An Interview With Chad Silverstein
The ROI of AI isn’t just efficiency. It’s what efficiency enables.
As a part of this series, we had the pleasure to interview Vera Modenova.
Vera Modenova is Chief Operating Officer at Udora, a fast-growing gifting platform connecting customers with local florists, confectioners, and artisan makers across 50+ countries and 1,500+ cities, including MENA, UK, Spain and LATAM. She joined the company at 21 and became COO within four years. Over a decade in operations, she built the customer support function from scratch to a 92% satisfaction rate, scaled the platform across Europe and the Middle East, and in 2025 led the business to 129% GMV growth. She speaks on AI, operational scalability, and cross-border e-commerce at conferences across Europe and the Middle East.
Thank you so much for joining us in this interview series! Before we dive in, our readers would love to “get to know you” a bit better. Can you share with us the backstory about what brought you to your specific career path in AI?
It’s a pleasure, thank you for having me! I started my journey when I was 21, handing out leaflets at an exhibition. That was my day one. Within four years, I’d stepped into the COO role. I think that trajectory says a lot about how we operate at Udora — if you show up with energy and a clear head, you grow. I never planned to lead operations for a gifting marketplace spanning 50+ countries. I just focused on solving the next problem in front of me.
Today, Udora connects customers with local florists, confectioners, and artisan makers in 1,500+ cities across MENA, Europe, and Latin America. What I really love about this business is the balance: we’re a technology platform, but we serve deeply human moments — birthdays, grief, celebrations, love.
Our AI chapter took off around 2021, just as ChatGPT was starting to make waves. I became an early adopter not because it was trendy, but because I’m constitutionally unable to ignore a tool that might solve a problem faster. I started testing it for internal memos, analysis, market research, and personal decision-making.
What I found was that the productivity shift was real — but only if you were clear about the problem you were actually trying to solve. With time, we came to a point when AI is embedded across Udora at every layer: marketing, design, localization, customer support, software engineering, legal. It’s not a project; it’s the operating system. Our engineering team has already automated 22,000 hours of manual testing. Our support AI, Anna, handles 46% of first-contact messages without human involvement. Our design team generates visuals four times faster than they used to. This is the result you get when you move past the ‘experimenting’ phase and actually start building.
Can you share the most interesting story that happened to you since you started working with artificial intelligence?
The one I keep coming back to is consumer experience and customer support in particular.
My team and I developed “Anna”, our AI support assistant. She was built to handle the high-volume, repetitive bits of customer service like tracking order statuses, delivery questions, or basic seller onboarding. The data is impressive: Anna handles 46% of all first-contact messages, resolves 41% of tickets entirely on her own, and has effectively replaced the equivalent of 3.5 full-time staff in routine workload. Ticket resolution time dropped 10.5%.
What actually surprised me most, though, was something that cropped up while we were testing for localization. Anna isn’t just multilingual; she operates across all our global markets. She’s navigating not just linguistic differences but cultural nuances in how people ask for help, how they express frustration, what they consider acceptable in a service interaction. We found that the same request, translated word-for-word, would land completely differently in the UAE versus Spain versus Brazil.
That pushed us to think about Anna not as a chatbot, but as a cultural intelligence layer. She now works with both our customers and our sellers, helping local artisans in newly launched cities to navigate our platform in a way that feels personal and local. What struck me most was that the hardest part wasn’t the technology. It was figuring out what “good” looks like across 50 different markets.
You are a successful leader in the AI space. Which three character traits do you think were most instrumental to your success? Can you please share a story or example for each?
I’d say it comes to a few things that might seem simple on the surface but can actually be pretty tough to master.
First, it’s all about prompting. I think it’s the defining skill of our era. AI is only as effective as the task you give it — a vague input produces a vague output, every time. What separates an effective prompt from a basic one is specificity, depth of context, and clarity of expected result. Most people only give AI one of these three. When you hit all three, the quality of what you get back is on a completely different level.
Second, data-driven thinking but never data alone. Once again: your AI is only as smart as the information it’s fed. If your data is messy, incomplete, or siloed, your AI is just going to repeat those mistakes. Investing in data quality, structure, and accessibility isn’t a technical nicety; it’s a prerequisite. Before asking “What can AI do for us?”, you have to ask “Is our data actually good enough? Is it something a model can meaningfully learn from?” Skipping that step is why so many companies don’t see their AI projects grow as expected.
Finally, staying on the wave. Just keep learning. I saw some research recently from ADP that only 3.8% of employees learn new skills on their jobs within two years. In a field where model capabilities evolve every few months, coasting means falling behind fast. I track major updates, test new tools before my team does, and I’m still the person most likely to have spent a weekend breaking a new feature just to understand how it actually works. That curiosity isn’t a personality quirk — it’s a professional necessity.
Can you share a specific example of how you or your organisation used AI to solve a major business challenge? What was the problem, and how did AI help address it?
I’ve got two examples that really show both ends of the spectrum.
The first is our engineering workflow. We had a problem that every scaling company faces: test case creation was slow, inconsistent, and stylistically dependent on whoever happened to be writing them. To fix that, we integrated AI-generated test cases through Qase AIDEN. Now, the AI drafts the cases based on our feature descriptions, and our engineers just do a review. As a result, test creation became 30–40% faster and cases became more structured. In 2025 alone, those automated tests prevented 67 risky releases. That’s 22,000 hours of manual work effectively automated. For a team focused on international expansion, that kind of leverage is a game-changer.
The second is design. Our designers now use AI to generate visual concepts four times faster than before. We trained a neural network on our brand identity — so it understands our visual language, our color approach, our aesthetic — and built a clear prompt guide so the whole team can use it consistently.
I’m a big believer in delegating repetitive, high-volume tasks to machines, but I’m just as picky about what we chose not to automate. We don’t use it for recruiting — the volume doesn’t justify the system. And we made a conscious decision to stay manual with analytics, because the cost of maintaining an AI for error detection outweighs the benefit at our current scale. Knowing what not to automate is as important as knowing what to.
What are some of the common misconceptions you’ve encountered about using AI in business? How do you address those misconceptions?
The biggest one I encounter is the idea that AI is a replacement strategy. Leaders see the cost savings, see the efficiency numbers, and start planning headcount reductions. Klarna is a common example here — they replaced 700 agents with a single AI assistant. By 2025, service quality had dipped enough so that they started rehiring. The lesson isn’t that AI failed; it’s that they optimized for cost over empathy, and customers noticed it immediately.
The second misconception is that AI is a project with a start and end date. You don’t “implement AI” and tick the box. The technology moves, customer expectations shift, and your use cases need to evolve with them.
In your opinion, what is the most significant way AI can make a positive impact on businesses today?
Honestly, I think the biggest win is just removing cognitive load from decisions that don’t require human deep check.
Businesses waste enormous amounts of time and talent on tasks that are pattern-matching exercises: routing support tickets, generating first drafts, flagging anomalies. These are tasks where consistency and speed matter far more than creativity or empathy. When you let AI handle them, you free people to do work that actually requires them.
At Udora, we ask one question for every process: ‘Is a human adding irreplaceable value here, or are they just burning energy on something a well-trained model can handle as well?’ The answers are often eye-opening. The compounding effect of this over time is significant — you don’t just save hours, you shift the entire character of the work your team does. A team doing more meaningful work produces a better product. The ROI of AI isn’t just efficiency. It’s what efficiency enables.
Based on your experience and research, can you please share “5 Ways AI Can Solve Complex Business Problems”?
I’d love to! These points are grounded in what we’ve done at Udora and what the data is telling us:
1. Automate the high-frequency, low-risk tasks first. The decision framework I use comes from a simple matrix: if a task is done frequently and is easy to automate — do it now (translations, visual creatives, support routing). If automation and maintenance costs don’t justify it — keep it human. If errors carry serious consequences — automate later, with oversight (financial reports, legal documents). If it’s done rarely and takes little time — ignore it entirely.
2. Make AI the default for every employee, not just technical teams. At Udora, every employee has at least one company-paid AI subscription. The choice is role-specific: engineers use Claude and GitHub Copilot; designers use Midjourney and Gemini; support teams use ChatGPT; legal uses NotebookLM. When AI is democratized across the organisation rather than siloed in one team, the compounding productivity effect is measurable.
3. Optimize your content for AI assistants, not just search engines. According to Adobe’s data, shoppers arriving from generative AI assistants spend 45% more time on site, view 13% more pages per visit, and convert significantly better than non-AI traffic. In the UK alone, retail traffic from GenAI tools grew 1,200% in under a year. This is the new SEO, and to win here, you need structured data and clear FAQs that an AI can understand and trust to further recommend your brand.
4. Build your data infrastructure before you scale your AI. You have to get your data quality and accessibility right before you write the big check for AI. Good data infrastructure makes every decision better, whether there’s a model involved or not.
5. Know where not to automate — and protect those boundaries actively. As mentioned before, we don’t automate analytics because of the cost of possible mistakes. What else? We don’t use AI for creative copywriting at Udora because it loses the human soul that matters a lot in such a warm category as gifting. The companies that extract the most from AI are usually the ones with the clearest list of what they’ve decided not to touch.
How can smaller businesses or startups, with limited budgets, begin to integrate AI into their operations effectively?
Honestly, my best advice is to start with a triage framework rather than the tech itself.
Before you ask “which AI tool should we use?”, ask “what is our most expensive recurring problem?” The answer to that question should drive every AI decision you make. Most early-stage businesses get this backwards — they find an interesting tool and look for a problem to apply it to. That’s how you end up with AI washing instead of actual results.
Whenever I speak on identifying high-impact opportunities, I share a simple prioritization checklist: How frequently is this task done? How complex is it, and can AI handle it today? What are the manual costs versus the automation and maintenance costs? What is the cost of an error? And can the task be broken down so that only part of it is automated? Running every candidate process through those five questions will save you from expensive mistakes.
For most startups, the highest-impact and most accessible starting points are: AI-assisted content and communications (first drafts, translations, product descriptions), basic support automation for frequently asked questions, and using existing frontier models — ChatGPT, Claude, Gemini — before building anything proprietary. One critical caveat: always review AI outputs before they reach customers. The biggest pitfall I see is deploying AI-generated content without a human review step. The automation saves time, but exactly the review process protects your reputation.
What advice would you give to business leaders who are hesitant to adopt AI because of fear, misconceptions, or lack of understanding?
I’d reframe the question entirely. The fear of adopting AI is understandable. But the more urgent question isn’t “What if we do this?” It’s “What is the cost of doing nothing?”
Every week of delay is a week your competitors are already experimenting to win. AI compounds — the organisations building capability now are creating institutional knowledge and optimized workflows that will be genuinely difficult to replicate later. McKinsey has some great data on this: high performers are 3.6 times more likely than other companies to have fundamentally redesigned their workflows around AI. You can’t close that gap by buying the same tool eighteen months later.
That hesitation usually comes from one of three places: fear of getting it wrong, fear of what it means for people’s jobs, and genuine uncertainty about where to start. For the first: start small and make failure cheap. For the third: use the prioritization framework — find your highest-frequency, lowest-error-risk task and automate that first.
In your opinion, how will AI continue to shape the business world over the next 5–10 years? Are there any trends or emerging innovations you’re particularly excited about?
The trend I’m watching most closely right now is agentic AI. This shift is already here and it’s messier than the headlines make it out to look.
The direction is clear: AI moves from the assistant role to an executor. It doesn’t just help you find a product — it finds it, orders it, arranges delivery. But the real-world friction is already revealing the tensions. Amazon just won a court order blocking Perplexity’s AI shopping agent — not primarily over data security, but because the agent bypassed ads and sponsored listings entirely. When an agent skips the aisle, everyone who monetises the aisle loses. That’s the business model conflict agentic AI creates, and platforms are already fighting over it.
On the consumer side, Target just updated its terms for their Gemini-powered agent: if the AI buys the wrong product, it’s the customer’s problem. The truth is, only 24% of consumers are comfortable letting an agent complete a transaction without final sign-off. Trust hasn’t caught up with capability yet.
For businesses, the real challenge is designing for both worlds. You have to be open enough for AI agents to access your platform, but you can’t get so tech-heavy that you lose the trust of the real people. And the longer arc: as AI handles more execution, genuine human interaction becomes scarcer (and more valuable). Human touch will become a premium feature. The brands that invest in that now will own the high end of every market they operate in.
How do you think the use of AI to solve business problems influences relationships with customers, employees, and the broader community?
On the consumer side, the shift has been less about the technology itself and more about what it makes possible for the people behind it. When our AI support assistant started handling the repetitive queries, it let our employees from the support team to focus on the cases that actually need them: complex situations, emotionally sensitive moments, where empathy matters a lot. That balance is what keeps our customer satisfaction rate at 92% positive reviews — well above the 80–85% e-commerce benchmark.
You are a person of great influence. If you could start a movement that would bring the most amount of good to the most amount of people through AI, what would that be?
My husband works at the intersection of IT and medicine, and it’s given me a clear view of where the most significant gaps are. If I could start a movement, it would be an AI-powered early detection in preventive health. Cardiovascular disease is still the leading cause of death globally — and so many lives could have been saved with an earlier intervention. The data and tools already exist. What’s missing is integration into the everyday touchpoints people already use for apps or devices. The technology is closer than most people realize, we just need to bridge that final gap to make it accessible to everyone.
How can our readers further follow your work online?
It’s been such a pleasure — thank you for the thoughtful questions!
The best place to follow my work is LinkedIn — I share perspectives on AI in operations, e-commerce, and what we’re learning at Udora as we scale across 50+ countries. For what we’re building as a company, the website is the place to go.
About The Interviewer: Chad Silverstein is a seasoned entrepreneur with 25+ years of experience as a Founder and CEO. While attending Ohio State University, he launched his first company, Choice Recovery, Inc., a nationally recognized healthcare collection agency — twice ranked the #1 workplace in Ohio. In 2013, he founded [re]start, helping thousands of people find meaningful career opportunities. After selling both companies, Chad shifted his focus to his true passion — leadership. Today, he coaches founders and CEOs at Built to Lead, advises Authority Magazine’s Thought Leader Incubator. Learn more at www.chadsilverstein.com
Vera Modenova Of Udora On How Artificial Intelligence Can Solve Business Problems was originally published in Authority Magazine on Medium, where people are continuing the conversation by highlighting and responding to this story.