Need for Speed in AI Sales

AI Doesn’t Just Change What You Sell. It Also Changes How You Sell It.

For years, enterprise sales has followed a well-defined playbook: educate the market, run a pilot, prove a business case, negotiate security and procurement, and then go live. With AI, the pace and needs of that classic sales process have changed. Today’s buyers of AI products come in eager to try and know more, demand more proof of ROI earlier, expect faster deployments, and have a higher bar on compliance requirements.

From conversations with a number of sales leaders across AI companies, we’ve distilled a few key takeaways on how leading GTM teams are navigating these shifting selling conditions.

The New Rhythm of the Enterprise Sale

First, let’s walk through what’s changed.

More buyers are in the market than ever before. As we all know, enterprises have carved out “AI budgets,” and board and leadership teams have pushed their companies to adopt AI at scale as soon as possible. Before SaaS replacement cycles were 5-7 years, at best. Now, everyone is in market looking at AI tools.

Buyers are self-educating. We are seeing stronger inbound interest: buyers are conducting their own product research and engaging vendors directly instead of waiting to be approached on their most important use cases. Buyers are no longer canvassing the entire market for all the potential products, with an in-depth comparison in an RFP. Instead, they may only consider one or two products, or scan a broad category of AI tools (horizontal and vertical), and then decide. There’s also a desire to do more on a self-serve basis than ever before.

Buyers want to try AI now, not next quarter. In a recent a16z Enterprise survey, 70 percent of buyers reported that speed of deployment was a top factor when engaging AI vendors. In this environment, long sales and implementation timelines are not acceptable. Many enterprise buyers are backlogged on implementation. This is amplified by buyer FOMO. When peers are making AI announcements and earning promotion-worthy wins from adoption, buyers feel pressure to capture similar results at their organization.

More startups than ever are doing the same thing. Now that it’s so much faster to build a product with AI, differentiation based on “what you do” has become thinner. Competition is higher across categories than ever before and new entrants can catch up more quickly on product. Instead, differentiation on “how quickly and reliably you do it” matters more.

Trust has new needs. Moving fast only works if the buyer believes you won’t blow up their controls. Trust used to mean “we log what humans do on our software.” Trust now means “we do the work correctly and can show how.” Many enterprises have even introduced new or increased compliance checks for AI tools, including red-teaming to proactively surface compliance and security vulnerabilities. If your product “presses buttons” on the customer’s behalf, you should expect this to become the norm.

Enterprise buyers only want to test AI once (for now). Many true enterprise buyers are not returning to market looking for another AI product if the first product didn’t perform up to their standards. This is especially true given that implementation timelines can be long. While this varies by company and industry, we’ve heard repeatedly about CIOs and enterprise buyers not wanting to commit resources twice to try a second product that claims to work.

Closing the sale: field notes from AI first teams

Special thanks to Wouter van den Brande (Decagon), Alex Lindahl (Clay), and Joe Morrissey (GP, a16z) for sharing their insights.

(1) Building a brand matters.

In this environment, brand matters more than ever: customers flock to the first players to reach scale and gain market recognition, a pattern already evident in end markets like legal and customer support. Just as “no one ever got fired for hiring McKinsey or IBM,” the same is true today for choosing well-known AI tools. Therefore being strategic and having well-known logos early on and investing in building a community earlier around the power users and often their enablers (e.g., advisors, fractional CXOs) matter.

(2) Demos do the work.

Demos have become more important when selling AI products. Showing how the work is actually done is far more compelling than a repurposed stock video that doesn’t reflect the real product in production.

The most compelling AI demos create “lightbulb” moments, showing in real time how software can handle complex, human tasks and therefore create immediate efficiency gain through the replacement of labor with software. Voice demos have done particularly well here because they replicate human work (e.g. call centers) in a believable way, and in many cases do better. Voice agents are infinitely patient, can speak in any language, and know the answer to any question. Listen to Happy Robot’s freight negotiation agent yourself!

In other areas of enterprise software, like procurement, askLio’s procurement agent demonstrates how generating a purchase request (a process that can take weeks in some enterprises) can be collapsed into just five minutes. Instead of filling out forms and emailing colleagues, users simply chat with a natural-language assistant that collects all required details and automatically generates a structured purchase request. Each request is then assigned a sourcing agent to find competitive offers, a negotiation agent to optimize terms, and a compliance agent to ensure the order stays within company policies. It resonates because no one, not even procurement teams, enjoys logging into SAP to enter data manually and getting stuck in endless email ping-pong with their teammates. Hours of productive work are lost here.

Demos are also taking over the traditional Proof of Concept (POC) role. The moment when buyers see tangible value has shifted much earlier, to the demo. The best teams prove their products work from the very first sales meeting by running demos in sandboxes that mirror production, seeded with real (or synthetically generated) customer data vs. waiting for the traditional POC timeline for that to happen.

(3) A new set of security standards needs to be met from the get-go.

With AI, buyers want continuous assurance, not just a certificate in a data room. When AI is performing actions core to business workflows, it’s unlikely an enterprise is going to take a bet on you without compliance in place.

The strongest teams come prepared with clear answers to the most common questions: Do you train on my data? How are prompts and outputs logged? What are the safeguards against hallucinations?

Some AI startups are going even further: new AI agent security standards, like AIUC, are offering enterprise testing, certification, and insurance to provide peace of mind to buyers from liability and AI hallucinations. Others, like truthsystems, are building guardrails for enterprises to monitor and flag non-compliant AI usage in real time across vendors.

(4) Deliver outcomes immediately and charge for it.

Enterprise sales has always been about starting with a narrow wedge. However, the time to prove value on that wedge has collapsed. The data supports this: 57 percent of buyers expect to see positive ROI within three months of purchase; 11 percent expect to see positive ROI immediately after purchase.

Winning companies start with a workflow where outcomes can be clearly measured. This also opens the door to charging for your service based on those outcomes rather than just monthly SaaS fees, aligning pricing with the customer’s goals. If a buyer is hesitant to commit to a large contract upfront, outcome-based pricing is a great way to land the deal.

In customer support, for example, Decagon focuses on improving CSAT (customer satisfaction) and deflection rates (the percentage of support tickets resolved by AI instead of human agents) on a week-over-week basis throughout any pilot. Customers can then opt for an outcome-based pricing model tied to the number of support tickets resolved.

The GTM Team is Changing.

As companies experiment with the right formula across sales, implementation, and customer success, GTM teams are evolving.

To start, the sales role has become more technical. Instead of understanding the people using your software, account executives (AEs) need to understand the technical details of how agents will actually do the work of humans. Sometimes this means hiring a more technical representative; in other cases, it means collapsing the AE and Solutions Engineer roles in order to create a better customer experience.

For implementation, we’ve previously discussed forward deployed engineers (FDEs) and the services needed to onboard an AI agent to a customer. FDEs help integrate and tune agents to a customer’s specific context, helping that customer realize ROI quickly.

Finally, success is also changing. Someone needs to ensure the AI deployment succeeds. This is a different skill set from traditional customer success, where the focus was on answering tickets or managing renewal cycles. AI-native success teams work directly with customers to design prompt logic and data integrations while passing feedback to GTM and product teams. Decagon calls this an Agent PM; Harvey calls it a Solutions Architect. Oftentimes, this person is an ex-consultant or PM with just a few years of experience, but with a strong presence in front of customers. As products become more self-serve with limited sales team interaction, this role becomes especially critical in the post-sales process as a core point of customer interaction. As business models increasingly align to outcomes or usage, AI companies must ensure customers are actively using their products. Selling a seat and checking back a year later no longer suffices.

The roles and scope of the team will change more at some companies than others. The change depends on the complexity of implementation and the type of support needed. Either way, reps need to know how to use the latest AI tooling to accelerate their own work. Sales teams are spending less of their time doing data entry and account research, especially with tools like Clay and Unity, outbound SDRs and meeting schedulers like 11x, and notetakers like Gong and Granola. Reps can now spend more time with the customer, understanding their specific needs deeply.

Where This All Lands

As AI products evolve, we expect to see the sales motion continue to evolve with it. Winning startups will build products that actually do the work, wrap them in governance that makes risk teams comfortable, and engineer their sales process to generate undeniable evidence quickly. In that world, the best go-to-market teams look a lot like the best product teams: fast, technical, and obsessed with customers. When new AI models launch, these teams don’t just update their products, they update their GTM processes (and quickly)!

---

This newsletter is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. Furthermore, this content is not investment advice, nor is it intended for use by any investors or prospective investors in any a16z funds. This newsletter may link to other websites or contain other information obtained from third-party sources - a16z has not independently verified nor makes any representations about the current or enduring accuracy of such information. If this content includes third-party advertisements, a16z has not reviewed such advertisements and does not endorse any advertising content or related companies contained therein. Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z; visit https://a16z.com/investment-list/ for a full list of investments. Other important information can be found at a16z.com/disclosures. You’re receiving this newsletter since you opted in earlier; if you would like to opt out of future newsletters you may unsubscribe immediately.