When marketers talk about optimizing their brand for AI, they default to focusing on ChatGPT. After all, it has the large majority of the market, and your grandmother probably even knows what it is.
But what about Claude? Perplexity? Should marketers just ignore them? Or are there certain industries and customer segments where these alternatives make up a significant part of LLM usage?
To find out, I used technographic data from Bloomberry, which tracks which B2B products companies subscribe to, to analyze 44K companies (customers of ChatGPT, Claude, or Perplexity) to see who actually picks what as their enterprise LLM.
Turns out, the type of company you are – your industry and the tools you use – strongly predicts which AI you choose. Here’s what the data shows.
- Do companies use multiple AI tools, or just one?
- Which industries prefer Claude and Perplexity over ChatGPT?
- What other tools do Claude and Perplexity users use vs ChatGPT users?
- What roles are companies hiring for that mention Claude only, Perplexity only or ChatGPT only?
- Are companies optimizing for Perplexity like they do for Google?
- Is Claude or Perplexity adoption more US-centric than ChatGPT?
1. Do companies use multiple AI tools, or just one?
Let’s start with the raw numbers. Of the companies in our dataset that use ChatGPT (Business or Enterprise), Claude (Team or Enterprise), or Perplexity Enterprise:
- ChatGPT: 41,087 companies
- Claude: 3,964 companies
- Perplexity: 1,116 companies
ChatGPT has roughly 10x the enterprise customers of Claude, and 37x Perplexity. No surprise there. But the overlap patterns tell a more interesting story:
| Tool | Also use ChatGPT | Also use Claude | Also use Perplexity |
|---|---|---|---|
| ChatGPT users | – | 4.1% | 1.2% |
| Claude users | 43.0% | – | 3.8% |
| Perplexity users | 45.1% | 13.6% | – |
ChatGPT users don’t feel the need for alternatives – only 4% also have Claude, and just 1% have Perplexity. But Claude users? 43% of them also pay for ChatGPT. For Perplexity, it’s 45%.
Why do Claude and Perplexity users also keep ChatGPT around?
Because Claude and Perplexity are specialists. Claude is great at code and long documents. Perplexity is great at research with citations. But neither tries to do everything.
ChatGPT, on the other hand, is a jack of all trades. It has image generation, a plugin ecosystem, voice mode, canvas, memory, and a dozen other features that Claude and Perplexity haven’t built yet. It’s also the tool everyone already knows how to use – your marketing team, your sales team, your CEO. When you adopt Claude for your engineering team, you don’t rip out ChatGPT from everyone else. You just add Claude where it’s needed.
The asymmetry makes sense when you think about it this way: companies adopt Claude or Perplexity because they have a specific need ChatGPT isn’t meeting. But ChatGPT is still useful for everything else. So they end up with both.
So who are the true believers? The ~2,300 Claude-only companies and ~560 Perplexity-only companies who chose an alternative and stuck with it?
2. Which industries prefer Claude or Perplexity only?
To understand who these loyalists are, I looked at what industries they’re in. I compared the ~2,300 Claude-only companies, ~560 Perplexity-only companies, and ~39,000 ChatGPT-only companies by sector. Three very different profiles emerge.
Claude-only users are mostly software companies.
| Sector | Claude-only | ChatGPT-only | Difference |
|---|---|---|---|
| Software Development | 23.3% | 10.1% | 2.3x more likely to be Claude |
| Financial Services | 8.7% | 4.5% | 1.9x more likely to be Claude |
| Insurance | 2.1% | 1.0% | 2.1x more likely to be Claude |
| Security | 1.9% | 0.9% | 2.1x more likely to be Claude |
| Hospitals & Health Care | 2.6% | 1.6% | 1.6x more likely to be Claude |
| Biotechnology Research | 1.2% | 0.9% | 1.3x more likely to be Claude |
Almost a quarter of Claude-only companies are in Software Development (which includes Taboola, Anaplan, and Redfin) – more than double the rate of ChatGPT-only. This isn’t that surprising. Claude has invested heavily in coding capabilities with Claude Code surpassing OpenAI Codex in almost every metric.
But it’s not just benchmarks. Anthropic has built an entire developer ecosystem: Claude Code for autonomous coding, the Claude Agent SDK, native integrations with VS Code and JetBrains, and support for 30+ hour autonomous coding sessions. If you’re building software, Claude has become the tool optimized for your workflow.
The financial services, insurance, and healthcare pattern requires a different explanation. These are regulated industries where security and compliance aren’t optional – they’re legally mandatory.
And Anthropic has built specifically for this: SOC 2 Type II certification for security, availability, and confidentiality; HIPAA Business Associate Agreements for healthcare customers; Zero Data Retention (ZDR) endpoints that ensure no prompts or outputs are stored; and private deployment options through AWS Bedrock and Google Vertex AI with VPC isolation.
Perplexity-only users are mostly investors, lawyers, and journalists (Examples: Bayer, HP, SAP, Visa, the US Patent and Trademark Office, O’Melveny & Myers (law firm), Clarivate)
| Sector | Perplexity-only | ChatGPT-only | Difference |
|---|---|---|---|
| Newspaper Publishing | 1.6% | 0.2% | 8x more likely to be Perplexity |
| Government Administration | 2.3% | 0.5% | 4.6x more likely to be Perplexity |
| Biotechnology Research | 3.0% | 0.9% | 3.3x more likely to be Perplexity |
| Law Practice | 2.3% | 0.7% | 3.3x more likely to be Perplexity |
| VC & Private Equity | 2.8% | 1.0% | 2.8x more likely to be Perplexity |
| Investment Management | 2.0% | 1.1% | 1.8x more likely to be Perplexity |
| Market Research | 1.1% | 0.3% | 3.7x more likely to be Perplexity |
The common thread here? Jobs where you need to find information, synthesize it, and cite your sources. Perplexity was built as an “answer engine” – fundamentally different from ChatGPT’s chatbot approach. Every response includes inline citations with links to sources. You can verify where information came from.
For a journalist fact-checking a story, a lawyer researching case law, or an investor doing due diligence, that’s not a nice-to-have – it’s the core requirement.
The use cases are documented. Gunderson Dettmer, a leading law firm serving high-growth companies and VC funds, rolled out Perplexity Enterprise firmwide. Within months, 80% of the firm’s lawyers were actively using it, processing over 35,000 queries per month.
For investors, IVP (Institutional Venture Partners) implemented Perplexity Enterprise across 62+ employees and saw 100% adoption. Their General Partner noted: “Perplexity Enterprise Pro helps the entire team at IVP accelerate investor productivity and our ability to rapidly evaluate new market segments. As a result, our diligence is significantly faster.”
The use cases include competitor fund analysis, researching active deals, and quickly getting up to speed on new market segments – all tasks where you need to synthesize information from multiple sources and know where it came from.
ChatGPT-only companies is everyone else (Examples: Carnival, Warner Bros Discovery, Reuters, KKR, Harbor Freight, and Massimo Dutti)
| Sector | ChatGPT-only | Claude-only | Difference |
|---|---|---|---|
| Marketing Services | 1.8% | 0.5% | 3.6x more likely to be ChatGPT |
| Real Estate | 2.0% | 0.9% | 2.2x more likely to be ChatGPT |
| Advertising Services | 4.9% | 2.4% | 2x more likely to be ChatGPT |
| Business Consulting | 3.7% | 1.9% | 1.9x more likely to be ChatGPT |
| Professional Training & Coaching | 0.8% | 0.4% | 2x more likely to be ChatGPT |
| Staffing & Recruiting | 0.7% | 0.4% | 1.8x more likely to be ChatGPT |
Marketing, advertising, consulting, real estate, recruiting – these are industries where the tool you use is less important than using something. They’re not building AI into products. They’re not doing research that requires citations. They’re writing emails, generating content ideas, drafting proposals. For these use cases, ChatGPT is good enough.
The implication for marketers: if your target audience is software developers, Claude might be their primary tool. If you’re targeting investors, lawyers, or journalists, check Perplexity. But if you’re going after marketers, agencies, or consultants? ChatGPT is almost certainly their default.
3. What other tools do Claude and Perplexity users use?
Okay, but “software companies” and “investors and lawyers” are broad. What kind of companies specifically? To get more granular, I looked at what other tools these companies are actually paying for. Not what they say they do – what software they use.
Claude-only companies are all-in on dev and design tools.
| Tool | Claude-only | ChatGPT-only | Difference |
|---|---|---|---|
| Cursor | 13.4% | 4.6% | 2.9x more likely to be Claude |
| Docker Hub | 17.5% | 5.5% | 3.2x more likely to be Claude |
| Figma | 8.0% | 2.6% | 3.1x more likely to be Claude |
| PagerDuty | 7.8% | 2.1% | 3.7x more likely to be Claude |
| Incident.io | 3.0% | 0.4% | 7.5x more likely to be Claude |
| Vanta | 5.0% | 1.1% | 4.5x more likely to be Claude |
The Cursor signal is the smoking gun. Cursor is an AI-powered code editor that has become the tool of choice for developers who take AI-assisted coding seriously – and it uses Claude as its default model. If 13.4% of Claude-only companies use Cursor vs just 4.6% of ChatGPT-only companies, that tells you Claude has won the hearts of the most AI-forward engineering teams.
These aren’t companies that “prefer Claude for chat.” They’ve built their entire dev workflow around Claude-powered tooling.
The rest of the stack tells the same story: Docker, GitHub Enterprise, PagerDuty, Incident.io, Linear. Engineering-led orgs with real ops infrastructure. Companies that have on-call rotations. Companies where “the site went down” is an emergency, not an inconvenience.
But it’s not just devs. Figma, Miro, Lucidchart, Dovetail (a user research tool) – design and product tools also show up. Claude-only companies aren’t code-only shops. They’re product companies with strong design functions. The kind of place where designers and engineers actually collaborate.
Then there’s the security/compliance cluster: Okta, Vanta, Kandji, OneLogin. Companies that care about SOC 2 and identity management. This lines up with the Financial Services and Insurance industries we saw earlier – and now we know why.
These companies need AI tools that meet their security requirements. Anthropic offers SOC 2 Type II certification, zero data retention options, and HIPAA compliance – features that matter when your compliance team has veto power over vendor selection.
Perplexity-only companies use research, deal flow, and security tools.
| Tool | Perplexity-only | ChatGPT-only | Difference |
|---|---|---|---|
| DealCloud | 0.9% | 0.04% | 22x more likely to be Perplexity |
| Luminance | 2.1% | 0.03% | 70x more likely to be Perplexity |
| Wiz | 3.4% | 0.3% | 11x more likely to be Perplexity |
| Zscaler | 7.6% | 1.5% | 5x more likely to be Perplexity |
| Box | 4.1% | 0.6% | 6.8x more likely to be Perplexity |
This is where it gets interesting. DealCloud is PE/VC deal management software – the tool investors use to track their pipeline and portfolio. Luminance is AI-powered legal research and contract analysis.
These tools don’t show up at all in the Claude or ChatGPT profiles. They’re Perplexity-specific. This confirms the industry data: Perplexity users aren’t just “people who like citations.” They’re professionals in research-heavy fields who have specific workflows around finding, verifying, and synthesizing information.
Box Enterprise is also telling. Not Google Drive, not Dropbox – Box. The enterprise document management platform that banks and law firms use because of its security and compliance features.
The security tools are heavy too: Wiz, Zscaler, Netskope, Sailpoint. Perplexity-only companies aren’t scrappy startups moving fast and breaking things. They’re established enterprises with serious security requirements.
ChatGPT-only companies? No strong pattern. The most over-represented tools are Google Tag Manager, Microsoft365, Yoast, Intune, Cloudflare. Generic enterprise stack. Marketing tools. The kind of software every company uses. There’s no signal because there’s no specific profile. ChatGPT is the default choice for companies that don’t have strong opinions about their AI tools.
They’re not wrong to use it – ChatGPT is good! But they’re not optimizing for anything specific.
That’s the real difference. Claude-only and Perplexity-only companies have distinct, opinionated tech stacks that tell you exactly what they do. ChatGPT-only companies look like… everyone.
4. What roles are companies hiring for that mention Claude only, Perplexity only or ChatGPT only?
To pressure test this further, I looked at a completely different dataset: job postings. I searched for jobs that mention “Claude” and “AI” in the description (but not ChatGPT, Perplexity, or Gemini), and did the same for each tool. Then I looked at which job titles are statistically over-represented for each. The results are unambiguous.
Claude job postings are almost entirely engineering roles.
| Job Title | % of Claude jobs | % of all jobs | Difference |
|---|---|---|---|
| Machine Learning Engineer | 17.8% | 0.04% | 445x more likely |
| Backend Engineer | 26.1% | 0.9% | 29x more likely |
| Research/Applied Scientist | 2.5% | 0.02% | 125x more likely |
| Data Scientist | 3.7% | 0.07% | 53x more likely |
| Frontend Engineer | 4.1% | 0.12% | 34x more likely |
| Product Designer | 1.5% | 0.06% | 25x more likely |
Almost 1 in 5 job postings that mention Claude are for Machine Learning Engineers. Add in Backend and Frontend Engineers, and nearly half of Claude job postings are for people who write code. Product Designer is the only non-engineering role that shows up, and it’s a weak signal. Companies mentioning Claude are hiring people to build AI products.
ChatGPT job postings are mostly marketing and sales roles.
| Job Title | % of ChatGPT jobs | % of all jobs | Difference |
|---|---|---|---|
| Performance Marketing Specialist | 1.7% | 0.01% | 170x more likely |
| Content Marketing Specialist | 2.3% | 0.04% | 58x more likely |
| Graphic Designer | 2.0% | 0.05% | 40x more likely |
| Copywriter | 0.8% | 0.02% | 40x more likely |
| Social Media Specialist | 0.8% | 0.02% | 40x more likely |
| Product Manager | 5.7% | 0.21% | 27x more likely |
Not a single engineering role in the top results. Companies that mention ChatGPT in job postings are hiring people to use AI for content and marketing – not to build with it.
Again, this isn’t random. ChatGPT’s strengths align with marketing workflows: consumer brand recognition means hiring managers use language candidates will recognize; its conversational interface is optimized for generating text like blog posts, social copy, and email drafts; and you don’t need to understand APIs or write code to use it.
But even within engineering, Claude and ChatGPT split the market.
I dug deeper into engineering roles specifically. For each role, I calculated how many job postings mention Claude vs ChatGPT (excluding jobs that mention both or other tools).
| Role | Claude mentions | ChatGPT mentions | Ratio |
|---|---|---|---|
| Machine Learning Engineer | 1,357 | 537 | 2.5x Claude |
| Backend Engineer | 1,990 | 1,416 | 1.4x Claude |
| Research/Applied Scientist | 187 | 164 | 1.1x Claude |
| Solutions Architect | 146 | 132 | 1.1x Claude |
| Mobile Engineer | 120 | 107 | 1.1x Claude |
| Frontend Engineer | 318 | 306 | Even |
| Data Scientist | 285 | 278 | Even |
| DevOps Engineer (SRE) | 200 | 232 | 1.2x ChatGPT |
| Data Engineer | 114 | 148 | 1.3x ChatGPT |
| QA Engineer | 144 | 252 | 1.8x ChatGPT |
| Data Analyst | 68 | 243 | 3.6x ChatGPT |
Claude dominates with engineers who are building AI products: ML Engineers, Backend Engineers, Research Scientists. These are the roles that are writing code that calls the Claude API, integrating LLMs into products, fine-tuning models. They care about benchmarks, API design, and developer tooling.
ChatGPT wins with engineers closer to data, testing, and operations: Data Analysts, QA Engineers, DevOps, Data Engineers. These are roles that use AI as a tool to do their job faster – writing SQL queries, generating test cases, debugging pipelines – not roles that build AI into products. They care about ease of use and don’t need to integrate AI into products they ship.
The dividing line isn’t “technical vs non-technical.” It’s “building with AI vs using AI.”
The bottom line: If a company mentions Claude in their job postings, they’re almost certainly hiring engineers to build AI products. If they mention ChatGPT, they’re either hiring marketers to use AI for content, or engineers who use AI as a productivity tool rather than a product component.
5. Are companies optimizing for Perplexity like they do for Google?
Here’s something I didn’t expect to find. When digging through Perplexity job postings, I kept seeing SEO roles pop up. Not people using Perplexity for research – but companies hiring SEO specialists to optimize their content for Perplexity as a search platform.
Some examples:
- “Lead global & local strategies optimised for Google SGE, Bing Copilot, Perplexity, and more” – Fisher & Paykel
- “Utilize AI tools (e.g., Perplexity) to conduct research and create structured content plans” – Compose.ly
- “Lead the evolution of Vio’s SEO playbook for emerging surfaces: e.g. SGE, Perplexity, zero-click results” – Vio.com
- “Experience with other forms of SEO: YouTube, Perplexity AI, etc.” – Reedsy
- “GEO Optimization: Design and implement strategies to increase discoverability through AI-driven experiences (e.g., Google SGE, Bing Copilot, Perplexity)” – ScienceLogic
To quantify this, I looked at 13,000+ SEO specialist job postings since 2024 to see which AI tools get mentioned:
| Tool | % of SEO jobs mentioning it |
|---|---|
| ChatGPT | 5.3% |
| Perplexity | 2.5% |
| Gemini | 2.0% |
| Claude | 1.1% |
ChatGPT leads – it’s the default AI tool that comes to mind when writing job descriptions. But Perplexity is already #2, ahead of both Gemini and Claude. This makes sense. Perplexity is fundamentally a search product – people use it to find answers, just like Google. And just like Google, marketers are starting to care whether their brand shows up in the results.
We’re already seeing this shift in how marketing teams are being evaluated. First it was “Are we ranking on Google?” Then it became “Is ChatGPT mentioning us?” Now it’s “Are we being cited in Perplexity?” Perplexity’s citation-heavy model makes this especially visible. When someone asks Perplexity about your industry and your competitor’s blog gets cited instead of yours, you notice.
Claude barely registers in SEO job postings (1.1%). It’s a tool for building products, not a channel to optimize for. For marketers, this is the signal: companies are already betting that Perplexity will become a discovery channel worth optimizing for. Whether that bet pays off depends on how Perplexity’s user base grows – but the early movers aren’t waiting to find out.
6. Is Claude or Perplexity adoption more US-centric than ChatGPT?
Here’s something that surprised me: over half of Claude-only companies are American.
| Country | % of Claude-only companies |
|---|---|
| US | 53% |
| UK | 9.4% |
| Australia | 5.2% |
| Canada | 3.7% |
| Germany | 3.0% |
53% is a lot. Our dataset has companies from all over the world – the US only makes up about 10% of all companies in our index. So Claude-only companies are 5x more likely to be American than you’d expect by chance. Anthropic is a San Francisco company, and it shows. They’ve dominated the US tech scene – as we saw earlier, Claude is the tool of choice for software companies, and the US has a lot of software companies.
But internationally? Claude hasn’t broken through yet. Only 3% of Claude-only companies are German. Less than 1% are Spanish.
Compare that to ChatGPT:
| Country | % of ChatGPT-only companies |
|---|---|
| US | 38% |
| UK | 6.5% |
| Germany | 5.0% |
| Canada | 3.8% |
| Australia | 3.2% |
Only 38% of ChatGPT-only companies are American – a 15 percentage point gap from Claude. OpenAI has built a genuinely global user base. Germany is at 5% (vs Claude’s 3%). Spain is at 2.8% (vs Claude’s 0.6%). The Netherlands, the Nordics, Brazil, Italy – ChatGPT has penetration everywhere.
This makes sense when you think about it. ChatGPT had a massive consumer launch that made global headlines. Everyone’s mom heard about it. That brand awareness translates to enterprise adoption – when a company decides to buy an AI tool, the marketing team has already been using ChatGPT personally for a year.
Claude’s growth has been more word-of-mouth, spreading through engineering teams and tech Twitter. That’s a powerful distribution channel in San Francisco. It’s less powerful in Munich or Madrid.
Perplexity tells a different story:
| Country | % of Perplexity-only companies |
|---|---|
| US | 43% |
| Germany | 9.2% |
| UK | 6.2% |
| Canada | 4.3% |
| Switzerland | 2.5% |
Germany punches way above its weight – 9.2% of Perplexity-only companies are German, nearly double ChatGPT’s 5%. Hong Kong (1.6%) and Singapore (1.2%) also show up more than you’d expect.
I don’t have a clean explanation for this. Maybe Perplexity’s search-and-citation model resonates more in markets where English isn’t the first language and verifying sources matters more? Or maybe it’s just noise in a smaller sample. Worth watching.
The bottom line for marketers: If you’re selling to US tech companies, assume Claude is in the mix – it’s their default. If you’re selling globally, especially in Europe, ChatGPT is probably what your customers are already using. And if you’re selling to German companies specifically, don’t sleep on Perplexity.
Conclusion
Three AI tools, three very different buyers.
Claude is the tool for builders. Software companies, engineering teams, ML engineers. Companies that use Cursor, Docker, PagerDuty – the infrastructure of modern product development. If you’re hiring people to build AI products, you’re probably using Claude. It’s heavily concentrated in the US, and it’s rarely the only AI tool in the stack – 43% of Claude users also have ChatGPT.
ChatGPT is the tool for everyone else. Marketing teams, agencies, consultants, real estate, advertising. Less technical industries where brand recognition matters more than benchmark scores. It’s the default choice – the one you pick when you haven’t done the research. And that’s not an insult. Defaults win. ChatGPT has the most global footprint, the broadest adoption, and the biggest moat: everyone’s already used it.
Perplexity is the tool for researchers. Investors doing due diligence, lawyers doing case research, journalists fact-checking stories, government workers researching policy. Jobs where you need to find information, synthesize it, and cite your sources. It’s also quietly becoming a search channel that SEO teams need to optimize for – 2.5% of SEO job postings already mention it.
For marketers, the implications are tactical:
- Building a developer tool? Your users are probably on Claude. Make sure your docs show up when engineers ask Claude how to implement things.
- Selling to enterprise? ChatGPT is the safe bet. It’s what their teams are already using.
- Targeting investors, lawyers, or analysts? Check if you’re showing up in Perplexity results. They’re using it to research you.
- Going global? ChatGPT has the international footprint. Claude is still mostly a US story.
📊 See which companies use these tools
Explore adoption data for dev tools, security platforms, and AI infrastructure used by Claude, ChatGPT, and Perplexity customers:
