A closer look at venture capitalists investing in AI startups reveals that while companies are experimenting, they are very slow to add AI solutions to their ongoing business processes.
However, there are some exceptions. And one of them appears to be a field known as AI Sales Development Representatives (AI SDRs). They use large-scale language models (LLM) and voice technology to create personalized outreach emails and make automated phone calls to potential customers.
Shardul Shah, partner at Index Ventures, said of the AI SDR boom, “In some markets, we're seeing five to 10 companies succeeding in a short period of time.”
It's certainly not uncommon for multiple startups to work on the same problem, but it's rare for all startups to experience rapid growth. But investors say this appears to be the case for startups that automate content creation for sales teams.
“When studying something, [these startups] Personally, I'm like, 'Wow, the product market fit is great,''' Shah said. “When all 10 companies have amazing product-market fit, it's hard to answer, 'What's going to happen?'”
The index has not yet invested in these companies, many of which are less than a year old. Even if the whole category is hot and customers are using them, it remains to be seen whether the growth will continue in the long term or whether it will be scrapped like so many other AI pilot projects once the wow factor wears off. It's still too early to know. It has not been proven to be more effective than human assistance.
Small businesses prefer AI Sales LLM
Arjun Pillai, founder of Docket, a startup that trains AI sales engineers, believes AI SDR adoption is increasing because small and medium-sized businesses can easily experiment with these tools. Prior to joining Docket, Pillai served as chief data officer at sales lead generation platform ZoomInfo.
“Over the past two years, cold email response rates have dropped by at least 50%,” says Pillai. “There are now a lot of companies that claim they can improve this rate, so everyone is willing to try their services.”
Some of the most well-known AI SDR startups include Regie.ai, AiSDR, Artisan, and 11x.ai, but incumbent ZoomInfo also released Copilot to compete with these and other virtual reseller startups.
Although these companies are experiencing rapid revenue growth, it is unclear whether they are actually helping companies improve their sales efficiency.
According to Tomasz Tunguz, founder of Theory Ventures, the chief revenue officer of a publicly traded company said that AI SDR helped him generate a significant amount of leads in nine months, but not in actual sales. He revealed that he was not connected.
“That doesn't mean AI doesn't work; that's what many of us would say. [still] I don’t know how to use AI,” Tungs said on stage at the SaaStr conference in September.
Will the incumbents crush them?
Chris Farmer, partner and CEO of venture firm SignalFire, believes applying AI to sales and marketing is a huge opportunity, but without access to differentiated data. , said AI SDR startups are at risk of being overtaken by incumbents such as Salesforce, HubSpot, and ZoomInfo. These companies' flagship products store customer data. Therefore, providing bots that allow customers to utilize their data can make such bots more effective.
Another venture capitalist who is eyeing this market but has not yet invested says that her company is looking at several AI SDR startups, all of which have an ARR of 1 million within a year. Said to have reached USD. She said the startup's impressive growth was appealing, but like Farmer, she was concerned that their solution would eventually be offered as a free feature by an established competitor.
Jasper, a copywriting startup last valued at $1.5 billion that fell into financial trouble after introducing ChatGPT and had to lay off 30% of its staff, It's a wake-up call.
Investors are not surprised by the rapid adoption of AI SDRs. They just wonder if adoption is sticky.
Update: This article was originally published on August 22nd and updated on December 26th with comment from Tomasz Tunguz.