If you argued a startup’s go-to-market (GTM) playbook is a commodity, I’d be hard-pressed to disagree. You likely know the gist: a founding team builds a minimal viable product (MVP), launches it, iterates ad nauseam until they find product-market fit, and then frantically aims to scale it as fast as possible.
Though few startups’ success stories follow the same linear path, this general playbook will be exceedingly familiar. Try, try, iterate, try, and hopefully, you’ll succeed. But AI startups don’t have this same luxury of time.
The exorbitant costs of building AI software—from renting GPUs, talent shortages, sky-high salaries, and API costs—stress the need to achieve positive momentum or risk running out of funds. Meanwhile, given the abundance of alternative competing options, early adopters’ patience wanes quickly. Combatting these pressures, founders constantly have to showcase their product’s differentiated improvements, and even that may not be enough. To breakthrough, AI startups have to build their brand loudly.
While the general framework of GTM strategy remains similar to the enterprise SaaS era, the increased time pressure is pushing many AI startup founders to rethink their approach.
Here are six major ways GTM is shifting in the age of AI:
1. A Defensible Brand is Paramount
Software as a service (SaaS) startups never set sail amid such a deluge of competition. The low barrier to building an MVP AI application is enabling hundreds of startups to launch weekly, often with similar value propositions. Many customers struggle to understand the differences between AI models and approaches. Their uncertainty can quickly shift to doubt or even fear. AI startups require a defensible brand as a core moat to differentiate. These brands must be steadfast in their convictions.
For founders, this means having answers to essential questions early on, such as:
- What does your brand stand for?
- Are you selling to the enterprise? What is your ICP?
- Are you considered to be the most secure?
- Are you the least expensive?
These are some of the questions leadership needs to consider before publicly launching the company. The answers should differentiate the company and be firmly rooted in the company’s values, which originate from the founders. Once brand attributes are defined, the startup needs to unflinchingly own them.
Consider how Runway strategically aligns its product marketing and brand marketing. In a video describing one of its models, while the narrative strategically builds excitement for the product, it also purposefully shows customers how Runway continuously improves the model. This demonstrates the brand’s commitment to constantly shipping and expanding the product’s capabilities and defines how the brand, Runway, stands to unleash a new universe of creatives through ongoing incremental improvement.
Loudly defining a brand sometimes means founders need to address thorny topics willingly. As Hollywood screenwriters picketed against AI, fearing it would replace them, Runway’s CEO and co-founder Cristóbal Valenzuela openly addressed the concerns, sharing his belief that the generative suite could enhance creativity, not take away from it.
2. Positioning Against Skepticism
Early in the adoption of SaaS, there was limited interest in the technology. No one believed the cloud would securely store Fortune 500 companies’ data and reshape the market. To garner attention, SaaS startups created controversy, “XX will bring the end to YY.” But AI startups are born controversial, and everyone accepts its adoption is inevitable. Instead, AI founders face the unique challenge of combatting fear and misunderstandings of the technology’s application.
Consider a recent Wall Street Journal article, “AI Startups Have Tons of Cash, but Not Enough Data. That’s a Problem,” which spotlighted CIOs’ reluctance to input their company’s proprietary data into AI solutions out of concern about its usage, no matter the benefits. Meanwhile, media headlines echo the public’s fear of AI taking jobs or reaping unexpected negative societal impacts. Proper positioning is paramount to educate buyers, dismay concerns, and build a bold brand.
Many buyers assume new AI capabilities are broadly available and simple to learn. “There’s now this preliminary layer for a lot of larger companies, where they review for safety and risk. This wasn’t really the case for selling B2B SaaS,” said Greg Garte, Head of Revenue at Runway, during an interview.
Education is one facet of positioning; another is differentiation. Take Armada. Positioning itself as a full-stack infrastructure and AI company, something that’s extremely difficult to accomplish for any company—might have been enough for Armada to stand out, but the company went further. While some Valley companies coyly work with the Department of Defense, at its launch, Armada published a manifesto unapologetically stating its intentions to support the United States in the innovation race against China.
Remember, people are drawn to movements, not tools. You have to position your company relative to the current wave of innovation and meet your customers' aspirations while addressing any of their misconceptions.
3. Product-Led Growth is a Must
Anytime you’re introducing something new, and everything in AI is new, you’re asking people to take on additional cognitive load. This means founders must make the product incredibly easy to use and experience. Humans do not like to be confused. Regardless of whether a founder is building a consumer or enterprise AI product, customers need a way to simply understand the value of the product. PLG has always been a hot topic for SaaS companies, but very few have been able to capitalize on it. People want to see the product in action before they buy.
ChatGPT is one obvious example. Prior to its launch, explaining the value of LLM was generally abstract and less impactful: “You know those text suggestions to auto-complete your sentence?” But with ChatGPT, technologists and luddites alike were delighted by a chatbot that demonstrated the power of OpenAI’s generative AI models. The chatbot was merely the demo to entice buyers to license OpenAI’s infrastructure.
For some time now, Weights & Biases has been the platform of choice for AI developers to build and experiment with their models. But the company didn’t get there overnight. They made it simple for any AI engineer to use their platform for personal projects (for free!) and also made it easy to dive into a live notebook. Yes, enterprise demos are available, but builders want to try things out for themselves, and Weights & Biases leaned into that approach.
Understanding how to leverage PLG and getting your product into the hands of builders and creators as quickly as possible is a must for any AI company.
4. Rising with the Media Hype Cycle
While most early-stage startups struggled for media attention, ChatGPT’s launch created a hype cycle of media coverage that afforded strategic AI founders ample opportunities to build their brand awareness and momentum.
Initially, coverage was hyperbolic, as the Columbia Journalism Review pointed out. Doomsday stories like “Bing’s AI Chat: ‘I Want to Be Alive. 😈’” were met by hopeful counterpoints, “How Hospitals Are Using AI to Save Lives.” Even historical moments paused for AI. The day The New York Times ran a headline on Trump’s indictment, below the front page fold ran a story on Runway.
While the media’s willingness to cover at the pace of AI innovation has waned for more discerning reviews—the appetite remains for persuasive opinions on the relevant issues. Recently, Casey Newton and Kevin Roose hosted Anthropic CEO Dario Amodei on their popular Time’s podcast Hard Fork to discuss, amongst many topics, why Anthropic created an AI constitution. Herein lies one example of current greenspace for AI founders to thoughtfully weigh in on the public’s concerns around AI and build credibility.
There is so much nuance to AI, and well-spoken, media-conscious founders have a chance to become beacons that help the media and the public wrap their minds around what you’re doing to change the world.
5. Seats, Usage, and the Infrastructure of Pricing
Historically, customer usage of software wouldn’t impact pricing. When customers subscribed for Microsoft Office and a new feature was released, the pricing didn’t change based on the customer’s usage of the new offering. But for AI startups, this framework can bankrupt a great idea prematurely.
Once a startup has scaled to significant distribution, generative AI can serve as a loss leader to capitalize on an incumbent’s market share.
For now, the high processing costs of foundation models mean AI startups need to rethink their pricing strategy. Usage is a critical determinant of an AI startup’s growth. More data leads to new insights and, in turn, more innovation. For most foundation model companies, pricing must be dynamic based on tokens used—mirroring the underlying costs of the infrastructure models. Initially, setting low prices is a strategic approach to attract usage and exceed customer expectations.
There are exceptions to this rule. Some companies are limiting the number of queries to a certain amount of time, and others, like Runway, have monthly credit allocations included in their per-seat license fee but allow users to purchase more a la carte. These tactics ensure that compute costs don’t grow out of proportion, revenue still grows, and product usage is optimized for customer efficiency. Additionally, once a startup has scaled to significant distribution, generative AI can serve as a loss leader to capitalize on an incumbent’s market share.
Traditional seat-based pricing likely won’t be enough for the most ambitious AI startups. Tokens and GPUs are expensive—so usage will be more of a factor than traditional SaaS. You need to have agility with your packaging and be able to iterate quickly on your pricing strategy based on demand, costs, and customer feedback.
6. Your Community is Your Best Sales Rep
While community has always played a role in SaaS and open-source software, it’s reaching new heights thanks to AI. Companies like Midjourney and Runway built early prototypes (or full-on revenue-driving products) into their Discord servers. This led them to get customer feedback and validation all at once and build a loyal group of ambassadors.
And with people craving meaningful experiences that gather like-minded individuals together, we’re seeing teams actively invest in building affinity around their product and brand. Hugging Face hosted a conference that felt more like “Woodstock for AI,” while Runway’s AI Film Festival and their GEN:48 short-film competition celebrated their community’s creativity and helped forge connections.
People love to learn from each other, and the network effect of people bringing in other like-minded individuals is the most positive momentum a company can get, which is why activating your community is a crucial lever for AI companies with large ambitions.
The Playbook is Dynamic
The abilities of AI will continue to evolve at a rapid pace. With it, new possibilities and startup opportunities will arise. At the risk of sounding cliche, it’s still early. New functions will come up that will continue to reinvent the GTM strategy for unique companies.
With this evolution, there remains plenty of uncertainty. Some of these GTM trends may shift. As processing costs decrease with hardware advancements, pricing models will likely adjust.
The only constant we can expect is demand for AI will remain high, ushering in more competition and innovation.
“We’re shipping things so fast that it’s a dangerous game to sell what future thing we might build because our plans could change entirely,” said Greg. “We actually stay rooted in what we’ve built today and highlight how fast we got to this point. How fast everything improved because that’s factual. The future is more a hope and a dream.”
