Startups are hitting a road block when selling their AI to big companies. Investors and founders explain how to get around it.
- As companies clamor to buy AI products, some startups are running into a sales barrier.
- Startups can struggle with procurement teams, which buy services for the company.
- Issues include data concerns, cloud providers, and education on AI basics.
After a decade of working at Microsoft and Qualcomm, Samir Kumar understands how difficult it can be to sell to large corporations.
In November, Kumar left Microsoft’s venture arm to co-found Touring Capital with former colleagues. He now advises Series B and C companies on how to sell to companies like his former employer as an investor in AI-powered software-as-a-service startups.
“People are seeing the potential in [AI] use cases, but that doesn’t mean we’ve made a ton of progress in the last few years in terms of getting it inside a company,” says Kumar.
Companies are eager to join the AI hype train, but the startups pitching them may become stuck at the bottom of the sales funnel at large tech firms. The procurement team, which is in charge of sourcing and pricing the services that the company purchases, can present significant challenges to enterprise startups. Sellers must negotiate a minefield of data concerns, rising price sensitivities, and other issues with procurement teams that may be unfamiliar with AI.
However, the right preparation – and extra patience – can speed up the process, according to two VCs and two founders.
The top concerns are data and cloud commitments.
Startups looking for inspiration for difficult procurement negotiations can look to an unlikely source: Barack Obama.
Despite security concerns, the 44th president, known for his BlackBerry devotion, publicly fought the low-tech White House to keep his beloved device.
Kumar emphasized the importance of C-suiters (or the president of the United States) influencing more cautious IT and procurement teams in the Great BlackBerry Battle of 2009.
“Find those teams and stakeholders who are your power users and internal advocates, and align that with who’s holding the budget purse strings,” Kumar went on to say. He believes that R&D teams can be the best advocates in regulated industries such as finance and healthcare.
Negotiations, like Obama’s BlackBerry battle, take time. According to a September survey conducted by New York-based SAAS financing startup Capchase, enterprise software deals are taking longer to close than they did last year.
The slowdown, according to Capchase founder Miguel Fernandez Larrea, stems from the negotiation stage, where procurement teams are more focused than ever on the value software provides. He claims that AI products, in particular, have lower retention than last year.
“It’s easier to launch a competitive product, especially when open-source AI models are used.” That means, first, it’s more difficult to demonstrate return on investment, and second, it’s more difficult to demonstrate differentiation,” Fernandez Larrea explained.
Customers’ cloud computing contracts are another source of procurement concern. Kumar advised cloud-based startups to consider their potential customers’ contracts with AWS, Azure, and other similar companies. Customers want to know if products have been approved by their cloud providers and are integrated into billing so that they can account for how much cloud computing the enterprise product will consume.
“It’s not a deal breaker, but it’s a hurdle they run into after the proof of value stage,” Kumar went on to say.
‘People just want to understand it better’
Sheila Gulati of Tola Capital, another Microsoft veteran, maps out sales processes before investing in a software startup. The resulting framework assists her portfolio companies in identifying the various parties involved (such as the buyer, procurement, and end user) and attempting to work with them all at the same time.
In some cases, procurement teams are using or evaluating AI-enabled products for their own use, so they are well-versed in enterprise data and customer and employee experiences. Other teams may take longer to get up to speed and may require education from software vendors.
Gulati stated that procurement teams are now focusing on data concerns related to AI, such as privacy, explainability, regulations, and other factors.
“Enterprise use of nondeterministic systems is pretty new,” she went on to say. “That’s leading procurement teams to ask sellers ‘what works here, tell me where I could have unanticipated outcomes?'”
Rachael Nemeth, cofounder and CEO of the AI-powered training platform Opus, stated that she and her cofounders prepare for sales meetings by creating FAQs and other teaching resources.
Insider previously reported that her startup has signed contracts with a diverse clientele, including Vanderbilt University and restaurant chain Wagamama.
Nemeth stated that teams in procurement are not yet asking about large language models or other AI specifics. Instead, she answers questions about AI fundamentals. According to Nemeth, procurement teams want to be educated and assist in the completion of deals.
“A lot of questions are around functionality and who it’s touching, but the level of sophistication of those questions has not reached its end,” she went on to say. “It’s not a case of ‘we’re skeptical of AI.’ People simply want to better understand it. It’s a good thing and a good sign for SAAS firms.”
