Ramp is doing something more interesting than content marketing

I have seen the name Ramp a lot over the past few years.

I spent around six years working in and around the Sage Intacct world, so I still keep an eye on the technology ecosystem around it: the platforms, integrations and companies that keep appearing around finance teams. Ramp is one of those names. It has a direct Sage Intacct integration, and the relationship developed into a multi-year strategic partnership.

Ramp is especially visible in the UK right now. On 15 September 2026, it formally launched here after a summer beta, opening corporate cards, expenses, bill payments and accounting automation to UK businesses. Earlier in the year it had also announced London and Stockholm offices as part of its European expansion.

That makes the timing interesting to me. Ramp is not just a US fintech I happen to be watching from the Sage Intacct ecosystem. It is actively trying to establish itself in the market around me.

But the thing that made me stop and look more closely was not the UK launch or the integration. It was Ramp Economics Lab, because it is a very good example of something I have been trying to describe with Editorial Intelligence.

The evidence already exists

Ramp Economics Lab describes itself as a think tank studying how companies spend and operate, powered by real-time business spending data. That starting point matters because the evidence is not being assembled after somebody has already decided what article they want to publish.

Ramp sits inside the financial activity of businesses, so its product generates a view of what companies are actually spending money on. The Economics Lab can turn aggregated and anonymised versions of that behaviour into research.

Its work includes the Ramp AI Index, which tracks business adoption of AI products, and the Ramp Spend Share Index, which looks at where business spending is moving. It has also published research combining Ramp spending data with workforce data to examine what happens to employment after companies adopt AI heavily.

That is a very different content sequence from the one most teams are used to.

Product → behaviour → evidence → question → research → story.

The evidence exists before somebody decides they need an article. That changes what the editorial job is because the value is not mainly in filling the page; it is in recognising which signal deserves to become a story at all.

Someone still has to notice what matters

The other thing that caught my attention was who Ramp is putting around this work. The Economics Lab team draws on economists, applied scientists, analytics and data science expertise, which might sound like the opposite of editorial until you look at the questions they still have to answer.

Having proprietary data does not automatically give you useful thought leadership. Someone still has to work out what the data is saying, what it is not saying and why anybody outside the company should care.

The questions are familiar:

• What is actually interesting here?
• Is this a genuine signal or noise?
• What can the evidence support?
• What can it not support?
• What is surprising?
• What deserves another question?
• What would somebody outside the company actually want to know?
• How do you explain the finding without overstating it?

Those are research questions, but they are also editorial questions. The interesting part is that the boundary between the jobs starts to matter less than the combined ability to recognise evidence, interrogate it properly and make it useful.

Ramp is hiring for the editorial layer

The job architecture makes this even more explicit. Ramp is currently hiring a Finance Storytelling Lead in New York, with a target base salary of $160,000 to $220,000 plus equity.

The salary is interesting, but the job itself is more interesting. This is not a conventional content role sitting downstream from the business waiting for briefs.

Ramp says the person will spend roughly half their time embedded with its Finance team, learning how the team actually operates, and the other half working on customer storytelling. The role is expected to sit through messy meetings, spot what matters, ask the question nobody else asked and turn the raw material into something useful.

It also reaches directly into the evidence layer. The person will work with the Economics Lab and Ramp’s proprietary spend data, identify questions the existing data cannot answer, develop original research and practitioner-led analysis, and help build a repeatable system for finding ideas, capturing the raw material, developing drafts and shipping them.

That combination is striking:

Reporting → operational access → proprietary data → editorial judgement → research → story → distribution.

Ramp even asks for someone who is “AI-native, but use[s] personal judgement”. That distinction matters. AI is there to reduce busywork and explore angles. The value of the person is still in judgement: recognising the signal, asking the better question and deciding what deserves attention.

The role also makes the ambition unusually clear. Ramp wants to become a company finance leaders learn from, not simply one they buy from.

That is a much bigger editorial brief than producing content around a product. It is about turning the organisation itself into a source of useful knowledge.

This is what I mean by moving upstream

I have written quite a lot recently about editorial work moving upstream. The easiest version of AI content production is to make the production layer faster: brief, draft, edit, publish.

AI can obviously accelerate all of that, but it does not necessarily make the underlying material more valuable. If ten companies use AI to write from roughly the same public information, we should probably expect a lot of competent content saying roughly the same things.

The more interesting opportunity appears earlier, before anybody asks for copy. It starts with questions such as what the organisation knows, where that knowledge lives, what evidence the product generates and what customers are revealing through their behaviour.

It also asks which of those signals are strong enough to become something external. Ramp Economics Lab is an unusually clear example because the proprietary evidence is sitting very close to the product, so the editorial opportunity begins with recognising the signal rather than generating the paragraph.

That distinction matters to me because it changes the centre of gravity of the job. Editorial judgement is no longer only applied to a draft; it is applied to the evidence, the questions and the decision about what deserves to exist in the first place.

Proprietary evidence becomes more valuable in AI search

There is another reason I think this model matters. AI systems make generic explanation incredibly cheap, which means another 1,500-word explainer on a well-covered topic is unlikely to become more distinctive simply because it was produced faster.

Ask an AI assistant what accounts payable is and it does not particularly need another software company to explain the definition. Ask what is happening to business adoption of AI products this month and the task changes because the answer now depends on evidence.

If your organisation produces useful, defensible evidence that does not exist elsewhere, you have something machines as well as humans may need to retrieve. That is a much stronger position than trying to win attention with another generic explanation.

Ramp has pushed this further than simply publishing articles. It has made parts of its economic data available through API, MCP and CLI access, while also building data partnerships that can put those signals inside other products, research and AI experiences.

That progression interests me enormously.

Product → evidence → research → story → structured knowledge → retrieval.

The article becomes one interface to the evidence rather than the final destination for it. That is very close to how I think Editorial Intelligence needs to work, especially as search becomes less about retrieving pages and more about producing answers and recommendations.

Not every company needs an economics lab

I would not take the lesson from Ramp that every B2B software company should hire economists and create a think tank. Most companies would be better served by first understanding what useful evidence they already generate and whether they have a credible reason to publish it.

The better question is:

What does our position in the market allow us to know?

The answer will be different for every organisation. It might come from transaction data, hiring behaviour, customer conversations, operational benchmarks, support questions, search behaviour, product usage, research programmes, practitioner interviews or hundreds of customer stories that currently disappear into the archive once the campaign has finished.

The evidence will be different, but the operating principle is the interesting bit:

• Find the evidence.
• Preserve it.
• Understand its limitations.
• Connect it to other signals.
• Interrogate it.
• Then decide what deserves to become content.

That is almost the reverse of starting with a content calendar and asking what should be published next. It starts with organisational knowledge and works forward from there.

The company knows things before the content team writes things

That is probably the idea I keep coming back to. Organisations are producing useful knowledge constantly, but most content systems only become active after somebody has already asked for an output.

Products generate signals. Customers reveal problems. Researchers uncover patterns. Sales teams hear objections. Events expose contradictions, and customer stories contain evidence that might matter again six months later.

Most content systems begin too late because they start when somebody asks for an article. By then a lot of the most valuable evidence has already been flattened into notes, decks, CRM entries, product telemetry or somebody’s memory.

Ramp Economics Lab is interesting because you can see what happens when the system starts earlier. The product creates evidence, people with the right expertise interrogate it, editorial judgement helps identify what matters, and publishing makes it useful outside the organisation.

Structured access then allows the knowledge to travel further than the original article. It can become something a journalist, analyst, customer, AI system or internal team can interrogate again later without starting from zero.

That is not really a content engine.

It is an intelligence system that happens to produce content. And I suspect we are going to see a lot more of those.