It’s no secret that Claude has emerged as a competitor to just about everybody.

For a while a customer’s threat to replace software with Claude seemed pretty empty — unless perhaps that customer was a micro-business. Now it’s much more doable through Claude Code and Cowork, MCP connectors, skills, and the Agent SDK.

But Anthropic isn’t content to stop there. They keep going deeper into apps with Claude Design, Claude Science, the Claude Legal plugin, Claude for Teachers, Claude for Small Business, and Claude Tag. All of these apps were released within the past few months!

“Can’t we just build this with Claude?” is a real question for software buyers. I won’t pretend to know who has the biggest moat against Claude or how durable those moats will be. But I did want to investigate how top apps are positioning against Claude.

Up today: what 50 of the best SaaS and AI-native apps are doing to convince prospects to buy rather than DIY.

This research would’ve taken me half a day to do manually. Doing it with Claude Cowork took less than 15 minutes. While the data was compiled with AI, I independently reviewed the materials and personally wrote the insights.

Here’s how I did it — feel free to copy this approach for your own GTM research:

If the research turned up nothing, that means Claude couldn’t find it, not that it doesn’t exist. I believe this finding is still compelling since it means the company probably isn’t influencing the AI search results on build-versus-buy.

The data was gathered in July 2026. Some of it could already be stale. You can see all the companies and assets in this Google Sheet.

Buyers might be comparing products against Claude, but vendors are hesitant to frame this comparison directly. Based on the marketing materials, few seem particularly worried about build-versus-buy as a competitive threat.

34 of the 50 have shipped MCP servers or connectors that make their product available inside LLMs like Claude. The bet: use Claude with these products, not as a replacement for them. But all of a sudden Claude is the main UI for your own product.

18 of the 50 have published nothing at all that could be picked up by Claude related to Claude and/or build-versus-buy. These aren’t isolated to certain product categories and include surprises like Cursor, Replit, and Wix where there’s a significant amount of third-party content framing the comparisons.

Only 12 of the 50 directly name Claude on their own domain. 7 of these are adversarial while 5 are framed as partnerships. Even fewer — just 4 — go a level deeper and publish DIY cost comparisons that frame the total cost of build versus buy.

Why this matters: When you don’t write the comparison, someone else does. That’s usually a dubious third-party with a referral link and no reason to be fair to you.

Just 7 of the 50 companies — Lovable, Framer, Glean, Dust, UiPath, GitLab, and GC AI — had a page on their site that directly compared their product to Claude.

Instead of framing why products are better than Claude, the more common approach was to frame products as complementary. Usually it’s a flavor of start with Claude, then switch when you’re serious.

Lovable, the vibe coding app ($6.6B valuation), starts off their comparison page highlighting that “Lovable runs on Claude” — essentially saying that if you like Claude, you can access it through Lovable.

They tell would-be users to choose Lovable versus Claude “based on what you’re trying to walk away with—a conversation or a product.” Lovable says a common workflow is to plan with Claude, then move to Lovable to build and ship a working app.

This framing is risky from a long-term perspective as Claude invests in the application layer. But it’s persuasive right now given how the average person uses Claude.

Zapier, the workflow automation company ($5B valuation), takes a similar approach, albeit with OpenAI standing in for Claude. Their key framing: Zapier makes OpenAI work better.

Zapier’s argument is that OpenAI’s Agent Builder is fine if all you need is a native connection to a handful of other tools. Zapier’s MCP expands that to an ecosystem of 9,000 connected apps, and it allows for capabilities like triggers and scheduling.

Clay, the GTM infrastructure company ($5B valuation), has a more subtle approach. Clay highlights Anthropic as a customer and features the customer story prominently on their website. (Fin does this, too.)

The case study shows off four specific Clay use cases along with tangible impact numbers. But it’s not all about Clay: the piece makes a prominent mention about how Anthropic also uses Claude within Clay and why that’s better than Claude on its own. The specific framing: this helps Anthropic customize industry tags programmatically and without active maintenance.

Why this matters: Complementary positioning helps ride the Claude wave and attract top-of-funnel. I worry that it could leave you vulnerable if (when?) Claude goes after your category.

For those who want to compete head-to-head against Claude, how should you frame the narrative?

The first starting point is domain expertise. GC AI, the legal AI startup ($555M valuation), makes the case that “Claude Cowork is a powerful engine” while “GC AI is the car built for in-house counsel.” The key evidence: Claude Cowork doesn’t have the legal-specific compliance needed for this use case.

I like this as a starting point, but I wonder how durable this argument will be. As Anthropic goes deeper into industry verticals like Claude Science and hires AEs with industry experience (as they’re doing), I suspect companies will quickly need other winning arguments.

Glean, the enterprise AI platform ($7.2B valuation), makes the case against Claude most aggressively and highlights better context, better token efficiency, and enterprise-grade compliance. Glean isn’t claiming to have better AI; they’re claiming that they have a better context layer around the AI.

Glean brings real data to the comparison based on workflow evaluations across 175 queries. It reminded me of the Coke versus Pepsi challenge, only brought into 2026.

I found Glean’s comparison page to be compelling because it didn’t shy away from the technical details. They were specific about the handling of sensitive data for unauthorized users, audit logs, compliance APIs, and single-tenant deployment.

Technical depth used to be overkill for marketing collateral. Now it positions Glean as the expert who can teach prospects what to look out for when buying enterprise-grade AI.

Why this matters: You need to find the right balance between focusing on what buyers care about today but also what they’ll care about tomorrow. As sentiment shifts from tokenmaxxing to getting AI costs in check, I suspect buyers will give purpose-built apps another look.

You could be fantastic at explaining what makes you better than Claude. There’s just one problem: the prospect probably doesn’t care. Those differences don’t matter unless they’re part of the evaluation criteria.

The best companies are teaching prospects how to compare AI products head-to-head, welcoming comparisons against Claude (or other tools) because they know their products will come out ahead.

Fin, the customer support agent I’ve profiled previously (to be acquired by Salesforce for $3.6B), tells prospects exactly how to evaluate products including 20 questions to ask during an evaluation. Fin proposes (a) how to evaluate products, (b) which questions to ask, (c) which metrics to use, and (d) what ‘good’ looks like across those metrics.

The questions are genuinely useful. But some are clearly positioned to work in Fin’s favor. A few examples:

Harvey, another legal AI product ($11B valuation), builds on this by owning the benchmark evaluation of LLMs for complex legal tasks. Harvey calls this BigLaw Bench and introduced it in August 2024. These evaluations single out Harvey alongside the top foundational model companies (OpenAI, Anthropic, Google) and don’t reference other legal AI competitors.

Harvey continues to update the evaluations and evaluation criteria with newer releases. They control the evaluation methodology, which means they’re seen as the experts in AI for legal.

Why this matters: Before doubling down on why you’re better than Claude, teach buyers why they should care about those differences. If you write the RFP, you influence which product gets selected.

I turned this positioning research into a simple decision tree you can use right now.

You can’t ignore the build-versus-buy threat. But you might just be able to come out ahead.

Related resources: