What is AI good for? Berlin-based startup Synthflow wants to automate repetitive tasks for busy people who run small businesses. The company is announcing a $7.4 million seed round for its no-code platform for AI voice assistants for small businesses.
The startup has raised a total of $9.1 million since it was founded around spring of last year, underscoring investors' continued enthusiasm for accelerating the applications of generative AI.
The startup also claims to be approaching 1,000 customers, boasting “double-digit” monthly growth since it secretly wrapped up development and released its browser-based “no-code” tool in December 2023. This suggests there's a strong appetite among small and medium-sized businesses to adopt (or at least try) generative AI tools that can easily deliver productivity gains.
The new funding will be put into research and development, and the team hopes to continue building on its early momentum by improving the product's viability and broadening the reach of small and medium-sized businesses it targets, according to Synthflow CEO and co-founder Hakob Astavazian.
“We have a lot of ideas. We know exactly what our customers need,” he told TechCrunch.
A serial entrepreneur with a business background, Astabazian is formerly with Rocket Internet. His latest venture brings him on board with his brother Albert, with whom he previously worked at a no-code startup, and Sassan Mirzakhan Sakhi, who brings a software engineering background and CTO expertise to the team.
Synthflow's product started off with call handling in English, as their largest market is English-speaking, but they've since added German and French versions (note: these are still in beta), so they're considering focusing on the latter market in Europe.
End-to-end experience
Call centers were early adopters of AI voice agents, leveraging APIs for large language models (LLMs) to enable systems that, while not always with perfect understanding, can answer calls in a human-like manner, 24/7, 365 days a year, with tireless energy and enthusiasm.
Synthflow takes this concept in a slightly different direction, targeting small and medium-sized businesses focused on the service industry directly, including smaller players in the category that offer DIY, “no-code” products. Astabatsyan says the goal is to provide an “end-to-end” experience for SMEs, and he argues that the return on investment from being able to automate core tasks like appointment scheduling will be immediately apparent to the resource-limited businesses he targets.
“AI can do it cheaper and more reliably, freeing humans to do other tasks,” is his succinct sales pitch for voice assistants.
He gives the example of handymen or mechanics who often answer their own phones when they're not working, meaning they inevitably miss many calls and lose some business as a result, or dentists who employ receptionists who only work limited hours, and who also aren't always available to take calls.
Astavazian argues that having tools that can handle basic customer inquiries can be a game changer for small businesses.
Synthflow's targeting of small and medium-sized businesses necessarily means that the startup's core focus is making AI technology accessible to non-technical people, so it has built a no-code interface to let customers design a voice agent to suit the needs of their business.
“I wanted to create something simple,” he explains, “with a no-code layer on top. [of AI agents] This will allow business owners and business-minded people to really try this out and understand and explore what the LLM can bring to their business.”
Synthflow's interface allows customers to drag and drop elements to configure the voice AI to perform specific tasks, such as scheduling an appointment, reviewing FAQs, or performing “information extraction” such as getting personal information from a potential customer so a human can return the call.
Image credit: Synthflow
“For example, AI can be extremely useful when someone calls and you need to ask them a specific set of questions or gather specific information, especially static information like their address or home address,” he argues.
Customers can choose to have their AI assistant reveal that it's a robot. “I think it's a really good habit to reveal that you're a virtual assistant,” Astabazian said. “My favorite opening line is, 'Hello. My name is [so-and-so]All lines are busy right now. Sorry. I'm the virtual assistant here. [the name of the business]. how can i help you?'.”
Another big benefit of voice AI, Astavazian said, is that it knows if a call needs to be transferred to a human agent — essentially using AI to filter incoming calls based on their complexity, further amplifying the benefits by handling simple requests through automation, freeing up human agents to spend their time on more complex customer inquiries.
He stresses that the goal is not to replace human jobs, suggesting that AI can help small businesses become more productive and efficient with limited resources.
To that end, Synthflow's system is designed to not only help customers onboard voice agents, but also handle post-call data entry tasks (such as adding events to calendar tools), so building integrations with third-party software is another big focus for the team.
“This is what AI is good at,” he argues. “It can make decisions from this information.” [extracted from a call] And let's say you update a specific field in a specific CRM, and when you do this at scale across hundreds or thousands of calls, all of a sudden you see this technical advantage that we saw. [when businesses first adopted] “computer”
For its voice agent, the startup is building on OpenAI's GPT LLM but also incorporating its own AI models, which Astavazian said are trained on its own data and fine-tuned for specific customer use cases.
He said the company has also built its own “voice orchestration layer” that converts customer speech to text and sends it to an AI model as prompts, which the system then replies with text-to-speech automated responses that the customer hears as a synthesized voice on the other end of the phone line.
For now, Synthflow is focused on using AI for inbound calls, which Astabatsyan suggests is the easiest avenue of automation for resource-limited companies, but he suggests the large seed round will spur research and development to develop more advanced features.
One thing he mentioned they're working on is enabling Synthflow's voice AI to do what they call “live action” or “connect,” meaning the AI could check live inventory in a warehouse while you're on the phone, or, in his words, grab other requested information and “push it elsewhere.”
He also envisions scenarios where task-focused AI voice systems could collaborate to expand their usefulness: These systems could hand off calls to other dedicated voice AIs trained for different tasks that customers request.
“The key here is to focus on who your customer is, because your products will be very different depending on who you're making them for,” he added.
One impact to consider is that if voice AI and voice assistant systems live up to the productivity-boosting hype and expertly deliver on their promise of efficiently handling all tiers of customer enquiries, including cleverly redirecting the more complex content to the right system or human, the average small business may find themselves with far more work than they can handle.
“I think that's an interesting question for a lot of managers and leaders to think about,” he said, discussing this scenario. “Like, if you have all this capacity and all this productivity unlocked, how do you redirect this human resource to other parts of the economy? I don't think that's a question that's been answered yet, but it's certainly a very interesting question.”
Synthflow's seed funding was led by Singular, with participation from existing investor Atlantic Labs and a number of other investors in the AI space, including the founders of Krisp AI.