A straightforward comparison for UK ops teams, agencies and SMBs weighing up Anthropic's Claude and Make.com (formerly Integromat) — and why most end up using both.
Claude Cowork and Make.com are not really competitors — they do fundamentally different jobs. Claude (by Anthropic) is a thinking and writing assistant: it reads documents, drafts replies, summarises reports, analyses data you paste in. Make.com is a visual automation platform: it connects your apps (Gmail, HubSpot, Airtable, Xero, Slack, hundreds more) and runs scheduled or event-driven workflows between them. The best setup for most UK businesses is both: Make handles the plumbing between tools, and Claude sits inside those workflows wherever the step needs reasoning, writing or classification.
If you only need one, start with Claude: it delivers value on day one for any professional doing document work. Add Make.com once you have a repeating cross-tool workflow worth automating (lead intake, invoice follow-ups, proposal pipelines). The two combine beautifully — Make orchestrates, Claude provides the thinking step inside the scenario via its official Claude integration module.
The first thing to clear up: Claude and Make.com are not alternatives to each other in the way that, say, Claude and ChatGPT are. They sit in completely different categories of software. Framing it as a head-to-head is a bit like asking whether a solicitor or a filing system is more useful for your practice — the honest answer is that you want both, and they do unrelated jobs.
Claude is an AI assistant. You open a chat window, paste a contract, and ask it to flag unusual clauses. You hand it a long client email and ask for a measured reply in your firm's tone. You give it the brief and the past three proposals and ask for a draft of the fourth. The value comes from a thinking, reading and writing partner that happens to be extremely fast and extremely patient.
Make.com (previously known as Integromat, rebranded in 2022) is an automation platform. You open a canvas, drag on a Gmail trigger module and a HubSpot action module, connect them, and when a new email arrives from a particular address a contact is created in your CRM. No one types anything into Make day to day; once the scenario is built, it just runs. The value comes from eliminating repetitive data shuffling between the tools your team already uses.
Make is one of the most capable no-code automation platforms available, and for UK ops people it is often the first serious upgrade on Zapier. The visual canvas shows your entire scenario at a glance — triggers on the left, actions flowing to the right, with routers branching when conditions diverge. For anyone who has ever outgrown a spreadsheet of manual tasks, it feels like switching from a push-along mower to something with an engine.
A few things Make does particularly well. Its pricing is operation-based rather than task-based, which means complex multi-step scenarios tend to work out much cheaper than they would on Zapier. It handles iteration gracefully — looping over a list of invoices, a batch of leads, a folder of files — in a way that competitors often charge a premium for. It has strong error handling, with built-in retry logic and dedicated error routes so scenarios degrade rather than fail silently. And its library of over 1,800 integrations covers every mainstream tool a UK SMB is likely to touch, from Xero and QuickBooks to Monday, Notion, Microsoft 365, Google Workspace, Slack, HubSpot, Pipedrive and the big helpdesks.
Perhaps most importantly for UK businesses: Make is a European company (Czech-headquartered, EU-hosted) and takes GDPR seriously. That matters when your scenarios are moving personal data between systems.
Make does not think for you. It can move data between places, transform it with simple functions, and apply conditional logic — but it cannot read a contract and tell you what to worry about, cannot draft a proposal in your firm's voice, and cannot look at a client's email and compose a thoughtful reply. These are language and reasoning tasks, and they sit outside automation platforms by design.
This is where the AI module library comes in. Make has official modules for Anthropic's Claude, OpenAI's ChatGPT, and several other AI providers, specifically so that when your scenario needs a thinking step, you can call out to a model. A typical setup: a lead form submission arrives, Make picks it up, passes the message to Claude with a prompt like "classify this enquiry and draft a first-pass reply in our house tone", then routes the classification into the CRM and the draft reply into the relevant inbox for a human to review before sending.
Without the AI step, that scenario is just data shuffling. With Claude dropped in, it becomes a workflow that actually reads what the prospect wrote and responds to it — but still inside Make's orchestration, logging, and error-handling.
The pattern we see most often with our clients is straightforward. Claude is the assistant people use directly — open the chat, paste the document, have the conversation. Make is the silent machinery behind the scenes, handling the workflows that do not need a human in the loop every time. And wherever those workflows need a reasoning step, Make calls Claude.
Three concrete examples from recent engagements. A recruitment agency uses Make to watch their job-board inbox; each new application is sent to Claude for a structured summary (candidate background, red flags, fit against the brief), and the summary is posted into a Slack channel where the consultant can act on it in seconds. A small accountancy practice uses Make to pick up client emails tagged as queries; Claude drafts a first-pass reply referencing the firm's response templates, and the draft lands in the partner's inbox as a saved draft to review and send. A marketing agency uses Make to pull new transcripts from Fathom, send them to Claude for an action-item extraction and client-ready recap, and auto-post the recap into the correct client's Notion workspace.
In each case, Make handles what Make is good at (triggers, routing, pushing things between tools) and Claude handles what Claude is good at (reading, reasoning, writing). Neither tool could do the job alone, and both become more valuable when wired together.
If you are starting fresh and can only roll out one, start with Claude. The return on a Claude subscription is visible on day one — your team does less rewriting, drafts land faster, documents get read properly, long emails stop blocking the afternoon. There is no setup, no scenarios to build, no app connections to manage. You simply open it and start using it.
Add Make.com once you can point to a repeating cross-tool workflow that is eating hours. Lead intake into the CRM. Invoice reminders. Proposal follow-ups. Weekly reporting pulled from three systems and pushed into a dashboard. Client onboarding. Once you have a clear target, Make becomes very compelling — a few afternoons of work replaces a recurring grind that was quietly taxing your team.
The sequencing matters because Claude's value is unconditional (anyone who writes benefits immediately), whereas Make's value is conditional on having a workflow worth automating. Rolling Make out before you have identified the workflows tends to produce a half-built scenario graveyard and a modest feeling of disappointment. Rolling Claude out first, letting your team get fluent, and then noticing "actually, this thing we keep doing manually should be a Make scenario with Claude in the middle" is a much happier path.
Treat them as complements, not alternatives. Standardise on Claude for everyday thinking and writing work — proposals, client correspondence, document analysis, meeting notes, reporting drafts — and build Claude Projects per client or workflow so outputs stay on-brand. When a specific cross-tool process is costing you real hours every week, bring in Make to automate it, and use the Claude integration module inside any Make step that needs language or judgement.
If you want a hand mapping out which workflows are worth automating, which should stay manual, and how to wire Claude into the ones that are — that is exactly the kind of advisory work we do alongside our training. Book a discovery call below.
Not really — they do different jobs. Make.com (formerly Integromat) is a visual automation platform that connects apps and runs scheduled or event-driven workflows between them. Claude is an AI assistant that reads, reasons and writes. Most UK businesses who use both run Make for the plumbing between tools and call Claude inside a Make scenario whenever that workflow needs a thinking or writing step.
No. Make.com can route data between HubSpot, Gmail, Airtable, Xero and hundreds of other tools, and it can call an AI model as one step inside a scenario — but it is not itself an assistant you talk to about a document, a proposal, or a client's email. For reading, reasoning and drafting, you still need Claude (or a similar model) sitting inside the workflow.
No. Claude does not sit between your apps and move records around on a schedule. It will not watch your inbox for new enquiries, push rows into a CRM, trigger an invoice in Xero, or sync data between tools. Those are integration jobs, and Make.com (or Zapier, or n8n) is what you want for them.
Make.com has an official Anthropic Claude integration module. Inside a Make scenario you can pass data (an email, a form submission, a row of a spreadsheet, a PDF) to Claude, ask it to summarise, classify, draft a reply or extract fields, then pipe the result into the next step — Slack, HubSpot, Notion, your helpdesk, wherever. Make handles the triggers and routing; Claude handles the thinking and writing.
For most UK SMBs, yes — Make is often more cost-effective than Zapier at volume because its pricing is operation-based rather than task-based, and complex multi-step scenarios tend to cost less to run. The free tier (1,000 operations a month) is enough to prototype most things, and paid plans start around £7-9/month. The main cost is a few hours of learning the visual builder.
No. Make is designed for non-developers — scenarios are built visually by dragging modules onto a canvas and connecting them. That said, there is a learning curve: you need to understand how data maps between modules, when to use iterators and aggregators, and how to handle errors. Most UK ops people can be productive within a week or two of focused practice.
Start with Claude. It delivers value on day one for any team that writes, reads documents, or communicates with clients — no setup, no scenarios, no app connections. Add Make.com once you have a clear, repeating cross-tool workflow worth automating (lead intake, invoice reminders, proposal follow-ups). At that point, wire Claude into the Make scenario and let them do what each is good at.
Book a free 30-minute discovery call or explore our Claude Cowork training — no AI experience needed.