It is a proper TMS.
It manages projects and jobs, translators and vendors, translation memories, terminology, automation, machine translation, AI translation, analysis, costs and quotes.
Clients can submit work through a portal. Providers can be assigned inside the system. Project managers can analyse costs and savings.
So if your argument starts with “Phrase does the words, but you need another system to manage the project”, anybody who actually uses Phrase is going to stop reading.
The real gap is narrower.
Phrase is very good at running localisation production.
A translation agency still has to run a business around that production.
Those two worlds overlap considerably, but they are not identical.
Start with what Phrase already does
A Phrase project can contain multiple target-language jobs and multiple workflow steps.
Project templates can preconfigure resources and workflows. Jobs can be assigned to linguists or vendor organisations. Status changes and automation can move work through the process without a PM manually pushing every button.
Phrase also has a Submitter Portal through which a client or internal requester can choose languages, upload files, specify a deadline and PO number, see pricing and later retrieve completed files.
That already covers a large proportion of what people casually mean by “translation project management”.
It also has real financial functionality
Phrase does not merely calculate word counts.
It has price lists, net-rate schemes and quote functionality.
Quotes can be generated from a Phrase analysis using the relevant price list and workflow steps. They can be produced for buyers or providers, emailed, approved and reported on.
Phrase Analytics also includes cost and quote dashboards based on the pricing attached to projects and providers.
So take a straightforward job:
20,000 English words
French and German
translation + review
Phrase can analyse those jobs, apply TM discounts, associate provider pricing and calculate costs.
It is perfectly capable of telling you far more than “20,000 words need translating”.
That is why the business-management boundary needs to be described carefully.
A quote is not necessarily an invoice
Suppose Phrase calculates that a project will cost your customer £4,800.
Excellent.
The customer approves the quote.
The translation work happens.
At some point, however, your agency may need to issue a legally numbered invoice with your payment terms and VAT treatment, send it through your accounting workflow, record a credit note if necessary and reconcile the eventual bank payment.
Those are accounting and accounts-receivable processes.
Likewise, Phrase may calculate that a linguist’s work costs £720.
That does not necessarily mean your agency has issued its own supplier purchase order, received the freelancer’s invoice, approved that bill and paid it from the bank.
This is the distinction between project financials and business accounting.
Phrase is strong at the former.
Most agencies still use another system for the latter.
Phrase itself lists business-management platforms including XTRF and Plunet among the third-party systems that integrate with Phrase TMS, which is a useful illustration of that division.
Provider management is also stronger than a simple freelancer list
Phrase can work with both individual linguists and vendor organisations.
Vendor connections allow projects or jobs to be shared with another Phrase organisation. The vendor can then manage its own translators without exposing them directly to the buying organisation. Pricing information such as net-rate schemes and price lists can be attached to vendor relationships.
That is substantial supply-chain functionality.
What an agency may still want outside Phrase is the broader relationship around the vendor.
- What services do they offer outside this particular Phrase workflow?
- What are their rates for certification, interpreting or DTP?
- When did we last use them?
- What is their payment term?
- Have they sent us the required compliance documents?
- What do we owe them across all current projects?
- What has actually been invoiced and paid?
Again, some agencies solve all of this through additional enterprise software.
Others solve it with a spreadsheet.
Phrase in 2026 is also an AI translation platform
This is perhaps the biggest omission in a conventional description of Phrase today.
Phrase still supports traditional MT and MTPE workflows through Phrase Language AI, including multiple MT engines and AI-based quality estimation. Phrase explicitly documents MTPE as a workflow where machine output is followed by professional editing and review.
But it now also offers AI Translation Agent.
Rather than simply returning one MT suggestion per segment, the agent uses generative AI together with translation memories, term bases and style information, then performs its own refinement and error-correction steps.
Phrase is increasingly building workflows around the idea that not every segment necessarily needs the same amount of human intervention.
Its QPS quality-estimation system can score machine-translated output, and Phrase has been adding AI-based quality rules and workflows that can automatically separate content that needs attention from content considered good enough to progress.
For an agency, that has commercial consequences.
A client may buy:
- human translation;
- MTPE;
- AI translation with human review;
- or a largely automated workflow with quality-based human intervention.
The production logic belongs in Phrase.
The corresponding sell rate, supplier rate and eventual margin still need to survive outside it.
Where things become awkward is when two systems disagree
The problem with running Phrase alongside a separate business system is not that either one is bad.
It is duplication.
If Phrase says Jonas Keller is translating German but your project-management system says it is Anna Weber, which is correct?
If the Phrase analysis changed after the source files were updated but your quote still uses yesterday’s counts, which number are you billing?
If Phrase says German is complete but your commercial project still says “in progress”, which dashboard does the PM believe?
This is where integration matters more than feature checklists.
Why the VectorLingo integration with Phrase goes deeper
Phrase is currently VectorLingo’s deepest CAT/TMS integration.
VectorLingo can create the Phrase project and jobs, push the source files and pull job status information back as production progresses.
The Phrase analysis can feed the weighted-word quote without being retyped.
Most importantly, provider assignments in Phrase are synchronised into VectorLingo, so assigning a linguist to a Phrase job can also keep the commercial workflow aligned with who is actually doing the work.
That last part matters.
There is little point integrating a TMS with a business-management system if the project manager must still maintain the supplier assignment twice.
Phrase is the only VectorLingo integration that currently synchronises those vendor assignments. Trados has two-way project and file integration but not vendor-assignment sync; memoQ currently has analysis import only.
What should stay in Phrase?
Quite a lot.
- The source files.
- The jobs.
- Translation memories.
- Term bases.
- MT and AI translation.
- Pre-translation.
- The CAT editor.
- Linguistic review.
- QA.
- Production statuses.
- Phrase’s own project automation.
Trying to reproduce those things in a translation business-management system would be pointless.
Phrase is already built for them.
What can sit around Phrase?
- The wider agency record.
- The original sales quote.
- The client’s complete rate card.
- Supplier rates beyond the particular Phrase job.
- Purchase orders.
- Commercial deadlines.
- Non-CAT workflow steps such as certification or courier delivery.
- Supplier liabilities.
- Client invoices.
- Payment status.
- Project and client profitability.
- Accounting integration.
The exact line will differ by agency.
What matters is that every object has an owner.
Do you actually need another system?
Not necessarily.
If Phrase already handles your operational requirements and you have connected it to something like XTRF or Plunet for the business layer, adding VectorLingo may simply duplicate software you already have.
Likewise, an internal localisation team that never invoices customers and has no freelance suppliers may need almost none of this.
The clearer VectorLingo use case is a smaller or growing agency that already likes Phrase for production but still runs the commercial side through some combination of spreadsheets, email and accounting software.
That agency does not need another enterprise localisation platform.
It needs the pieces around Phrase joined together.
The short version
Phrase can manage translation projects very well.
It can automate projects, assign providers, analyse content, calculate buyer and provider quotes, track costs and run increasingly sophisticated MT and AI translation workflows.
So the case for adding business-management software is not that Phrase lacks project management.
It is that running an agency eventually extends beyond the production project.
The Phrase job tells you what is being translated, by whom, for how much and how far through the workflow it has progressed.
The agency still needs to turn that work into supplier liabilities, client invoices, cash, margin and a longer-term client and vendor relationship.
For agencies using both Phrase and VectorLingo, the integration is intended to keep those two views of the same project aligned rather than forcing the PM to maintain them twice.