I made the case that AI has changed everything about sourcing enterprise customers. Enterprise buyers have AI budget and want to spend it. They discover solutions in a chat window instead of on a trade show floor. And by the time they take your first meeting, they have already done three of the four stages of the buying journey with an AI.

If an AI is the one member of every buying committee that is consulted at every stage, how does a founder at a new startup get consideration?

The short answer is, AI is trying to answer two questions on the buyer's behalf: what is the best possible answer to this query, and who is the best possible source for it? Everything below is about becoming that source. It's a survival imperative for today's startup founder.

The question AI is actually answering

A CFO does not ask an AI "what is a compliance platform." She asks, "What should a platform like this cost for 200 seats?" The security lead asks which vendors are SOC 2 Type II and support SSO. The AI then fans each of those out into a half-dozen sub-queries, searches every one, and synthesizes a single answer from whichever sources covered each sub-query best. You can reverse engineer this to be the best source for a diligent AI researching real questions your buyer needs answered.

The big idea: purchases are decided in that long tail. "Best CRM" is a query with enormous volume and brutal competition. "Best CRM for a 12-person agency migrating from HubSpot, under $50 a seat" is a query with small volume, little competition, and a buyer on the other end of it. Two buyers asking the same question with different context get different shortlists. Your content has to match the context, not the keyword.

That reframes the work of ranking in the AI era. You are not trying to rank. You are trying to be the one source that answers the buyer's specific situation so completely that the AI quotes you. In B2B software, that is unusually doable: 68.4% of AI citations in the category come from the brand's own product and company content, the highest share of any industry.[^1] The AI wants to cite you. You have to give it something to cite. Here's how.

Making sure AI can describe what you sell

Start by running your top 25 buyer questions through ChatGPT, Claude, Gemini, Perplexity and Google's AI Mode, and reading what comes back. Most founders are shocked. The AI gets your pricing wrong, names a competitor you beat in every bake-off, or describes a product you shipped two versions ago.

By the way, how do you know what questions your buyers are asking?

This is where it gets gorgeous, in my point of view--smart founders are already tracking and recording all customer conversations. If you've been doing this for a while, putting those transcripts in a single location where you can run queries on them from your Claude or internal AI makes coming up with realistic scenarios actually braindead simple. I strongly recommend running "top customer concerns / questions / interest / engagement" queries against your transcript treasure trove weekly, and standing up an AI that comes up with outstanding, detailed, research and product driven answers and videos as one of the ways you keep the long tail lengthening.

But back to the top 25 question run (which is also a scheduleable task, my friends). Fix the facts first. Everything else builds on them. That means your positioning, your ICPs and use cases, your pricing and packaging, and the answers your best rep gives on every call: comparisons, objections, ROI, security, compliance and integrations. Then make it technically readable: pages AI crawlers can access and parse, structured data, and the same facts stated consistently everywhere they appear.

The mental model is to enable AI the way you would enable a new sales rep.

Everything you would teach that rep in week one, market knowledge, customer insight, product, pricing and objection handling, gets published where a model can learn it. Then schedule a repeat cycle to make this an automated process, and track the results.

You could stop reading here, and already be winning, but I've got a couple more strong tips you'll want up next.

Answering the follow-up question for every ICP and scenario

Take those 25 questions, multiply by your ICPs (often founders aren't sure what their ICP is yet, I get it, so you allow me to allow you to work with all 2 or three of your top running candidates), and two scenarios each, and you have hundreds of answers the AI needs to be able to find.

A CTO running approval on a big new AI initiative is running a separate process than the CEO at the same size company, their competitor, running an AI augmentation initiative--and both can be your potential client. So, ungate your long-form content. A PDF behind a form is invisible to AI, which means it cannot quote it, cite it or recommend you from it. When I ran my top 10 tech PR marketing agency over a decade ago, I loved gating content--that behavior, smart then, is stupid now.

The same guide published as a long-form page on your site is fully usable. And make your sales enablement assets public: case studies, the pricing guide, comparison pages, the ROI calculator, the implementation guide, the integrations directory, the security FAQ. Every asset your reps send after a call is an answer AI could be giving before the call.

Quick aside here: be aware of where AI looks for material. It's the web, sure, but also LinkedIn, Reddit, and -- drum roll please-- YouTube. So for those real market movers and shakers among you, be sure to prompt a beautiful way to make short, potentially animated videos of highlight points in your weekly content run. This, too, can be automated with AI and scheduled as a task.

Getting others to back up your claims

With AI, reputation is currency. Before it recommends you, the model checks review sites, communities, professional social, video and press to see whether what you say about yourself holds up.

Reddit, LinkedIn, Wikipedia and YouTube alone carry 40% of citations across the major AI engines.[^2]

Two cautions. The weighting swings without warning: Reddit's share of ChatGPT citations fell from roughly 60% to 10% in a few weeks in late 2025 before settling near 28%, so never bet your visibility on one source.[^3] And the fastest-growing citation source is YouTube, up from 1.8% to 4.6% of all AI citations in ten months, while almost everything else stayed flat.[^4]

What our founders do: ask for the G2 or Capterra review at the moment of value, right after a win, a renewal or a great support interaction. Answer critical reviews and Reddit threads in public, quickly, as a person. Publish customer stories in the customer's own words on multiple platforms. And track sentiment by theme. If AI is telling buyers your product is "hard to set up," that is a product fix, a content fix and a proof fix all at once. Then re-run the same prompts and watch whether the score moves. Automate your reporting. You do not have to purchase this service; it's buildable inside your own organization with Claude and some cognitive curiosity and sweat.

Publishing beyond your product

Four signals tell AI you are the expert: relevance to a specific buyer, presence across every channel AI reads, engagement, and authority built over time. Most companies stop at content about their own product. The iconic brands in a category cover every critical pain point their buyer has, including the ones their product does not solve, with a diagnosis and a prescription for each.

Shopify's YouTube channel teaches aspiring merchants how to use Perplexity and sell on Facebook Marketplace. Jobber runs a podcast for home-service owners about running the business, not about Jobber. That is what expertise looks like to a model: a source worth trusting on the buyer's whole problem, not just the slice you sell.

1. Automate audit, gap, execute, and track

Run it like a sales pipeline, because it is one.

Audit your baseline. Run your target queries across every AI platform and score each one: cited, mentioned, competitor named instead, or nobody named. Benchmark against competitors on mention rate, citation rate and share of voice.

Find the biggest gaps. Rank by revenue impact, not search volume. The highest near-term ROI is a query where a competitor is cited and you are not. The second is where your content is cited but your product is not recommended.

Specificity, buying-stage proximity and citation-gap size all outrank search volume, which is a tiebreaker at best. Most agencies and outsourced experts still focus on "search volume"--that is yesterday's metric. Do not let the winning moves of the prior decade hobble you in this one.

• Track over time. Retire the old dashboard. Rankings become share of voice in AI answers to your ICP, nothing more, nothing less. Organic sessions are morphing into--citation frequency across AI engines.

• Click-through rate, the old metric, is becoming branded search volume after you are cited.

• Traffic as the goal modifies into revenue from AI-influenced touches.

Make your AI visibility a priority for your startup

The founders I back are allergic to marketing generally speaking--I love them, but they love the product and they love the problem and rarely does that radical combination also include loving marketing. However, they do love winning, and this is the nerdiest way to win yet. Check out the facts of today:

• Brands cited inside an AI Overview earn 35% more organic clicks and 91% more paid clicks than brands that are not, because the buyer already recognizes them when the ad appears.[^5]

• AI-cited traffic converts at a multiple of traditional organic traffic, and an AI visitor is worth several times a traditional organic visitor.[^6]

More to the point, every new channel mints new giants, and the giants of this one have not been crowned. Search gave us TripAdvisor and Zillow. The App Store gave us Uber. Meta ads gave us Warby Parker. The B2B enterprise software leaders of the next decade will be whichever companies are most visible in AI search. Latecomers pay more. Natives win big.

A closing note of thanks. This post was built on the work the four pillars of Core, Contextual, Reputation and Expertise, and the audit-gap-execute-track process, from Shiv Narayanan, founder and CEO of How To SaaS, and his 2026 deck "The AI Visibility Imperative." Shiv has spent a career building growth engines for private-equity-backed software companies, and he has now written the playbook for the AI era. If you want the full version, with the 25 questions, the roadmap and the KPIs, go to howtosaas.com and ask him for it. Tell him Valor sent you. If you need a smart human helping build this out for your startup, he's top notch.

[^1]: AirOps, AI Citations by Vertical, May to June 2026, as presented in Shiv Narayanan, "The AI Visibility Imperative," How To SaaS, 2026. Share of AI citations from brand and product content: B2B software 68.4%, HR and workforce 67.5%, legal services 60.6%, fintech 54.9%, consumer and retail 43.8%, travel 33.5%. [^2]: Semrush, domains cited by ChatGPT search, Perplexity and Google AI Mode, January 2026, as presented in Shiv Narayanan, "The AI Visibility Imperative," How To SaaS, 2026: reddit.com 11.29%, linkedin.com 11.03%, wikipedia.org 9.53%, youtube.com 8.77%. [^3]: Shiv Narayanan, "The AI Visibility Imperative," How To SaaS, 2026: Reddit's share of ChatGPT citations fell from roughly 60% in early August 2025 to 10% by mid-September 2025, then stabilized near 28%. [^4]: AirOps, AI Citation Analysis 2025 to 2026, as presented in Shiv Narayanan, "The AI Visibility Imperative," How To SaaS, 2026: YouTube's share of all AI citations rose from 1.8% in August 2025 to 4.6% in June 2026. [^5]: Shiv Narayanan, "The AI Visibility Imperative," How To SaaS, 2026: brands cited inside an AI Overview earn 35% more organic clicks and 91% more paid clicks than brands that are not. [^6]: Digital Bloom and Ahrefs, 2026, and Semrush, 2026, as presented in Shiv Narayanan, "The AI Visibility Imperative," How To SaaS, 2026: a 23x conversion premium for AI-cited traffic versus organic, and an AI visitor valued at 4.4x a traditional organic visitor.