Case study / Business services / Agency operations
Smoothire: staffing agency software built around the client to placement workflow
Smoothire is staffing agency software built by Logic Square and run as its own product. It models the work an agency actually does, client, mandate, candidate, submission, client feedback, placement, so the real state of every search lives in the system rather than in one recruiter’s inbox. The stable generation has run in daily production at a working agency since mid 2023: 25 to 30 recruiters, about 180,000 candidates, about 250 placements. An AI assisted generation, built to offer Smoothire to other agencies in the US and India, entered pilot in August 2026.
- Product
- Smoothire, Logic Square’s own
- Industry
- Business services, staffing and recruitment agencies
- Status
- Stable generation in daily production since mid 2023; AI assisted generation in pilot since August 2026
- Scale
- 25 to 30 recruiters, about 180,000 candidates, about 250 placements
- Markets
- US and India
- Stack
- React, Node.js, PostgreSQL, Prisma
01 / The Challenge
The agency’s real state lived in inboxes and heads
A recruitment agency does not run on applications. It runs on mandates from clients, candidates who fit several of them at once, submissions, feedback, follow ups and the placement at the end. Most tools agencies buy were built for a company hiring for itself, so the pipeline that matters, client to placement, ends up reconstructed by hand.
The cost shows up when someone is away. A client calls, the recruiter who owns the relationship is out, and nobody can see where the search stands, what was promised, or which candidates have already been sent. When a recruiter leaves, that history often leaves too.
02 / The Decision
Model the agency, not an in house hiring department
The first decision was the data model. Client, mandate, candidate, submission and placement are first class records with ownership, status and history. A candidate can sit in several mandates. A client can run several mandates. A submission carries its own feedback trail.
The second was continuity by design. Follow ups, reminders, email history and ownership stay with the record, not the person, so an authorized colleague can pick up a mandate mid search and see everything the original recruiter saw.
The third was sequencing. Build the operational workflow first and prove it in daily use. Add AI assistance on top of a workflow recruiters already trust, and keep it in pilot until it earns its place.
03 / What We Built
The agency pipeline as the spine
Smoothire is an ATS and CRM for founder led specialist agencies of 1 to 50 people. It covers the full desk, from first contact with a candidate to submission to a client: a candidate database with AI resume parsing and a recruiter verification step, client and opening management for every search, a Chrome extension for capturing candidates while sourcing, a branded career portal for each agency, and configurable roles, permissions and teams.
How a search moves through Smoothire
- 01
Client.
The relationship, contacts and history in one record.
- 02
Mandate.
The open role, its requirements, its owner and its status.
- 03
Candidate.
One profile, sourced or imported, with ownership and update history, matchable against any open mandate.
- 04
Submission.
The candidate sent to the client for this mandate, with the client’s feedback captured in the workflow.
- 05
Placement.
The outcome, recorded against the mandate and the candidate.
Continuity when a recruiter is out
Ownership can be reassigned. Reminders, follow ups and message history travel with the mandate and the candidate. An authorized colleague takes over without a handover document.
Candidate data at scale
Structured sourcing and import, AI resume parsing with bulk upload, a central profile per candidate, and the import of a large existing database into an operational system rather than a static list.
Search, outreach and the team
Advanced search to find and match candidates against open mandates. Email templates with scheduling and sequences. Client profiles that can be shared. Task management, activity and email logs, and follow ups that keep the team coordinated. Access control, plus reporting on team performance, individual performance and revenue.
The AI assisted generation (pilot since August 2026)
On top of the established workflow, the pilot generation adds assistance: proposed candidate matches against open mandates, drafted outreach, and summaries of long candidate profiles for a submission. Recruiters make every consequential decision. This generation is in pilot; the production figures on this page belong to the stable generation.
It is also the generation that turns one agency’s tool into a product other agencies can run. All of the following is built:
- One platform, many agencies. Each agency gets its own users, teams and data on shared infrastructure, with roles and permissions configurable from day one, so a two person shop and a 40 person firm each set access the way they work.
- AI parses, a recruiter verifies. A parsed resume is searchable at once but marked unverified, and only a recruiter’s check lets it go to a client. There is no bulk verify, on purpose: a button that approves 200 profiles at once teaches people to stop reading them.
- Nothing lost by accident. Every candidate, client and note is soft deleted and can be recovered. When the product is years of relationship history, losing it is not an option.
- Rules first, learning later. Automations are rule based, so they behave predictably. Every action they take and every correction a recruiter makes is recorded as training data for later models.
- Outreach that protects the account. WhatsApp runs at 75% of Meta’s tier ceiling. Email runs with per domain caps and bounce monitoring. When a recruiter hits a daily limit, Smoothire says so and lets them decide; it never pushes past quietly.
- Two markets from the start. Multiple currencies were built in, not added later, so agencies in the US and India work in their own.
- Fewer screens, not more. Recruiters tracked work in four places: a calling queue, follow ups, an older task list and a daily planner. Instead of a fifth screen, one layer now coordinates all four. The planner became a My Day tab, and the old task list retires with a migration path.
- Sourcing and hiring surfaces. A Chrome extension for capturing candidates while sourcing, and a branded career portal for each agency.
04 / Engineering Notes
What matters before the framework list
The data model is the product. Getting the client, mandate, candidate, submission and placement records right, with ownership and history on each, is most of the value. Everything else is a screen on top of it.
Stack. React on the front end, Node.js on the backend, PostgreSQL with Prisma ORM for data.
Migration. A large legacy candidate set brought into a relational model with ownership and update history, rather than a flat import.
05 / The Results
In daily production at a working agency
The adoption test is the one that matters for agency software: recruiters either work in the system or around it. They work in it. Smoothire is now being offered to specialist agencies in the US and India; the Smoothire demo shows it working.
Does your agency’s real state live in one recruiter’s inbox?
Book a Strategy Call06 / Who this pattern fits
Any business whose work is client mandates fulfilled from a shared pool of people: staffing and recruitment agencies first, executive search, RPO and retained search, talent marketplaces with a submission step.
The test is whether the system still knows where things stand when the person who knew is unavailable.
It is also a close match for a different buyer: a business that runs on a tool it built in house and thinks others would pay for it. That includes cases where off the shelf software fights the workflow instead of fitting it, where AI is wanted but cannot be allowed to guess on customer data, or where the product has to sell into more than one market or currency. It is probably not a fit for a template site or a quick prototype with no plan to scale.
More on how we approach software for staffing and recruitment agencies and the software products Logic Square builds and runs.
FAQ
Common questions
What is staffing agency software?
+
Software that runs the agency’s actual pipeline, client, mandate, candidate, submission, feedback and placement, with the records owned by the business rather than by individual recruiters. Smoothire is one, built and operated by Logic Square.
How is staffing agency software different from an in house ATS?
+
An in house ATS models one company hiring for itself: jobs, applicants, stages. An agency has many clients, each with mandates, and candidates who fit several at once. Smoothire models that, including the submission and the client’s feedback, which an in house ATS has no place for.
Is Smoothire an ATS or a recruitment CRM?
+
Both, on one data model: an ATS for recruitment agencies and a recruitment CRM in the same system. The ATS side holds candidates, submissions and placements. The recruitment CRM side holds clients, mandates, contacts and follow ups. Agencies that run the two in separate tools end up reconciling them by hand; Smoothire never splits them.
What happens to client and candidate history when a recruiter leaves?
+
It stays. Follow ups, reminders, message history and ownership live on the mandate and the candidate, so an authorized colleague takes the search over with the full picture.
Can an agency migrate a large existing candidate database?
+
Yes. Smoothire’s pool is about 180,000 candidates, brought in through structured sourcing and import into a relational model with ownership and update history.
How do you stop AI from putting bad data into a candidate database?
+
AI proposes and a recruiter confirms. In Smoothire’s current generation, parsed candidates are searchable at once but marked unverified, and must be verified by a recruiter before they can be submitted to a client.
Can one platform serve agencies in the US and India?
+
Yes. Smoothire’s current generation supports multiple currencies from the start, so agencies in each market work in their own.
Where should AI sit in recruitment software?
+
On top of a workflow recruiters already trust, proposing matches, drafting outreach and summarizing profiles, with the recruiter deciding. That is where Smoothire’s pilot generation puts it. See our applied AI service.
Should a recruitment agency build its own ATS or buy one?
+
Most agencies should buy. Building makes sense when your workflow is your edge, generic tools keep fighting it, and the software could become a product others pay for. That is how Smoothire started. Size matters too: past about 15 recruiters, with multiple clients and shared candidates, the cost of working around a generic tool usually exceeds the cost of building.
Running your business on a tool you built yourselves?
If other companies would pay for it, we can tell you what it takes to turn it into a product.
