E17: Everyday AI, Empowerment and Inclusive Design with Mave Social

From Corporate Life to Making AI Accessible

Sumathi's route into AI has been shaped by a long career in financial services and technology. Having previously worked as a managing director and principal at Bear Stearns, she later took a career break to look after her children.

That experience helped shape her understanding of what happens when people step outside the corporate environment. Technology continues to move rapidly, and those who are no longer immersed in workplace technology can easily find themselves falling behind.

This is why Sumathi uses the term “curious non-corporates” to describe a group she particularly wants to support: mothers, founders, freelancers, trustees and others who are not necessarily going into an office every day and therefore may not receive the same AI training as employees in larger organisations.

Her concern is that without accessible opportunities to learn, these people risk being “written out of the AI-fuelled future.”

Curiosity, Confidence and Getting Hands-On

For Sumathi, AI literacy is not simply about understanding what AI is — it is about actually using it. She argues that people often approach AI as if it were binary: either an extraordinary opportunity or something dangerous that should be avoided. In reality, she describes AI as existing in “layers of greys in between.”

The best way to understand those possibilities and limitations is through experimentation. Her workshops therefore focus heavily on hands-on learning, allowing participants to use AI, build with it and discover what it can and cannot do. Human expertise remains central to this process. People who understand their own field are often better equipped to use AI because they know what a good answer looks like. That knowledge allows them to assess AI outputs critically rather than simply accepting them.

Why SMEs Should Start Now

Sumathi believes small and medium-sized businesses are entering a particularly important moment for AI adoption. Two years ago, she argues, many SMEs faced a challenging environment in which AI platforms were expensive and cumbersome. Today, the market is far more accessible, with hundreds of tools available for businesses of different sizes. AI is therefore no longer something reserved for large corporations with dedicated technology departments.

For smaller businesses, Sumathi's advice is straightforward: start small, learn and experiment. Rather than trying to transform an entire organisation at once, businesses can identify one small part of their workflow that could be improved, introduce a tool, evaluate the results and then decide whether it is worth scaling.

Creating Permission to Experiment

One of the most important steps for organisations is establishing an AI policy. Employees may already be experimenting with AI without their employers knowing. A clear policy can therefore provide boundaries around what employees are allowed to use while giving them the confidence to explore the technology safely.

For Sumathi, this is about creating a framework in which experimentation is encouraged rather than punished. An AI readiness assessment can also help businesses identify which parts of their workflow are suitable for automation. Importantly, not every experiment needs to succeed. If a particular process does not work with AI, organisations can learn why and move on.

The underlying data may not be suitable, the process may not yet be ready for automation, or the technology itself may simply not be appropriate. The objective is to learn without making every experiment a major investment.

AI in Everyday Business

Sumathi highlights several practical examples of how AI can already save time and improve productivity. For one workshop, she took a simple spreadsheet containing attendees' names, email addresses and industries and automated research into their professional backgrounds. This allowed her to understand the audience's AI expertise, business size and industries before delivering the workshop, helping her tailor the session more effectively.

She also describes a business providing online lessons that uses AI to transform teaching materials into concise bullet points and generate multiple-choice questions for students. The process saves around an hour and twenty minutes each day — roughly ten hours a week.

Other examples are even more everyday: using ChatGPT's voice mode to draft emails while commuting, or using AI to quiz children on school subjects. These examples illustrate Sumathi's broader point: AI does not always need to involve a dramatic transformation. Small efficiencies can collectively give people significant amounts of time back.

Why Gender Equity in AI Matters

Sumathi's work with Women Defining AI is driven by a concern that the consequences of underrepresentation in technology are already well understood. She points to areas such as clinical trials and car safety, where historically insufficient consideration of women's bodies has contributed to systems that do not work equally well for everyone.

AI presents a similar challenge, but potentially on a much larger scale because it is increasingly influencing healthcare, transportation, recruitment and many other areas of everyday life. If women are not represented among the people building and making decisions about AI, important forms of bias can remain invisible.

Sumathi illustrates this through an experiment involving a CV with a name that does not clearly indicate gender. When she asked an AI system to assess the career prospects of the same person as both “her” and “him”, she received dramatically different recommendations. For her, this demonstrates why diverse perspectives are not simply desirable within AI — they are necessary for identifying and correcting bias.

Moving Women from Users to Builders

Women Defining AI aims to increase women's participation in AI by moving them from being users of AI to builders of AI. Sumathi argues that part of the challenge comes from how AI is presented. Framing AI purely as a STEM or technology subject can make it appear inaccessible to people who do not identify with those fields.

Even terminology can influence participation. Sumathi questions the phrase “prompt engineering”, arguing that the word “engineering” can unnecessarily make the activity sound more technical than it is. Changing the language around AI could therefore help broaden participation and encourage more people to see themselves as capable of contributing.

AI Adoption Is Not Black and White

Sumathi identifies one of the biggest risks in AI adoption as the tendency to treat AI outputs as definitive.The first answer an AI system produces is not necessarily the right answer. Users need to understand how to give AI sufficient context and direction, and then critically evaluate the result.

She compares this to searching for a road on Google Maps. Simply entering “London Road” is unlikely to identify the right destination because there are numerous roads with the same name. AI similarly needs context and intentionality from its user. This makes AI literacy particularly important. Understanding how to question, refine and evaluate AI outputs is just as important as knowing which tool to use.

Governance, Privacy and the Need for Consequences

Alongside literacy, Sumathi identifies data, privacy, deepfakes and governance as major risks. She believes policy needs to catch up with the pace of AI development, while also ensuring that a wide range of voices are represented in the policymaking process.

Importantly, she argues that AI governance should have meaningful consequences. Organisations should not only take privacy and responsible AI seriously once something goes wrong or a fine is imposed. The scale of AI's potential impact means that governance needs to become an integral part of adoption rather than an afterthought.

The Everyday Efficiencies That Excite Her Most

Despite the risks, Sumathi is particularly excited by the mundane tasks AI can now make dramatically faster.

She remembers the amount of time spent manually creating diagrams and aligning presentations in PowerPoint. Tools such as Gamma can now turn written material into presentations within seconds. For Sumathi, this represents a fundamental change in how people work. AI means that users no longer necessarily need to begin with a blank page. They can create a first draft and then apply their own judgement and expertise to improve it.

But that human contribution remains essential; AI-generated material still needs to be checked, refined and judged against a standard of quality. Without human expertise, there is a risk of AI-generated content simply being recycled back into future AI systems.

Building a More Inclusive AI Future

Sumathi's approach to AI is ultimately grounded in accessibility. Whether she is helping an SME automate a small part of its workflow, teaching someone outside the corporate world how to use generative AI, or encouraging more women to become AI builders, the focus is the same: people need the confidence and knowledge to participate in the technology that is reshaping their lives.

For organisations, that means creating space for experimentation, establishing clear policies and starting with manageable use cases. For individuals, it means remaining curious, getting hands-on and learning how to evaluate AI critically. And for the wider AI ecosystem, it means ensuring that the people building these systems reflect the diversity of the people who will ultimately use them.

As Sumathi puts it, the opportunity is already here. The challenge is making sure everyone has the opportunity to take part.