AuthBridge CEO on AI-driven solutions for Digital Identity Verification and market expansion

AuthBridge, a technology company focused on digital identity verification and risk management solutions, aims at helping organisations improve onboarding and compliance processes across various industries, including banking, financial services, and insurance. The company operates in three main segments: Workforce Solutions for candidate onboarding and verification, Financial Intelligence Solutions for digital onboarding, and Business Solutions for Know Your Customer (KYC) and vendor due diligence.

In a conversation with TechCircle, CEO Ajay Trehan discussed how AuthBridge uses Artificial Intelligence (AI) and automation to streamline employee screening and customer onboarding for over 2,500 businesses. As AuthBridge expands into new markets, he shared insights on the company's approach to enhancing trust and security in business operations. Edited Excerpts:  

Can you tell us about your company and its core areas of expertise?

We've been in business for 19 years. We are the largest company focused on simplifying trust through technology, working as a B2B firm. We serve over 2,500 companies across India, including some of the biggest names in industries like IT, banking, airlines, and hospitality.

For these companies, we solve various trust-related problems. Our main service is employee screening. We offer platforms and instant Application Programming Interfaces (APIs) that allow our clients to screen employees during onboarding or on an ongoing basis. For certain categories of employees, companies may want to continuously monitor their criminal records, and we provide the technology and platforms to make this possible. 

Another key service we offer is for banks, financial institutions, and insurance companies. We help them digitally onboard customers, conduct KYC, KYB (Know Your Business), video interviews, and video-based KYC, as permitted by regulators. Our system automatically detects liveness and flags document fraud during the process.

The third major area we address is for corporates, helping them onboard and screen vendors, distributors, freelancers, consultants, or even customers. When a company engages with a larger ecosystem, there are multiple external touchpoints. Before doing business, they need to know key details about these vendors or partners, like their financial capabilities, practices, and the background of their owners. We assist companies in this process. 

Recently, we have also expanded our business to the US and the Middle East. 

How do you use technologies like AI and automation in your processes to serve various industries?

One of the key problems we solve is processing millions of data records quickly. In just a few seconds, we can give information about an individual. 

India currently has around 30 to 32 crore cases across 3,800 courts, involving about five and a half crore individuals. None of these cases are linked to a national ID.

When clients ask us to check for criminal records, our AI algorithms go through these 300 million records, identify personas, and provide a match. The system might tell us with 95% or 97% probability whether a case matches the individual we’re looking for. This would have been impossible a few years ago with so much unstructured data.

We use artificial general intelligence to read, extract, and interpret these documents. The system can determine if the nature of a case could negatively impact a client’s reputation, and tag it as such. 

We also use technology to mine data from social and public websites through robotic process automation. This allows us to scan thousands of websites in seconds, extract data, and clean it with AI engines. At any given time, we handle 30,000 to 40,000 new onboardings every day, which wouldn’t be possible without the deep use of technology. 

How do you ensure the accuracy of these technologies?

At AuthBridge, we work with five different large language models (LLMs). Different LLMs produce different output quality based on their training data. For example, ChatGPT is good at analysing and interpreting text, while other models handle numbers better. Depending on the problem we're solving, we choose the right model or combination of models to ensure the best quality output.

After the AI generates results, we have a human review the output at every stage. In the past, we had to manually read documents and extract data, but now AI extracts the data, and humans compare it with the original dataset to verify its accuracy. We check accuracy at multiple levels: first, we assess how much of the needed data was extracted, and then we evaluate how accurate that extracted data is. This human review improves the overall process continuously.

When we first started using these AI engines, accuracy was around 85-86%. Today, most of our engines operate at 95-96%, and for some data extractions, we reach 98-99%. However, since AI can never be 100% accurate, and given the sensitive nature of our business, we always keep humans in the loop to ensure the decisions made by the AI are reliable.

When it comes to document reading, employment verification, and background checks, what technological advancements do you foresee becoming game changers in the next few years?

AI has had two major impacts on our operations. First, we used to rely heavily on manual processes, but with AI, we’ve automated many repetitive tasks, especially around data extraction and interpretation. This has allowed us to redeploy our workforce to more important tasks.

Second, AI has enabled us to develop products that would have been too complex without it. For instance, conducting thorough criminal record checks across India is now possible. For 70% of cases, we can complete checks in seconds. For the remaining 30%, which need human involvement, it takes just a few hours. This is a game changer because it allows us to offer new services that weren’t possible before.

Previously, clients only asked for background checks during onboarding, but now they ask us to monitor employees, vendors, and customers continuously. This helps them detect issues in real time.

We’ve also seen improvements in our SaaS (Software-as-a-Service) offerings. When we onboard smaller companies, we used to struggle with configuring workflows for specific use cases. Now, with AI, creating a workflow or rule engine is as simple as chatting with the system. A configuration analyst or even the customer can input a request like: “I need Aadhaar verification, followed by PAN, followed by GST and MCA documents, and if the liquidity ratio is below a certain level, mark it red.” The system automatically creates the input, workflow, and rule engine based on that command, which was impossible a year ago.

This capability has opened up new markets for us, particularly the SME market, which was previously hard to serve. It’s also made it easier for us to adapt workflows for specific countries and regions. 

How does your company differentiate its fraud detection and verification solutions from competitors? Additionally, how does your system keep up with fraudsters in today's rapidly changing digital landscape?

In different verticals, we compete with various companies, but none offer a comprehensive suite of trust simplification products across all use cases like we do. When I work with a customer, I can address any trust-related issues they face, without them needing to find separate solutions elsewhere. This is one of AuthBridge's biggest differentiators compared to competitors, thanks to our extensive use of technology.

As for staying ahead of fraudsters, it's a difficult challenge. The best way is to stay in close contact with our customers. They often alert us early about potential fraud risks. We then assess if we can solve the problem. We work with many top companies across industries, and through our close engagement, we gain insights into their regulatory conversations and industry challenges, which could lead to fraud and impact their business.

Each meeting gives us new ideas to take back to our lab and develop solutions. But in reality, we're often playing catch-up with fraudsters, which is unfortunate but a part of the landscape.

What are your main growth initiatives for the next two to three years? Can you share any upcoming projects, product innovations, or partnerships that you have in the pipeline? 

These days, we focus on growth over shorter timeframes, especially with the fast pace of AI and technology advancements. It’s crucial for us to stay aware of changes in the customer landscape and new regulations, so we can develop products that make things easier for our customers. One key area of growth is deepening relationships with our existing customer base, which drives a significant portion of our expansion.

At the same time, we are organically acquiring new customers in India, and we see growth opportunities in Dubai and the US, where we’ve recently opened offices. Many of our current customers are already doing business in these regions, so expanding there allows us to strengthen those partnerships. 

We also have inorganic growth plans. We're always on the lookout for companies doing innovative work with technology, and we explore potential partnerships or strategic opportunities with them. While there’s nothing concrete at the moment, we expect developments soon.

How do you plan to adapt your new services or solutions for new regulatory environments or some other emerging markets?

Because we are a technology and digital-first company, our solutions are configurable with or without AI. For example, when we expanded into the UAE, we didn’t need to create new products or platforms. Our existing enterprise solutions were already capable because we have access to international data, which is integrated into our systems. This made it straightforward for us to operate in those markets.

The US, however, has specific regulations, so we had to adjust some of our tools to meet those requirements. But this wasn't difficult, given our experience handling millions of screenings for large customers in India, so we were able to adapt.

On the regulatory side, we engage closely with our customers, many of whom work directly with regulators. Regulators often consult with groups or councils that include our customers when considering new regulations. Through these customers, we learn about emerging areas and work with them on proofs of concept (POCs). We’ve also presented our solutions to regulators in the past to explore their potential adoption for the wider industry. 

What technology trends are you seeing across the industries you serve, and what specific technologies are you considering for future research and investment?

These days, the big question for us is how to leverage AI in business. This is also a priority for our customers. We're constantly looking for more problems that AI can help solve. 

We might be one of the few organizations in the country with our own AI Council and a Generative AI Centre of Excellence at AuthBridge. We're also rolling out an enterprise-wide AI tool for our employees to use in their daily work. 

In the short term, I see AI significantly changing how technology is used worldwide, not just in India. Even conservative industries like banking and financial services, which usually stick to traditional technology, are starting to seriously consider AI tools. They realise that if they don’t, others who adopt AI will gain a major competitive edge by reducing costs. 

From text analysis and number crunching to internal audits, face matching, and location tracking, AI is now a strong component in many technologies. For example, Airtel recently used AI to detect spam calls, identifying millions of numbers in just 18 days of trial. This shows how powerful AI is compared to traditional technologies.