AI sales rep startups are booming. So why are VCs wary?

When you actually survey venture capitalists investing in AI startups, they’ll tell you that businesses are experimenting wildly, but they’re too slow to do so. add AI solutions to ongoing business processes.

But there are some exceptions. And one of them appears to be an area known as AI sales development representatives, or AI SDRs. These use large language models (LLM) and voice technology to craft personalized informational emails and make automated calls to potential customers.

“In some markets, we’re seeing all five to 10 companies succeeding in a fairly short period of time,” Shardul Shah, partner at Index Ventures, said of the AI ​​SDR boom.

While it’s not uncommon for multiple startups to target the same problem, it’s rare to see them all grow so quickly. That’s the case for startups that automate content creation for sales teams, according to investors.

“When one examines any of (these startups) individually, it’s like, ‘wow, this is an amazing product market fit,'” Shah said. “When all 10 of them fit into an amazing product market, ‘What’s this going to be like?’ difficult to answer.”

Index has not yet invested in any of these companies, many of which are less than a year old. While the entire category is on fire and customers are using them, it’s too early to know if their growth will continue in the long term or if they’ll be scrapped once the wow factor wears off like many other AI pilot projects because they haven’t proven to be more effective than human interactions.

Small businesses love AI sales LLMs

Arjun Pillai, Founder DocketA startup that trains AI sales engineers is confident that the adoption of AI SDR is high because small and medium businesses can easily experiment with these tools. Prior to Docket, Pillai was chief information officer at sales lead generation platform ZoomInfo.

“Cold email response rates have dropped by at least 50% in the last two years,” Pillai said. “Now there’s a group of companies claiming they can increase that rate, with everyone willing to try their services.”

The most popular AI SDR startups include Regie.ai, AiSDR, Artisan and 11x.aibut current president ZoomInfo also a co-pilot it competes with other virtual sales agent startups.

Although these companies are experiencing rapid revenue growth, it is unclear whether they are actually helping businesses sell more effectively.

According to Tomasz Tunguz, founder of Theory Ventures, the publicly traded company’s chief revenue officer explained to him that while AI SDR helped generate a significant amount of leads over nine months, it did not lead to actual sales.

“So it’s not to say that AI won’t work. It means that many of us (still) do not know how to use artificial intelligence”, – Tunguz He said this at the SaaStr conference in September.

Will officials crush them?

Chris Farmer, partner and CEO of venture firm SignalFire, said he believes AI applied to sales and marketing is a huge opportunity, but that AI SDR startups without access to differentiated data could be overtaken by incumbents like Salesforce, HubSpot and ZoomInfo. The main products of those companies are the protectors of their customers’ data. So if they offer bots that allow their customers to use their data, such bots can be more effective.

Another venture capitalist who has been looking at this market but has not yet invested said his firm has looked at several AI SDR startups, all of which have reached $1 million in ARR in less than a year. The startups’ impressive growth was attractive, he said, but like Farmer, he worried that their solutions might eventually be offered as a free feature by established competitors.

Jasper, a copywriting startup that was last valued at $1.5 billion, faced speed bumps and forced out. lays off 30% of its employees ChatGPT has served as a warning to some investors since its introduction.

Investors are not surprised by the rapid adoption of AI SDRs; they just doubt adoption is sticky.

Updated: This story was originally published on August 22 and was updated on December 26 with comments from Tomasz Tunguz.