How a 200 person Company Competes with a $160B Giant in AI Search

There is a common fear among smaller SaaS companies that the rise of AI search will simply cement the dominance of the biggest players. The logic is simple: the giants have the most data, the most backlinks, and the most brand recognition. It feels like a game that is rigged from the start. The same pattern also shows up in structured data, where the practical question is how the signal becomes visible.

But the data tells a different story. Descript, a company with fewer than 200 employees, is currently punching well above its weight class. They are not the most strong video editing tool on the market, nor are they the most popular. Yet, when you look at AI visibility scores, they are competing head to head with behemoths like Adobe and CapCut.

This is a critical realization for any growth lead or founder. You do not need a billion dollar budget to be the answer an LLM provides. You just need a different strategy than the one used for traditional SEO. Descript has found a way to be the preferred answer for specific queries, and their approach provides a blueprint for anyone trying to maintain visibility in an AI first world.

The power of narrow niche messaging

For a long time, Descript has avoided the trap of trying to be everything to everyone. Instead, they have anchored themselves as a podcast editing tool. This specificity is exactly why they show up in AI answers. When a user asks an AI for the best software to edit a podcast, especially for someone who lacks professional video skills, Descript is often one of the first mentions.

This happens because of how LLMs handle queries. Unlike traditional search engines that surface pages based on keywords and authority, AI search often uses query fan outs. The AI generates multiple internal searches to find the most direct and relevant match for the specific intent of the user. Because Descript has spent years focusing their messaging on podcasters, they are the most direct match for that specific intent.

Their product pages and blog posts do not use generic language. They speak directly to the person who wants to edit spoken audio without the technical headache. This clarity makes it easy for an AI to categorize them and recommend them for a specific job.

Practical step: Define your narrowest viable market

If you try to be a general solution, you become a commodity. AI is less likely to recommend a generalist when a specialist exists for a specific use case. To replicate this, look at your current customer base and identify who is getting the most actual value from your product. Who are your most vocal advocates? What do they have in common in terms of their role, industry, or specific workflow?

The goal is to find the overlap between high ROI and high advocacy. Once you find that, stop describing your product as an all in one platform for everyone. Instead, anchor it to a specific job. Instead of a general AI tool, you might be the AI tool for legal researchers who need to summarize case law.

Expert Interpretation: The tradeoff here is between total addressable market and immediate visibility. By narrowing your focus, you are intentionally ignoring some segments of the market. However, the decision you must inspect is whether your current broad messaging is actually attracting anyone, or if it is just making you invisible to the AI. It is better to own a small, specific category than to be the tenth most relevant result in a massive one. A useful companion note is AI Recommendation Sets Leave Some Brands Out, because it looks at a nearby part of the same system.

Creating content that is actually useful

Helpful content is not a new concept, but its importance has shifted. Google has already moved to penalize low quality, shallow content, reducing its presence in search results by 45 percent in recent updates. In the context of AI search, depth is the primary currency. LLMs do not cite surface level summaries; they cite expert perspectives and detailed instructions.

Descript does not produce generic blog posts. They create instructional content that answers real world questions and Help Center pages that actually solve the problems users are facing. They also venture into topics that are adjacent to their product but highly relevant to their audience. For instance, if a user asks an AI about how much YouTubers earn or the cost of starting a channel, Descript's detailed, expert guides often appear as cited sources.

By providing complete answers and embedding their own video content, they prove to the AI that they are an authority in the creator space, not just a software vendor.

Practical step: Build content based on real friction

To move away from shallow content, stop guessing what your audience wants. If you cannot talk to your customers directly, talk to your sales and customer success teams. These people hear the same frustrations and questions every single day. Those sticking points are your content roadmap.

Focus on creating content that solves a specific problem from start to finish. If you are writing a guide, do not just list steps. Explain why those steps matter and what the common pitfalls are. The more unique and expert the insight, the more likely an LLM is to cite it as a primary source.

Expert Interpretation: This is a high effort, medium impact strategy. The tradeoff is time. It takes significantly longer to write one expert guide than five generic posts. The decision to inspect here is your content production pipeline. Are you optimizing for quantity to satisfy an old SEO checklist, or are you optimizing for depth to satisfy an AI's need for authority?

Leveraging visual evidence for AI models

One of the most overlooked aspects of AI visibility is that LLMs are no longer limited to text. Models like CLIP allow AI systems to interpret what is happening inside a screenshot or a video. This means that the visuals on your site are now part of your SEO strategy.

When users ask for the best CRM for a small business, AI results often include actual images of the product interface. This represents a shift in what matters. Highly polished, abstract marketing mockups are less valuable than real, in product visuals. Descript leverages this by showing the actual product in action, which allows the AI to understand the user experience and present it directly to the user.

Practical step: Replace abstractions with screenshots

Audit your key product and marketing pages. For every feature you describe, ask if a user can actually see it working. If you have an abstract diagram or a stock photo, replace it with a real screenshot of the interface.

Extend this to your blog and help center. If you mention a specific workflow, include a screen capture of that process. If you are teaching a complex task, a short video is far more valuable than a wall of text. The goal is to provide the AI with visual data that confirms your product does what you claim it does.

Expert Interpretation: This is a low effort, medium impact move. The tradeoff is aesthetic perfection versus utility. Many brands fear that raw screenshots look less professional than polished illustrations. However, the decision to inspect is whether your visuals are designed to look pretty or designed to be understood by both humans and AI.

Developing citable middle and bottom of funnel content

AI search performs exceptionally well when it can find content mapped to different levels of user awareness. Descript does not just focus on top of funnel awareness. They create detailed content for people who are already solution aware or product aware.

For example, when users search for video editing tools in ChatGPT, Descript often appears not just as a recommendation, but as a cited source. They achieve this by creating their own best of lists and comparison articles. Instead of generic lists, they break tools down by specific use cases and explain exactly who each option is best for. This makes their content highly citable because it provides the nuanced analysis that LLMs strive to deliver.

Practical step: Create comparison and how to content

To increase your visibility for product aware users, create content that directly addresses the competition. Ask your sales team which competitors are brought up most often in calls and what the common points of comparison are. Use those insights to build honest comparison pages.

For solution aware users, create how to content that naturally integrates your product. Instead of a sales pitch, show how to solve a specific problem using your tool. When an AI looks for a way to reduce background noise in a recording, it will look for a guide that explains the process and mentions the tool that makes it happen.

Expert Interpretation: This is high effort and high impact. The tradeoff is the risk of mentioning competitors on your own site. Some marketers fear this gives away traffic. But the decision to inspect is whether you would rather avoid mentioning competitors or be the one who defines the comparison in the eyes of the AI.

Establishing consensus through digital PR

Your own website is only one part of the equation. LLMs look for consensus across the web to determine if a brand is trustworthy. This means positive sentiment must exist on third party sites that the AI already trusts.

Descript builds this consensus through two main channels: digital PR on high authority sites and a creator friendly affiliate program. By getting mentioned in articles on sites that AI frequently cites, they create a web of validation. When the AI sees Descript mentioned across multiple trusted sources, it gains the confidence to recommend the tool as a top choice.

Practical step: Target the sources the AI trusts

You do not need a massive PR firm to start this. Begin by identifying which sites are currently being cited by LLMs in your category. Use tools to see which domains dominate the citations for your core keywords. Once you have a list of these trusted sources, focus your outreach efforts there.

Building these relationships takes time, but it is the only way to create the external validation that AI requires. Focus on providing value to these publishers rather than just asking for a link.

Expert Interpretation: This is high effort and high impact. The tradeoff is the lack of direct control. You cannot control how a third party describes you. The decision to inspect is whether you are relying too heavily on owned media. If the only place your product is praised is your own blog, the AI will likely view you as a biased source rather than a market leader.

The danger of ignoring community sentiment

Even with a strong strategy, there are gaps. For Descript, one of those gaps is Reddit. Reddit is one of the most cited sources in Google AI Mode because it represents real human conversation.

Currently, Reddit is a mixed bag for Descript. While there are positive mentions, they are often buried under negative sentiment in subreddits like r/podcasting. Because AI search prioritizes these community discussions to provide a balanced view, negative sentiment in these forums can act as a ceiling on your visibility, regardless of how good your own content is.

This highlights a critical point: AI visibility is not just about marketing, it is about reputation. If the community consensus is negative, the AI will reflect that.

Final takeaways for SaaS visibility

Descript is not winning because they have the most resources. They are winning because they are focused, helpful, and consistent. They have aligned their content strategy with the way LLMs actually process information.

Introduction

The key issue here is At just under 200 employees, Descript is not the biggest name in video editing software. It's not the most strong or the most popular, either. But it's punching way above its weight, competing with much bigger companies (like Adobe, and CapCut) in LLM search. My read is to treat it as a decision point: what signal needs to become clearer, what part of the system is currently weak, and what evidence would show that the work is improving visibility rather than only adding activity.

That is the difference between reacting to a trend and building a useful search system. Connect this point back to the page template, internal linking, entity signals, content depth, crawl accessibility, and the way the brand is represented across the wider web before deciding what to change first.