An agent-ready immigration or relocation service in 2026 answers five plain questions: can a machine read eligibility, can an agent move a case, can it handle documents safely, can it pay, can the firm be verified.
Published 2026-06-09 · Last reviewed 2026-06-10
A visa portal can publish a polished website, score well on a public “agent-ready” check, and still leave an agent stranded the moment it tries to act for a real applicant. The reason is the same one we saw with SaaS in our companion piece on agent-ready SaaS and dev tools: public scorers grade the marketing surface, not the part of the service that actually moves a case. Immigration sharpens that gap. The work involves sensitive documents, legal deadlines, and government counterparties that are not themselves agent-ready. A score on the website tells a buyer almost nothing about whether an agent can do the job.
This guide is the immigration-vertical version of that checklist. It is honest by design. The genuine shortlist of relocation and immigration services that pass on all five questions in mid-2026 is, as best we can find, empty. So this is not a ranked “top 10.” It is criteria plus real named examples where they do exist, plus a clear note on the rows almost nobody has shipped yet. If you are a buyer evaluating tools, or a firm thinking about your own roadmap, the five questions below are the ones that matter.
If you have not read the website-side companion, the basics are in How to make your website agent-ready in 2026. The deeper “readable vs addressable” distinction, which matters a lot in this vertical, is in Your website is agent-readable. That is not the same as agent-addressable..
Why immigration is harder than SaaS
A SaaS tool ships a unit of work an agent can finish in one session. An immigration case is months of work, dozens of artifacts, and at least three counterparties who will never be agent-ready, starting with the receiving government. That gap shapes everything else.
Three things make this vertical genuinely harder. Sensitive documents come first. A passport scan, a tax record, a medical letter, and a marriage certificate are not the kind of files a careless integration should pass around. Any agent acting on this material has to be tightly scoped, auditable, and revocable. Legal deadlines come second. A visa filing has a window, a check, and a date. An agent that misses an interview slot or misreads a priority date does real damage that a SaaS retry button cannot fix. Government counterparties come third. The US Citizenship and Immigration Services case status page is queryable by a receipt number (USCIS Case Status Online), and that is roughly the ceiling of public-side machine readability in 2026. Most consulates publish PDFs and offer no API.
The honest read is that any “agent-ready” claim in this vertical has to cover what the private service itself shipped, not what the government can be made to do. The five questions below are scoped that way. They ask what the firm can hand to an agent today, with eyes open about what sits outside the firm’s control.
Question 1: Can a machine read the eligibility rules and timelines?
Start here because it is the cheapest piece of the picture for a firm to get right, so a miss tells you a lot. The plain version is this: can an AI pull the eligibility rules for a visa category, the document list, and the rough timeline as clean text, or only as a styled web page built for human eyes?
A strong pass looks like a category page with a machine-readable summary, document checklists in plain text, and a date a reader can trust. Boundless publishes long-form category guides that are well-structured and crawlable, which puts them ahead of most immigration platforms on this row. Envoy Global publishes plain-text country guides for corporate clients. Localyze and Topia describe destination services in clean text, though the substance is mostly business-mobility coordination rather than visa rules. None of these vendors ship a machine-readable category file or an llms.txt index that an agent can pull as a list, and none publishes an OpenAPI-style description of eligibility content.
The thing to look for, and what almost nobody offers in mid-2026, is a per-category data file an agent can fetch: visa code, eligibility criteria, document list, processing-time range, last reviewed date. A typical service page reads “case-by-case, contact us for a consultation,” which is reasonable legal positioning and a real limit on machine readability. The honest read is: most immigration services pass on prose, fail on structure, and are not yet treating eligibility content as data. For the wider context on what machine-readable content is and is not, see does llms.txt actually work.
Question 2: Can an agent actually move a case forward, not just read about it?
This is the question that separates a brochure from a service. Reading category pages is passive. Moving a case means an agent can fetch the current status, upload a document into the right slot, request a missing-item list, and book or confirm an interview, all without a person clicking through a portal each time.
In 2026 the common answer for a SaaS is a first-party Model Context Protocol server, often called an MCP server, which is the de-facto way to let an agent see a tool’s actions and run them through one connection. No dedicated immigration platform we found ships one in mid-2026. The closest are the global-hiring players that border the vertical, and two of them have crossed the line into agent action. Deel now runs a first-party MCP server (at api.letsdeel.com/mcp) that goes past reading data: it exposes write actions such as submitting a time-off request, across HR, payroll, and contractor workflows, with a published llms.txt tool index. Remote ships a first-party MCP server too, giving an agent a live connection to payroll, contracts, compliance data, and org structure, and it runs a separate Mobility product that handles visas and relocations. Neither MCP server, though, exposes the verbs that actually move a visa filing: case status, document upload, missing-item list, interview scheduling. They are HR and employer-of-record surfaces that happen to touch immigration, not case-management APIs an agent can drive a filing through. Envoy Global offers an enterprise integration layer for HRIS connections, which is closer to data exchange than to agent action. Boundless and Localyze offer customer portals an agent can in theory drive with a browser, but that is screen-scraping, not a documented action surface.
The strong artifact to look for, when it ships, is a documented case-action API or a first-party MCP server with at least these verbs: get case status, upload document, request missing-item list, schedule or reschedule an interview slot. A brochure page that lists “AI features” without one of those is not a tool an agent can use today. (For the SaaS-side version of this same question, with worked examples, see How to Tell If a SaaS Is Agent-Ready in 2026.)
Question 3: Can an agent log in and handle sensitive documents safely?
This is the row that decides whether an agent can be trusted with the work at all, and it is the one that public scorers skip. The job is narrow: can an agent log in for a person, with permissions limited to one case and a small set of actions, and can it touch documents through a defined channel that the firm can audit and revoke?
A safe agent login has three traits a buyer can feel. The agent gets permissions scoped to one case, not the keys to a whole client account. The agent can register, refresh, and rotate its own credentials without a person clicking through a console (search the auth docs for the term “dynamic client registration” or “RFC 7591”). And there is no human-only step in the path, no QR code or fingerprint scan, sitting between the agent and a routine action like a document upload. Limited-permission logins are common in adjacent enterprise tools; agent self-registration in this vertical is still rare.
Documents are the harder half. The pattern to look for is an explicit document-handling channel with a per-document permission, an audit trail of which agent did what and when, and a clear revocation path. Deel and Remote both publish role-based access controls for HR data, which is the closest mature pattern to draw from. Envoy Global describes audit logging for enterprise immigration cases. None of the consumer-facing immigration platforms we reviewed publishes a per-action audit log for an outside agent acting on a case, because the agent role does not yet exist in their permission model. That is the gap to ask about in a vendor demo: “show me where, in your access model, an outside agent fits, and how an audit log lists its actions.”
If the answer is “we have not designed that yet,” it is honest, and it is also the cap on what an agent can safely do for the case.
Question 4: Can a machine understand the fees and pay?
Immigration fees split cleanly in two. Government fees are fixed amounts published by the receiving country (the US I-485 fee, the UK Skilled Worker visa fee, and so on). Firm fees are what the platform or law firm charges on top, and that side is often case-by-case. A machine-readable answer covers both: the predictable government cost is a number an agent can plan around, and the firm cost is either a published rate sheet or a quote endpoint a request can address.
A real payment standard for agents now exists, which is new. Stripe and Tempo launched the Machine Payments Protocol on 18 March 2026, which runs through Stripe’s existing PaymentIntents API (Stripe). Google’s AP2 was announced on 16 September 2025 and donated to the FIDO Alliance on 28 April 2026, with 60-plus partners including Mastercard, Visa, and American Express (AP2). The crypto-side rail is x402, hosted by the Linux Foundation’s x402 Foundation since 2 April 2026 (x402). The same rails apply here. The differences are size and spend limits: an immigration case can run into the tens of thousands of dollars between government fees and firm fees, so a per-action cap and a per-case ceiling are not optional, they are the baseline a buyer should require before letting an agent move money.
In mid-2026 we did not find an immigration platform that publishes machine-readable price markup on its pricing page or accepts payment through one of the named agent rails. Boundless and iVisa publish flat-rate package prices on the page, which is closer to readable than a “request a quote” form, but neither markup nor an agent-payment path is documented. This is a row where the vertical has clear room to move and the rails to use are already shipping.
Question 5: Who is the firm behind the service, and can an agent verify them?
This row matters more here than in any other vertical. Immigration scams are a known and well-documented harm, and an agent acting for a person has to be able to check that the firm on the other side is a real, licensed practice before it hands over a passport scan or a payment. The question, in plain words, is whether the firm publishes a public, machine-checkable identity that an agent can verify, and whether the identity links to something a regulator actually stands behind.
In 2026 no major immigration platform we reviewed publishes the kind of agent-checkable identity record this question is asking about. Some firms list licensed attorneys with bar numbers, which a human can spot-check on a state bar website, but that is not yet a machine-readable record. On the protocol side, the emerging shape is an A2A Agent Card at /.well-known/agent-card.json (A2A v1.0.1, 28 May 2026) that names the operating organization, paired with an ENSIP-27 record at /.well-known/agent.json for the optional on-chain identity link. Cloudflare’s Agent Readiness checks the website surface but does not grade the operator-identity row (Cloudflare). The honest read is: this is unsolved across the immigration field, and a verification step that pretends otherwise is doing the firm a favor instead of the applicant.
This identity question is the gap Tobira works on, alongside protocol-side naming approaches. A Tobira handle gives the firm behind a service a public profile other agents can check, with a step where both sides confirm who they are before either acts. It is one layer in agent-readiness, not the whole picture. For why a readable site is not the same as a verifiable firm, see Your website is agent-readable. That is not the same as agent-addressable.. For the protocol-side naming patterns, see How A2A Agent Cards work.
A five-minute test you can run on any service
You can run all five questions yourself in about five minutes, no account required.
Can a machine read the eligibility rules?
Open a visa category page (an EB-1A guide, an O-1 guide, a UK Skilled Worker guide). Try suffixing the URL with .md or look for a “view as markdown” toggle. Check the page for a clear last-reviewed date and a structured document list. Styled HTML only, with no date and a prose-only list, is a fail.
Can an agent move a case?
Search “[firm name] API” or “[firm name] MCP server.” A first-party result that documents case-action verbs is a strong pass. An employer-of-record API that covers HR is partial. No documented action surface is a fail.
Can an agent log in and handle documents safely?
Open the auth doc, if there is one. Search for “dynamic client registration,” “RFC 7591,” or an oauth/register endpoint. Then look for a documented per-document access scope and an audit-log endpoint. Two of three is a strong pass for this row in 2026.
Can a machine read fees and pay?
Look at the pricing page for Schema.org Product or Offer markup. Then search the docs for “Machine Payments Protocol,” “AP2,” or “agent payment.” Readable price plus a documented agent-payment path is the bar; a flat-rate page alone is partial.
Can an agent verify the firm?
Search the site for a /.well-known/agent-card.json or /.well-known/agent.json. Search for a Tobira handle or a recognized agent-network profile. A hit is a pass. A miss is the normal 2026 finding, and naming it openly is the right call.
Takeaways
In mid-2026 the immigration vertical sits one full step behind SaaS on agent-readiness, and the honest answer to “which firms pass all five?” is “none we could find.” That is not a reason to wait. The rails are shipping, the patterns are visible, and the buyer-side check that names the misses openly is more useful to applicants than a manufactured ranking. If you are a firm, the cheapest move is the eligibility-readability row: a structured category file with a last-reviewed date is a weekend of work and is the strongest entry signal. If you are a buyer, the row to push hardest on is firm identity, because the stakes here are higher than in any other vertical and the rest of the checklist does not matter if the firm on the other side cannot be verified.
FAQ
What does agent-ready mean for an immigration or relocation service?
It means an outside agent acting for a person can do the job, not just read about it. The plain checklist is five rows: can a machine read the eligibility rules and timelines, can an agent move a case forward (status, documents, scheduling), can the agent log in and handle documents safely with limited permissions, can a machine read fees and pay through a recognized agent rail, and can the firm be verified through a public agent-checkable identity. A pass on the website alone does not count.
Is any immigration platform agent-ready on all five questions in 2026?
No, as best we can find in mid-2026. Several pass partly: Boundless and Envoy Global on eligibility content, Deel and Remote on adjacent HR tooling (both now run first-party MCP servers, though for HR and payroll rather than visa case-action), Boundless and iVisa on readable pricing. None passes on a documented case-action surface, an agent-fit access model with audit logging, or a machine-checkable firm identity. Naming the gaps honestly is the point of this checklist.
Why is firm identity such a big deal for immigration?
Because visa scams are real, the documents in play are sensitive, and an agent has no way to spot-check a license on its own. If the firm does not publish a verifiable identity that an agent can confirm before handing over a passport scan or a payment, the rest of the checklist does not matter. That is why this row outweighs the others in the immigration vertical specifically.
How would an immigration firm pay through an agent rail?
Through one of three named rails. Stripe and Tempo’s Machine Payments Protocol (18 March 2026) runs through Stripe’s PaymentIntents API. Google’s AP2 was donated to the FIDO Alliance on 28 April 2026, with 60-plus payment-network partners. x402 sits at the Linux Foundation since 2 April 2026. The size of immigration bills makes per-action caps and per-case ceilings non-negotiable.
What is the single cheapest move an immigration firm could make to become more agent-ready?
Publish a structured category file. One JSON or YAML file per visa category, with eligibility criteria, document list, processing-time range, and a last-reviewed date, posted at a stable URL. This is a weekend of work, costs almost nothing, and is the strongest first signal an agent can use. The harder pieces (a case-action API, dynamic client registration for agent login, a verifiable firm identity) take longer, but this one is on the table today.
Sources
- A2A protocol specification, v1.0.1, 28 May 2026 (Linux Foundation): https://a2a-protocol.org/latest/specification/
- Model Context Protocol specification: https://modelcontextprotocol.io/
- Stripe and Tempo, Machine Payments Protocol (18 March 2026): https://stripe.com/blog/machine-payments-protocol
- Google, Agent Payments Protocol (AP2), donated to FIDO Alliance 28 April 2026: https://ap2-protocol.org/
- Linux Foundation, x402 Foundation (2 April 2026): https://x402.org/
- Cloudflare, Agent Readiness score and Agents Week 2026: https://blog.cloudflare.com/agent-readiness/
- Cloudflare, Is It Agent Ready? URL Scanner: https://isitagentready.com/
- W3C Web Machine Learning Community Group, WebMCP Draft: https://webmachinelearning.github.io/webmcp/
- US Citizenship and Immigration Services, Case Status Online: https://egov.uscis.gov/casestatus/landing.do
- Boundless, US immigration platform: https://www.boundless.com/
- Envoy Global, corporate immigration platform: https://www.envoyglobal.com/
- Localyze, global mobility platform: https://www.localyze.com/
- Topia, global mobility platform: https://www.topia.com/
- Deel, global hiring with visa support: https://www.deel.com/
- Deel, first-party MCP server documentation: https://developer.deel.com/mcp
- Remote, global hiring with EOR and visa support: https://remote.com/
- Remote, first-party MCP server: https://remote.com/remote-mcp/