document what you do, then do what you’ve documented.”

The reasons those types of programs exist is because the documented process is rarely the actual process. But today, given the massive scale and breadth of change over the past year, it’s likely many of your processes have been altered in one way or another. In fact, a survey by McKinsey found that organizations have accelerated their digitization efforts by three to four years during the pandemic. So, before enterprises can change, improve, digitize, or automate a process, it’s imperative they first understand what’s really going on. 

A recent article in VentureBeat, “Why process mining is seeing triple-digit growth”, points to the past year’s massive change, along with many other factors, as what’s driving explosive demand for process mining tools and technologies. Gartner estimates that the market for these solutions has already tripled since 2018, and there’s more to come. Here’s a quick overview of why, according to VentureBeat.

RPA Isn’t Living Up to Expectations

Robotic process automation (RPA) initiatives and their expected ROI are based on the underlying process. Saving an hour per day for a documented process completed by your expensive procurement team could generate a nice return, for example. But, if Procurement is, in reality, already working around the systems and tasks that slowed them down, the return could be considerably lower, or it could even cost you money. Or, maybe that entire process is done differently in other regions or isn’t really valuable for Procurement, so you’ve spent money automating a bad process you’ll eventually need to spend more money to rework.

Companies are now finding that, in their haste to automate, they neither understood the actual process nor developed an optimized, scalable process built for their current needs. As the article states, “Many enterprises are finding it difficult to scale beyond a few software robots or bots because they are automating a bad process that cannot scale.”

Process intelligence, however, is giving enterprises detailed information on why their RPA isn’t driving more value. It’s also giving visibility to the true as-is processes, how they vary across the business, and where optimizations might make those processes better well before RPA is considered. 

Download our ebook: 6 Strategies to Drive Successful Automation

Benefits Beyond Simple Process Mining

Process mining, which “involves mining data logs from applications like ERP and CRM to assemble an accurate model of how a business process, like order to cash (OTC), works,” is inherently limited. Many tools in modern enterprises, such as Microsoft Office, virtual desktops, and email, don’t produce data logs. Process intelligence, or what VentureBeat refers to as “task mining,” uses computer vision, artificial intelligence (AI), and machine learning (ML) to directly record how a worker accomplishes any given task.

See our related post: What is Process Intelligence?

Since process intelligence doesn’t rely on data logs, it can capture more data from more applications, but also how today’s workers jump between applications. For example, maybe the OTC process requires a clerk to copy order data from an ERP system and paste it into an invoicing system, which then generates a PDF invoice that’s manually emailed to a vendor. Process mining would likely capture the ERP and invoicing logs, but process intelligence would provide insights across the end-to-end process and all applications, including the manually created email. 

Those cross-application actions are more indicative of how today’s workers work, and that’s the level of data enterprises are now demanding. Process intelligence adds AI and ML to uncover the nuance of processes and surface opportunities to optimize before you might apply RPA to automate suboptimal subtasks. 

New Applications Across the Enterprise

The inclusion of advanced AI and ML, along with the benefits of the cloud, enables enterprises of any size to quickly capture and evaluate data on enterprise-scale processes. But it’s not just limited to processes, per se, since other insights can be gleaned from knowing how your enterprise operates today and at a detailed level. VentureBeat calls it “a new sensory system for organizations.”

“In the ecology of companies battling for market share, a company with even a primitive capability of seeing invisible workflows better than the competition has a huge advantage over companies that cannot see. As Erasmus stated in 1500 AD: ‘In the land of the blind, the one-eyed man is king.’”

Some examples of these in-demand capabilities of process intelligence include:

  • Understanding and mitigating complex cybersecurity threats
  • Optimizing the physical logistics of warehousing
  • Improving manufacturing processes
  • Breaking down organizational silos
  • Reducing training time
  • Identifying the root cause of quality issues

It’s also being used to enhance how humans work, such as increasing safety and automatically altering workers when steps are missed by mistake. So while the word “automation” may sound like jobs are at risk, one of the most human benefits of process intelligence is in making humans better at how they work. 

“Rather than just looking at process mining as something to impose on workers, companies may see the biggest gains by finding ways to include and reward employees as part of the adoption. After all, thousands of eyes in the field may see some opportunities that a few experts in the office might miss.”

You can read the complete VentureBeat article here.