Publishers

Need unique free news content for your site customized to your audience?

Let's Discuss

By: citybiz
October 8, 2025

Curated TLDR

Q&A with Hakob Astabatsyan, CEO and Co-Founder of Synthflow

Hakob Astabatsyan is the CEO and Co-founder of Synthflow, the world’s #1 voice AI agent platform. He’s scaled the company 13X in ARR, powering over 45 million AI-driven calls for 1,500+ companies around the world—without a single second of hold music.

Before Synthflow, Hakob built ventures across multiple markets with Rocket Internet and shaped growth strategies at BCG. He’s spent the past decade solving complex problems at speed, turning strategy into execution and ideas into adoption. At Synthflow, that means building voice AI infrastructure that isn’t just technically advanced—but usable, scalable, and built for real-world impact.

Why did you start Synthflow?

I have a business background – I’ve worked in strategy consulting at BCG and venture building at Rocket Internet – but I’ve always been interested in how emerging technologies can unlock new levels of customer experience and operational efficiency.

AI, and particularly Generative AI, clearly has a major role to play in those areas. The way GenAI has evolved in recent years led me, along with my two partners, Albert and Sassun, to start Synthflow in early 2023.

The catalyst was the emergence of large language models, in particular ChatGPT, into the mainstream. While we could see that it had significant potential, we also knew that it needed a better interface for business users. At first, we focused on building a no-code layer on top of GPT for text, but we quickly realised that the real opportunity was in voice.

It’s certainly harder to develop a voice interface with a great experience than it is with text. Delivering lower levels of latency, reducing the number of interruptions and getting timing right takes a lot more work, but it’s the sort of challenge we love. We pivoted fully to voice and, by the summer of 2023, had launched the current version of Synthflow.

How has voice AI evolved over the last few years?

Up until 2023, automated voice was largely confined to interactive voice response (IVR) trees: press one for support, two for sales, that sort of thing.

That all changed with the release of OpenAI’s GPT-4 model and the shift of LLMs into the public consciousness. In voice, we started to see how we could create conversational AI in practice, rather than the theory that had dominated the industry.

Of course, at first it was still relatively niche: developers integrated speech-to-text and text-to-speech commands, layering specialist speech programs on top of general LLM to create technology that simulated real-time, human-like conversations.

Now that there are more voice-to-voice models, we’ve been able to move away from that. We’ve decreased the latency of conversations, we’re managing interruptions effectively and we’re able to produce natural, human-like conversations. That includes our new Memory feature, which allows AI voice agents to retain context and knowledge across customer interactions.

The update is designed to reduce repetitive verification, streamline calls, and support more efficient resolutions, whether for booking appointments, managing accounts, or handling follow-ups.

Will voice AI replace contact center agents?

Whenever we’re asked about whether voice AI will replace human workers, or even whether AI in general will remove the need for workforces, I always think about where else we’ve been automating. If we look at customer service, we’ve been automating for years, whether it’s emails, support tickets or internal workflows.

Voice was left behind, but now that it is catching up, most people’s first reaction is fear and distrust: it’s going to lead to more spam calls, or take human jobs.

I think part of the problem is that, when it comes to voice, most people’s initial experience is the worst aspect of automation: scripted, robotic, literally a computer saying no. We’ve been very conscious of that from the beginning of Synthflow, and that’s why we’re always focused on delivering real, useful conversations, not robocalls or mass sales blasts.

We’re being used to help people reschedule an appointment, get support when the office is closed, or quickly confirm an insurance detail. The sort of calls that are so important to customer satisfaction, but are also expensive for businesses to service. Up until now, this put a lot of companies in a bit of a bind; if they try to turn customers away from those sorts of queries, it can impact overall satisfaction, but deploying human agents for every small query impacts margins.

That’s why, when deployed correctly, voice AI can keep everyone happy: customers get their service they’re looking for, while businesses can give their teams relief from repetitive calls and spend more time on truly valuable tasks: helping people with complex problems and delivering that critical human touch at the right moment.

You recently announced a new round of funding, just a year after securing seed funding. How has the market for voice AI changed in that time and how was that reflected in the conversations you had with investors?

To have secured a total of $30 million in funding in a little over two years highlights how rapidly voice AI is evolving: both the underlying infrastructure and the agentic layers people interface with are improving every day, and the cost of deployment is coming down.

What was particularly clear in our conversations with both our new lead investor, Accel, and our existing backers that participated in this round was the level of awareness of the market opportunity. Two years ago, it was about what could theoretically be achieved; this time around, it’s very much about what’s been achieved, and what that means for future use cases, particularly in highly regulated industries such as financial services and healthcare.

Co-Founder / CEO Hakob Astabatsyan

Connect »

Company Synthflow AI

Learn More »

How is Synthflow different to competitors?

There are two main ways we differ from competitors: speed and being AI-native.

First, speed. Most voice AI companies we see take months to launch, require custom engineering, or push customers into rigid setups that reflect vendor needs, not those of the buyers.

We deliver fast deployments, low latency, full integrations and complete reliability. Our customers have a choice of a no-code builder or our full API suite, depending on their needs, but both are fully compliant with SOC 2, HIPAA, and GDPR and will go live in under three weeks.

Second, we’re AI-native. We’re built for the LLMs available today, rather than AI-washed legacy IVR bots. Our responsiveness, conversations and interruption handling are driven by AI, making it more natural and trustworthy.

The AI era is characterised by being able to move quickly without compromising on reliability. Many of our competitors are pushing out-dated solutions that have been rebadged for AI. We’re AI to our core, and that means we can adapt as the underlying infrastructure evolves.

Thinking more broadly, how do businesses deploy voice AI while retaining customer trust?

You need to be clear on why you’re deploying voice AI – is your aim for the conversation to be the focus, or for it to be the means to an end? In a customer service setting, it’s usually the latter: you want the conversation to generate an action.

That doesn’t mean businesses should see it as a way of pushing customers through as quickly as possible. They should be looking at it as an opportunity to build trust. Anything too robotic can derail that process, but the technology is continually improving with solutions becoming more human-like.

That’s why it’s so important to be clear on where you are deploying voice AI. In a B2B context, where conversations can be more straightforward, having an automated element can deliver significant trust and efficiency gains. More complex issues will require human interactions, but as teams will have been freed up from repetitive and straightforward calls, they’ll have more time to focus on these more challenging instances.

Businesses are already benefiting significantly from this technology, and there is still immense potential as new advancements continue to break through. Our customers are experiencing impressive results, such as 2.5x more qualified appointments, a 24% increase in answered calls, 12% more closed sales and a 25% increase in satisfaction.

What are the three things you would advise any business considering deploying voice AI to do?

First, define your business objectives. What are you trying to achieve: efficient operations, lower cost to serve, increased customer satisfaction? You need to be clear on your strategic goals before you even consider finding a solution. Don’t get distracted by the latest shiny new technology; otherwise, you might end up with a solution looking for a problem.

Second, be clear on what’s required to implement and manage the solution that’s going to help you achieve your goals. Do you have the right talent in-house? What will it mean from an organisational change perspective? Say AI, and people do start to worry about their jobs, so have a plan in place to communicate what it means to the relevant teams and demonstrate how your approach is going to make their lives better.

Third, choose your partners carefully. Voice AI might be a new technology, but you should approach it with the same rigour as any other procurement process: look at the company and what it offers, test the technology, and ensure it’s compliant. Make sure you understand how much it will cost you as well; there are extremely affordable options available that democratise voice AI technology and can be live in a month or two.

What do you think the next big trend in voice AI will be?

We’re rapidly reaching a point where more than half of B2B conversations will be managed by AI. We may well get to a point where one AI speaks to another. People laugh, but look how quickly things have evolved; I always think that, however fast you think AI is going to change, you need to divide that time by two.

It’s important to remember that technology goes through phases where it is considered niche, but it then reaches a tipping point where it goes mainstream. How many people had heard of Gen AI three years ago? Now look at its reach. The application of voice AI is having a similar moment.

We’ll also see an increase in voice AI’s problem-solving capabilities. As mentioned before, right now the deployment works best with straightforward queries, but as the technology develops, we’ll start to see it take on more complex tasks.

The post Q&A with Hakob Astabatsyan, CEO and Co-Founder of Synthflow appeared first on citybiz.

Blockchain Registration, Verification & Enhancement provided by NewsRamp™

This contant was orignally distributed by citybiz. Blockchain Registration, Verification & Enhancement provided by NewsRamp™. The source URL for this press release is Q&A with Hakob Astabatsyan, CEO and Co-Founder of Synthflow.

{site_meta && site_meta.display_name} Logo

citybiz

citybiz is a publisher of news and information about business, money, and people - including interviews, questions and answers with thought leaders. citybiz reaches business owners, C-level, senior managers and directors in 20 major U.S. city markets.