Fini is an accuracy-first AI for enterprise customer support. Fini recently raised $3.6 million in seed funding led by Matrix Partners (with Y Combinator and notable angels) to solve AI’s biggest problem in regulated industries: hallucinations. Pulse 2.0 interviewed Fini CEO Deepak Singla to gain a deeper understanding.
Deepak Singla’s Background
Could you tell me more about your background? Singla said:
“I graduated from IIT Delhi with a degree in Mechanical Engineering and a minor in Business Management. After graduation, I started my career at PwC, which gave me a strong foundation in business operations and strategy. But I was drawn to the tech world, particularly the challenge of scaling products that serve millions of people.”
“That led me to Uber, where I spent several years working on customer experience and growth. At Uber, I was responsible for resolving over 4 million customer interactions monthly. That experience was transformative, I saw firsthand both the power of AI to transform customer support and the massive gaps in existing solutions. We were using sophisticated ML models at Uber, but when I looked at what was available to other enterprises, the options were either basic chatbots that frustrated customers or expensive, low-quality solutions that didn’t actually solve problems.”
“My co-founder Hakim Khalafi and I met at Uber. Hakim has a background in software engineering and analytics with over 10 years of experience, and a master’s degree in Engineering Mathematics from École Polytechnique. Together, we saw an opportunity to bring enterprise-grade AI support to companies that couldn’t build it themselves. That vision became Fini.”
Formation Of The Company
How did the idea for the company come together? Singla shared:
“The idea for Fini emerged directly from our frustrations at Uber. We had built incredibly sophisticated machine learning systems that could understand customer intent, and make decisions. These systems were resolving millions of queries with high accuracy, but this capability was locked inside Uber’s engineering infrastructure.”
“Meanwhile, most enterprises were stuck with two bad options: generic AI chatbots that could barely answer FAQs, or expensive support teams that couldn’t scale. There was a massive gap in the market for AI that actually worked, AI that could understand complex queries, access real-time data, and take meaningful actions like processing refunds or updating account details.”
“The breaking point came when we realized that customer support teams are often the most ignored part of an organization, treated as a cost center rather than a strategic asset. They’re sitting on a goldmine of customer insights, but product and leadership teams are disconnected from that feedback loop. We knew AI could bridge that gap and transform support from a cost center into a competitive advantage.”
“We joined Y Combinator’s Summer 2022 batch. YC helped us validate that this problem was universal, companies of all sizes were struggling with the same issue. Within months of launching, we had our first paying customers, and we haven’t looked back since.”
Favorite Memory
What has been your favorite memory working for the company so far? Singla reflected:
“There are two moments that stand out.”
“The first was our Y Combinator Demo Day in 2022. Standing there after three intense months of building, learning, and iterating, with a product that customers actually loved, was surreal. We had gone from an idea to a real business solving real problems, and the YC network opened doors we never imagined possible.”
“The second, and honestly more meaningful, moment was when we received feedback from one of our first customers – ColumnTax. They told us they’d automated over 94% of their support queries in just three months with Fini, and that it was “a complete no-brainer for any business.” Hearing that kind of impact, especially from a customer who had evaluated six different AI tools before choosing us, reinforced that we’re not just building technology, we’re genuinely transforming how businesses serve their customers.”
Core Products
What are the company’s core products and features? Singla explained:
“Fini’s flagship product is Sophie, our AI support agent that works like your best human agent, but 24/7 and at scale.”
Differentiation From The Competition
What differentiates the company from the competition? Singla affirmed:
“What makes Sophie different:
- RAGless Architecture: Most AI support tools use Retrieval-Augmented Generation (RAG), which is essentially fancy search and summarization. Sophie doesn’t just retrieve and summarize; she understands, reasons, and acts. Our proprietary RAGless infrastructure delivers 98% accuracy with zero hallucinations, unlike fragile RAG-based systems.
- Agentic AI Execution: Sophie doesn’t just answer questions; she takes real actions. She can issue refunds, check KYC status, pause subscriptions, update account details, and execute complex workflows using our proprietary Agentic AI framework. This is miles beyond what traditional chatbots can do.
- Self-Learning Knowledge Engine: Sophie learns from every interaction. She doesn’t just reference a static knowledge base, she actively builds and improves her understanding over time, getting smarter and more accurate with each customer conversation.
- Enterprise-Grade Security: We’re built for enterprises with the most stringent regulatory compliance requirements. Fini is compliant with SOC2, ISO 27001, PCI, and HIPAA standards. We’re also GDPR compliant and have EU data residency available for companies operating in Europe.
- Seamless Integration: Sophie works on top of your existing help desk, whether it’s Zendesk, Intercom, Salesforce, Front, or others. Deploy in minutes, not months.
Results: Sophie resolves 80%+ of customer queries with zero human intervention, boosts CSAT by 10%, and reduces support costs by 50%. We’re the highest-rated AI support agent on G2 with a perfect 5/5 star rating.”
Challenges Faced
Have you faced any challenges in your sector of work recently, and how did you overcome those challenges? Singla acknowledged:
“The biggest challenge in the AI support space is trust. Every vendor claims their AI is accurate, but most enterprise customers have been burned by chatbots that hallucinate, give wrong answers, or make customers more frustrated.”
“We tackled this head-on with radical transparency and accountability. First, we built our Trust Metrics framework, we openly share our accuracy rates, resolution rates, and CSAT scores, and we benchmark ourselves against competitors. Second, we introduced the Enterprise Fini Guarantee: for companies handling over 1 million annual tickets, if our AI doesn’t deliver 80% resolution rate, improve CSAT, and maintain under 30-second response times within 90 days, they pay nothing.”
“This was risky, but it forced us to be genuinely confident in our technology. It also differentiated us in a market full of empty promises.”
“Another challenge has been educating the market about the difference between basic RAG-based chatbots and true agentic AI. Most customers don’t understand the technical differences, they just know that most AI tools don’t work well. We’ve addressed this through customer storytelling and transparent case studies. When potential customers see that ColumnTax automated 94% of their queries with 98% accuracy, or that Qogita achieved an 88% resolution rate and 121% SLA improvement, the value becomes immediately clear.”
Evolution Of The Company’s Technology
How has the company’s technology evolved since launching? Singla noted:
“We’ve made three major technological leaps since launching in 2022:
Phase 1: Knowledge-Based AI Agent – We started with Sophie as a knowledge assistant, basically a very smart FAQ bot that could pull answers from documentation. This got us to around 60-70% resolution rates for simple queries. It was good, but not transformative.
Phase 2: Agentic AI with RAGless Architecture – This was our breakthrough. We moved away from traditional RAG approaches and built our proprietary RAGless infrastructure. Instead of just retrieving and summarizing information, Sophie could now reason through complex scenarios and understand context deeply. We also introduced our Agentic AI Execution Framework, allowing Sophie to take real actions, issuing refunds, checking account statuses, and updating records. This pushed us to 80%+ resolution rates with 98% accuracy.
Phase 3 (Present): Self-Learning Knowledge + Product Intelligence Now Sophie doesn’t just execute, she learns and provides strategic insights. Our self-learning knowledge engine means Sophie continuously improves her understanding from every interaction. We’ve also added Product Intelligence capabilities that turn customer conversations into actionable insights for product, CX, and operations teams. This transforms support from a reactive cost center into a proactive strategic asset.
The evolution has been about moving up the value chain: from answering questions → to taking actions → to generating business intelligence.”
Significant Milestones
What have been some of the company’s most significant milestones? Singla cited:
“Two milestones define our journey:
Breaking the 1 Million Barrier: After 9 months of intensive R&D, we achieved a major technical breakthrough, resolving 1+ million support issues monthly while maintaining an 80% average resolution rate and 98% accuracy. This wasn’t just a volume milestone; it proved our RAGless architecture could scale without degrading quality. Most AI support tools break down at scale. We got better.
Earning Enterprise Trust: We signed marquee customers, including the US Chamber of Commerce Foundation, Swisscom, Bitdefender, Atlas, and Found. These aren’t typical SaaS buyers; they’re all in highly regulated industries with stringent security, compliance, and accuracy requirements. Winning their trust validated that Fini meets the highest standards for mission-critical customer support.
Other key moments: Y Combinator S22, achieving SOC2/ISO 27001/GDPR compliance, raising our seed round from Matrix Partners and angels from Uber/Intercom/SoftBank, and becoming the highest-rated AI support agent on G2.
The milestone I’m most proud of isn’t a number, it’s the fact that in an industry full of inflated claims, we deliver on our commitments.”
Customer Success Stories
Can you share any specific customer success stories? Singla highlighted:
ColumnTax evaluated six different AI support tools before choosing Fini. They needed to enhance customer experience, achieve high automation for complex tax queries, and automatically create new content. Within three months, they automated 94% of their support queries with 98% accuracy. Their CX Lead, Lori McCool, called it “a complete no-brainer for any business.”
Qogita, a B2B wholesale marketplace, needed to improve its response times and resolution rates while maintaining high accuracy. Sophie now resolves 88% of their support queries with 97% accuracy and improved their SLA performance by 121%. Their RevOps Lead, Clara Girardeau, noted that being able to make tweaks themselves to continuously improve Sophie’s performance was a game-changer.
Atlas, a fintech offering a 0%-APR rewards credit card, needed to automate sensitive workflows like address changes and transaction queries without compromising compliance standards. Fini’s AI (Sophie) now automates 80% of their key journeys with zero compliance exceptions, delivering answers in under 60 seconds. Their Head of Customer Experience and Operations, Crystal Singh, noted that Fini proved automation could safely handle “real work” with compelling accuracy, proving that fintechs avoiding AI support are “leaving money on the table.”
These stories share a common thread: our customers aren’t just reducing costs, they’re actually improving customer experience. That’s the holy grail of AI support.
Funding/Revenue
Are you able to discuss funding and/or revenue metrics? Singla revealed:
“We’ve raised funding from Y Combinator, Matrix Partners, and a stellar group of angel investors from companies like Uber, Intercom, SoftBank, McKinsey, and Twitter. While we don’t disclose exact funding amounts, we’ve been capital-efficient and focused on building a sustainable business.”
“On the traction side, we’re currently solving 1M+ support issues every month, growing 15-20% month-over-month and are on track to cross $10M ARR in 2026.”
“What’s more important than the raw numbers is our unit economics. Our Enterprise Fini Guarantee shows our confidence: we only get paid if we deliver results. That forces us to build technology that actually works, and it aligns our success with our customers’ success.”
“We’re profitable on a unit basis and could be profitable overall if we chose to be, but we’re reinvesting heavily in product development and expanding our enterprise capabilities. The AI support market is massive, estimated at over $50 billion, and we’re positioning ourselves to capture a meaningful share.”
Total Addressable Market (TAM)
What total addressable market (TAM) size is the company pursuing? Singla assessed:
“The global customer support market is estimated at over $50 billion annually, with AI support representing a massive opportunity to transform how businesses handle customer interactions.”
“More specifically, we cater to mid-market and enterprise B2C companies in 2 core verticals:
- Fintech, Banking, and Insurance: Companies handling sensitive financial queries, KYC processes, transaction support, and account management. This segment requires the highest accuracy and security, which is exactly where we excel.
- Healthcare & Fitness Tech: Platforms managing memberships, bookings, and personalized health/fitness guidance. These require empathetic, accurate support, another sweet spot for Sophie.
Within these verticals, we focus on companies handling 100,000+ support tickets annually. That’s roughly 50,000+ companies globally representing a $50B+ addressable market for our solution.
But here’s what excites me most: we’re not just competing for existing support budgets. We’re enabling companies to scale their customer experience in ways that weren’t possible before. That expands the market beyond pure replacement; it’s about unlocking new growth opportunities through better customer retention and satisfaction.”
Differentiation From The Competition
What differentiates the company from its competition? Singla affirmed:
“The AI support space is crowded, but most solutions fall into two camps: basic chatbots or expensive, generic ‘AI’ that doesn’t actually work. Here’s what sets Fini apart:
- Actual Results: Our competitors claim 30-60% resolution rates. We deliver 80%+ consistently. ColumnTax hit 94%. Qogita hit 88%. These aren’t marketing numbers, they’re verified customer outcomes.
- RAGless Architecture: While everyone else uses fragile RAG systems that hallucinate and break down on complex queries, we’ve built proprietary RAGless infrastructure that delivers 98% accuracy. This isn’t incremental improvement; it’s a fundamentally different approach.
- True Agentic AI: We don’t just retrieve information, we take action. Sophie can issue refunds, check KYC status, update subscriptions, and execute complex workflows. Most “AI agents” are still glorified chatbots.
- Enterprise-Grade Security: Regulated industries that we work with have high compliance standards and security measures, and so do we. As of today, Fini is 100% compliant with GDPR, SOC2, ISO 27001, PCI, and HIPAA standards. We don’t train on customer data, and we take security seriously from day one.
- Performance Guarantee: We’re the only AI support vendor offering a guarantee: 80% resolution rate, CSAT improvement, and < 30-second response times within 90 days, or you pay nothing. We put our money where our mouth is.
- Customer Obsession: We’re a small, lean team of 8, which means every customer gets white-glove treatment. Our customers work directly with our founders and engineering team. We’re not just a vendor, we’re a strategic partner.
The competitive landscape is simple: Intercom Fin claims 60% resolution, Zendesk AI achieves around 30%, Salesforce AI sits at 40%, and Ada tops out at 50%. We deliver 80%+. The gap isn’t closed.”
Future Company Goals
What are some of the company’s future goals? Singla emphasized:
Near-term:
- Launch Fini 2.0, where we reimagine support experiences from rule-based agents into context-aware concierges
- Scale to $10+ million ARR while maintaining our lean team structure and high efficiency
- Expand our enterprise customer base in fintech and healthcare, where security and accuracy are paramount
- Launch Sophie Voice, a voice-enabled AI support for phone channels
- Deepen our Product Intelligence capabilities to support a strategic, revenue-generating function
Medium-term:
- Achieve 100+ million support interactions resolved annually
- Expand into new verticals like SaaS, gaming, and education
- Build industry-specific AI agents pre-trained for fintech compliance, e-commerce operations, and healthcare privacy requirements
- Launch Sophie Workflows, a visual builder for creating custom AI-powered support processes without code
Long-term Vision: We believe customer support is fundamentally broken. It’s reactive, expensive, and disconnected from the rest of the business. Our vision is to transform support from a cost center into a strategic growth engine.
Imagine a world where every customer interaction generates insights that improve your product, where support teams are empowered by AI to be 10x more effective, and where customer satisfaction goes up while costs go down. That’s the world we’re building.
We’re not just automating support, we’re reimagining the entire customer experience stack for the AI era. That’s the future we’re chasing, and we’re just getting started.”
Additional Thoughts
Any other topics you would like to discuss? Singla concluded:
The Responsibility of Building AI: As AI becomes more prevalent in customer-facing roles, we have a responsibility to build it ethically and transparently. That’s why we created our Trust Metrics framework, openly sharing accuracy, resolution, and performance benchmarks. We believe transparency builds trust, and trust is the foundation of AI adoption.
We’re also committed to responsible AI deployment. Sophie is designed to assist and augment human agents, not replace them entirely. The best outcomes happen when AI handles repetitive queries while human agents focus on complex, high-value interactions that require empathy and judgment.
Why Amsterdam?: People often ask why we’re based in Amsterdam rather than Silicon Valley. The answer is simple: Europe has incredible engineering talent, a strong regulatory framework that forces us to build secure products, and proximity to enterprise customers who value privacy and compliance. Plus, Amsterdam gives us a European beachhead while remaining well-connected to the global tech ecosystem.
The AI Support Revolution: We’re still in the early innings. Most companies are just starting to understand what’s possible when AI can truly think, reason, and act. The winners in this space won’t be the companies with the fanciest marketing; they’ll be the ones with the best technology and the most customer obsession.
At Fini, we’re building for the long term. We’re not chasing hype, we’re solving real problems with real technology that delivers real results. That’s what sustainable businesses are built on, and that’s what our customers deserve.
