What is artificial intelligence useful for? Automation of repetitive tasks for very busy people running small businesses, says a Berlin startup Synthflowwhich pronounces a $7.4 million seed round for its SME-focused no-code AI voice assistance platform.
The startup has raised a total of $9.1 million since its founding last spring, underscoring continued investor enthusiasm for accelerating applications of generative AI.
The startup also claims it’s closing in on 1,000 customers – touting “double-digit” monthly growth rates since it ditched stealth development in December 2023 and launched a browser-based “no-code” tool. This suggests that there is a healthy appetite among SMEs to adopt – or at least experiment with – generative AI tools that promise easy-to-achieve productivity gains.
The recent funding will go towards research and development, in keeping with Synthflow CEO and co-founder Hakob Astabatsyan, who says the team is keen to further fuel early momentum by increasing the product’s usability and expanding the scope of SMEs to whom it is attractive.
“We have a lot of ideas. We know exactly what customers need,” says TechCrunch.
A serial entrepreneur with a business background, Astabatsyan is a former Rocket Internet member. Joining him in his latest enterprise are his brother Albert, who also worked with him on a previous no-code startup; and Sassun Mirzakhan-Saky, who brings software engineering experience and CTO knowledge to the team.
Although the Synthflow product began with English calling support because its largest markets are English-speaking, it has since added German and French language versions (note: these are still in beta). The plan is subsequently to extend the focus on the latter markets in Europe.
A comprehensive experience
Call centers were the first to adopt AI voice agents, leveraging large language model (LLM) APIs to power systems that might answer phone calls humanly – with boundless energy and enthusiasm 24/7, if not at all times with perfect understanding.
Synthflow is going in a barely different direction, targeting its concept directly at SMEs involved in the service industry, including those in the smaller end of the category offering a do-it-yourself, no-code offering. The goal is to supply SMEs with an end-to-end experience, in keeping with Astabatsyan, who says the ROI from having the ability to automate basic tasks reminiscent of meeting scheduling shall be immediately obvious to focus on enterprises with limited resources.
“Artificial intelligence can do it cheaper and more reliably, and people can do other things,” is how he succinctly describes voice assistance.
He gives the example of a handyman or mechanic who often answers the phone himself when he isn’t working, which suggests he inevitably finally ends up missing a lot of calls and missing out on some business as a result; or a dentist who employs a receptionist who works limited hours, so again, she’s not at all times around to reply the phone.
Having a tool that may handle basic customer queries may be a game-changer for small businesses, Astabatsyan argues.
Synthflow’s target market is SMBs, which necessarily implies that the startup’s primary goal is to make AI technology accessible to non-technical users – which is why the company has created a no-code interface for its customers to design voice agents that meet the needs of their businesses.
“We wanted to try to build something simple,” he explains. “There is no layer of code on top [of AI agents] so that… business owners, business-minded people, can mess around with it, explore it and discover what LLMs can do for their firms.
Synthflow’s interface lets customers drag and drop elements to establish a voice AI that may perform specific tasks for them—reminiscent of scheduling meetings, reviewing often asked questions, or performing “information extraction,” reminiscent of obtaining personal information from a potential customer so a human can call them back.
“Suppose someone needs to make a call and needs to ask specific questions and collect specific information — especially static information such as address, home address, etc. — artificial intelligence works very well,” he argues.
The customer can configure the AI assistant to point that it is a robot. “I think it’s very good practice to disclose that it’s a virtual assistant,” says Astabatsyan. “My favorite opening is: ‘Hello. my name is [so-and-so], currently all our lines are busy. Sorry about that. I’m a virtual assistant here [the name of the business]. How can I assist you?’.”
According to Astabatsyan, one other necessary tool for voice AI is recognizing when to transfer a call to a human agent. Essentially using AI to filter incoming calls based on complexity – with automation caring for easy requests, which in turn increases the advantages by freeing up employees to have more time for more complex customer queries.
It emphasizes that the aim is not to interchange human jobs, but fairly suggests that AI might help SMEs be more productive and efficient than can be possible with their limited resources.
So, in addition to enabling customers to deploy voice agents, Synthflow is designed to also handle post-call data entry—for example, adding appointments to the calendar tool. Therefore, the next necessary task of the team is to build integration with third-party software.
“This is what artificial intelligence is so good at,” he argues. “Because he can accept this information [extracted from a call] and, say, update specific fields in a particular CRM system – and if you do this at scale, making hundreds or thousands of calls, suddenly we will see this technological advantage that we have seen [when businesses first adopted] computers.”
When it involves voice agents, the startup relies on OpenAI’s GPT LLM platform, but also uses its own artificial intelligence models, which Astabatsyan says have been trained on its own data and tailored to specific customer use cases.
It says it has also built its own “voice orchestration layer” that converts a customer’s speech into text, which might then be fed into an AI model as a prompt, returning an automatic response that the system converts from text to speech, which the customer hears as a synthesized voice on the other side of the telephone line.
For now, Synthflow is focused on using AI for inbound calls — which Astabatsyan suggests are easy targets for automation for resource-constrained firms. But it suggests more advanced capabilities in development, with R&D being fueled by a fat seed round.
He mentions that one of the things they are working on is a feature that may enable Synthflow’s voice AI to perform what he calls “live action” or “calls” – meaning that while on a call, the AI will have the ability to ascertain live broadcast of stocks in the warehouse. Or you possibly can take other information you wish and “move it somewhere else,” as he puts it.
It also sketches a scenario in which task-oriented AI voice systems will have the ability to collectively increase their usefulness. They could hand off the call to a different dedicated voice AI trained for various tasks requested by the customer.
“The key here is to focus on who your customers are. Because depending on who you’re building it for, your product is going to be very, very, very different,” he adds.
One impact to think about is whether voice AI and voice assistance systems can live as much as productivity expectations – efficiently delivering on the promise of efficiently handling the entire layer of customer inquiries, including by expertly routing more complex things to the appropriate system or human , which shall be dealt with – this might mean that the average SME will find that they have much more work than they will handle.
“I think that’s an interesting question for a lot of managers and leaders to think about, right?” – he replies, discussing this scenario. “For example, if there is so much production capability and productivity is released, how do we direct these human resources to other sectors of the economy? Because I think this query hasn’t been answered yet, but it’s a really very interesting query.
Synthflow’s seed funding is led by Singular, with participation from existing investor Atlantic Labs and a variety of investors in the AI space, including the founders of Krisp AI.