Voiceflow Added $15M in New Funding on the Back of Rapid User Growth and Generative AI

New valuation climbs to nine figures on a doubling of users in 2023

Voiceflow just added another $15 million in new funding (first reported in Voicebot.ai 😎) led by OpenView, an investment firm that calls itself, “The Expansion Stage VC.” OpenView partner Blake Bartlett highlighted the firm’s 75-person “expansion team” designed to help portfolio companies grow in a podcast appearance last year. This seems like a good match for Voiceflow, a no-code software platform for conversational experiences currently experiencing rapid user adoption.

The new funding brings the total investment to $35 million, with a current valuation of $105 million. This reflects a lot of progress since the company raised $15 million in July 2021 (not including the venture debt that was never drawn upon) and a 50% valuation climb. It also highlights how some software companies are better positioned than others to take advantage of the generative AI wave.

Recognizing the ChatGPT Moment

Braden Ream, the founder of Voiceflow, and I spoke several times in December 2022 and January 2023 about how people were reacting to ChatGPT. While many of us that live in the deep end of the conversational AI swimming pool knew a great deal about GPT-3 and generative AI models, we had not anticipated that plugging OpenAI’s InstructGPT model—which had been out for nearly a year—into a common chat interface would yield such a reaction.

We both agreed that this was going to have a big impact on the conversational AI market. On the Synthetic Media year-in-review episode of the Voicebot Podcast, recorded in late December, Ream discussed Voiceflow’s perspective on how Generative AI would be incorporated into the technology stack.

The Four-Layer Generative AI Enterprise Model

In the video above, Ream describes the December 2022 model Vocieflow adopted as a framework for generative AI enterprise adoption.

• Layer 1: Large Langauge Models (LLM) - Knowledge of the world

• Layer 2: Company knowledge bases - Knowledge of your company

• Layer 3: Dialogue manager - Conversational interaction capabilities

• Layer 4: Customer knowledge bases - Personalized user context

I’d add a fifth layer to this and specifically split out the application and user interface layers while combining Ream’s layers 2 and 4, but you can argue that these concepts are included in layer 3. Also, Ream was actually talking about the order in which the layers are likely to b adopted. It is a fair assumption that the first three capabilities will precede deep integration with customer databases because of the organizational and technical complexity that it introduces.

You may want to tweak this model further, but it was a good starting point for a company’s immediate reaction to what was shaping up to be a massive change. The more impressive aspect was that Voiceflow adopted the framework and immediately started building product features.

I have known Braden since early 2018. That was before the company acquired Invocable's (AKA Storyline) assets and morphed from a no-code voice app builder for storytelling into a collaboration hub for enterprise conversational AI design.

Voiceflow was always innovative, but it operated at a deliberate pace, timed around new product releases. It was clear in December that Ream sensed the importance of the ChatGPT Moment on Voiceflow’s business. Instead of waiting for a big new product release, the team just started building new LLM-based features, demonstrating them in LinkedIn posts, and shipping beta releases.

From PLG to LLM LG

Voiceflow has always oriented itself around a product-led growth (PLG) strategy. In the first half of this year, we saw LLM-led growth within a PLG framework. And looking at the metrics, it has significantly expanded Voiceflow’s market size and awareness. The company now claims 130,000 users. That is double where the company was at the end of 2022. It has also led to a substantial uptick in monthly usage rates.

For conversation designers and developers, Voiceflow made it easy to try out LLMs and envision how they could incorporate the technology into their own applications. Denys Linkov, Voiceflow’s head of machine learning, demonstrated some of these features at Voicebot’s Model Mania Enterprise Generative AI Showcase in April.

Converting the Moment

The next step for Voiceflow is converting all that LLM enthusiasm into sustainable revenue growth and broader market awareness. Voiceflow has the advantage of a strong core business with happy customers and revenue. Generative AI expands the use cases and benefits the company can deliver.

The downside of this market upheaval will surely be increased competition. However, Voiceflow may be in a good position to withstand the entry of new competitors.

Voicebot.ai recently worked with Unparsed Conference to survey conversation designers about the software solutions they employ. Voiceflow shared the top spot with Google’s Dialogflow regarding conversation designer familiarity with the tools. It showed a clear lead as the most likely software to be recommended to colleagues.

Also, remember the position that Voiceflow holds in the enterprise conversational AI value chain. It has become a preferred tool for design and prototyping. Those features are complemented by runtime operational capabilities. It takes no big leap of imagination to see many of its design customers start using more runtime capabilities as a low-friction approach to faster production capabilities. That could dramatically increase Voiceflow’s total serviceable market size.

I expect the new funding will help Voiceflow stay out in front of the conversation design market by investing more in new features that undergird its PLG strategy. That, in turn, should drive the conversion of LLM feature interest into revenue and increased interest in its runtime capabilities. Together, these factors are likely to make Voiceflow even more attractive to new enterprise customers as well as potential acquirers.