The default stack for a modern tech-company recruiting team in 2026. It is built on one assumption: AI handles the high-volume, low-judgment work — sourcing, scheduling, screening synthesis — while humans keep the high-judgment work of relationships, evaluation, and closing. It is tuned for the 100-1,000 employee tech company hiring at meaningful but not enterprise scale, with a recruiting team small enough that every coordinator hour matters.
How the pieces fit
- Ashby is the ATS backbone. Native analytics deep enough that most teams never bolt on a separate BI tool, a modern API, and structured-interview scorecards as a first-class object. This is where funnel data lives and where every stage transition gets recorded. Ashby now publishes a floor — Foundations at $400/month flat for companies up to 100 employees — while Plus (101-1,000 employees) and Enterprise stay quoted on size, usage, and commitment. The line to model is the AI meter: 1,500 credits per month on Foundations, 2,500 credits per seat per year on Plus, 12,500 on Enterprise, with the AI Notetaker, advanced scheduling, and SSO ($100/month) priced as add-ons on top.
- Gem is the recruiting CRM. Long-cycle candidate pipelining, outbound sequence management, talent-community building. Gem now sells in two shapes, and only one of them belongs in this stack: “Gem + Your ATS” keeps Ashby as the system of record, while Gem All-in-One ships its own ATS and overlaps Ashby outright. Both are quoted on FTE count. Seats split into Standard and AI-Powered, with the sourcing, application-review, talent-rediscovery, and fraud-detection agents sitting behind the AI seat — so the agent roster, not the headcount, sets the Gem bill. The startup program is the only published number: $270/month list, $130/month on an annual term at 1-10 FTE.
- juicebox (PeopleGPT) is the AI sourcing layer. Natural-language candidate discovery against 800M+ profiles, which replaces hours of Boolean LinkedIn search with a plain-English query. Seats are published: Starter at $99 and Growth at $159 per seat per month on annual terms. The autonomous Agent is metered separately at $199 per agent per month, or $300/month for two agents on an annual term — that is what keeps a hard-to-fill req collecting fresh matches without a sourcer re-running the search.
- BrightHire is the interview-intelligence layer. It records and AI-analyzes every interview, produces decision-grade evidence for debriefs, and flags interviewer-behavior patterns that hurt candidate experience and quality. Zoom acquired BrightHire in December 2025 and it still ships cross-platform — capture works on Zoom, Microsoft Teams, and Google Meet, so the acquisition does not force the rest of the stack onto one video vendor. Its three Interview Intelligence tiers are all quote-based, and BrightHire Screen, the AI interviewer, sells standalone or bundled with them.
- ModernLoop is the interview-scheduling layer. It automates multi-person, multi-stage loop coordination and removes the coordinator scheduling tax that otherwise burns 1-2 FTE at this scale. The product line has split: the scheduling platform stays quoted, while the Taylor AI agent publishes credit pricing — $450/month for 500 credits on Starter, $1,250/month for 1,500 credits on Pro, $1 per credit beyond that. A completed AI interview costs 5 credits and an assistant conversation 0.5, which makes the automated phone-screen layer priced per interview rather than per seat.
- Claude is the horizontal AI surface. JD writing, interview loop design, debrief synthesis, rejection feedback, offer prep, and the weekly recruiting digest all run on it. Team seats are $20/month standard and $100/month premium on annual billing ($25 and $125 billed monthly) for teams of 2 to 150, and seat types mix within one team; Enterprise is a seat price plus usage at API rates. It is the connective AI that reads the other tools’ output and turns it into drafts a recruiter edits instead of writes.
The handoffs that make it a stack
The value is not six tools; it is the chain between them. A sourcer runs a plain-English search in juicebox, then pushes the matched profiles into a Gem project. When a Gem sequence gets a reply, the recruiter converts that prospect into an Ashby candidate — Gem’s sync dedupes against the ATS so nobody double-tracks. When Ashby advances a candidate to the onsite stage, ModernLoop picks up the loop and books the panel from live interviewer availability. The interview runs on whichever of Zoom, Teams, or Meet the panel already uses, BrightHire captures and structures it, and the debrief note lands back on the Ashby scorecard. Finally, Claude reads those scorecards and BrightHire notes and drafts the debrief summary, the rejection note, or the offer packet. Each arrow is an event that triggers the next action, which is why the stack compounds instead of just co-existing.
Why this combination
Pre-AI recruiting stacks at this scale were Greenhouse + LinkedIn Recruiter + Calendly + Zoom + maybe Gem — five tools that did their individual jobs but did not compound. The AI layer (juicebox + BrightHire + Claude) inverts that: each piece extends the recruiter’s reach, and the data moving between them surfaces insight no single tool holds. Trying to collapse this into one all-in-one platform (Greenhouse Enterprise, SmartRecruiters Talent Cloud) gives up the specialist depth — juicebox’s sourcing recall, BrightHire’s interview analysis — that is the whole reason to run the stack.
The structural change to plan for in 2026 is that the AI in this stack is metered, not bundled. Ashby meters credits, ModernLoop meters credits per completed interview, Gem gates its agents behind a costlier seat type, and juicebox charges per always-on agent. Seat count no longer predicts the bill; volume of AI actions does. Price the meters against your actual req volume before signing, and re-forecast at the midpoint of the term — a hiring spike now raises the software bill in the same quarter it raises the workload.
What it costs
Budget roughly $80K-$300K a year all-in for a 100-500 employee company hiring 50-150 roles, with a 3-8 seat recruiting team. Split it in two. The pieces you can price without a sales call run about $15K-40K a year: juicebox Growth at $159/seat/month across 2-4 sourcer seats plus the two-agent add-on at $300/month, ModernLoop’s Taylor AI at $450-$1,250/month, and Claude Team at $20-100/seat/month depending on how many recruiters need the premium seat. Everything else is quoted — Ashby Plus or Enterprise, Gem, and BrightHire — and those three carry most of the bill. BrightHire publishes no price at all; third-party marketplace data clusters it between roughly $15K and $100K+ a year and it scales on interview volume rather than seats, so treat that as a range to negotiate against rather than a price.
The costs that miss the spreadsheet are the Gem-to-Ashby sync setup, the recruiter ramp on juicebox query patterns, and mid-term credit true-ups when a hiring push burns the AI allowance early. The stack earns its cost only if it retires headcount you would otherwise hire: at 50+ roles a year, ModernLoop alone offsets the 1-2 coordinators it replaces.
Common variations
- Earlier-stage variant. Under 100 employees, Ashby Foundations at $400/month flat covers the ATS, and Gem All-in-One replaces both the CRM and the ATS if you would rather run one vendor — companies under 30 employees get it free for six months, then at half price. Drop ModernLoop and use native scheduling. Cuts cost 60-70% at that stage. Swap back the moment coordinator time becomes the bottleneck or you need Ashby’s analytics depth.
- Enterprise variant. Replace Ashby with Greenhouse Enterprise or SmartRecruiters; add Beamery or Phenom for talent CRM at enterprise scale; add HireVue for high-volume video screening. The trigger is compliance and req-volume complexity outrunning what Ashby’s workflow engine handles cleanly.
- Engineering-heavy hiring. Add HackerRank or CodeSignal for technical assessment; add Holly for autonomous engineering sourcing when recruiter capacity is the constraint.
- Autonomous-sourcing variant. If continuous, hands-off sourcing matters more than a sourcer-driven search UI, swap or supplement juicebox with SeekOut or GoPerfect (Pin is a lower-cost entrant in the same category). Choose this when the constraint is sourcer hours, not search quality.
- AI-screening-first variant. If phone screens are the bottleneck rather than scheduling, buy the AI interviewer directly — BrightHire Screen alongside Interview Intelligence, or ModernLoop’s Taylor at 5 credits per completed interview. Choose by where the evidence has to land: BrightHire Screen keeps it in the same record as the human interviews, Taylor keeps it in the scheduling flow.
- High-volume hourly hiring. A different category entirely — see Fountain plus Harver plus a corporate ATS for the salaried layer above the hourly-hiring stack.
When this stack is the right pick — and when it isn’t
Pick this stack when you are a 100-1,000 employee tech company, hiring 50-200 mostly-professional roles a year, with a recruiting team too small to absorb the coordination load by hand and enough hiring-manager discipline to act on structured evidence. That is its center of gravity.
Do not pick it if any of these is true. Under ~50 hires a year, the six-tool overhead and quote-based contracts cost more than they return — run the earlier-stage variant. If your volume is hourly or frontline, this is the wrong category; go to Fountain-class tooling. If your hiring managers will not run structured debriefs, BrightHire and Ashby scorecards become expensive recording devices, so fix the process before buying the tools. And if procurement forces a single-vendor suite, accept the all-in-one and its shallower AI rather than fighting for six contracts you cannot win.
What this stack does NOT replace
- A real recruitment marketing function — career site, content, and employer-brand work that drives inbound application flow.
- A formal structured interviewing discipline — the tools enable it but do not enforce it; manager and interviewer training still required.
- A diversity recruiting program — the tools surface candidates and signals, but the strategic and operational work stays human.
- A bias-audit obligation. NYC Local Law 144 and Illinois HB 3773 are in force today; the EU AI Act’s Annex III high-risk duties, which cover hiring, were deferred to December 2027. Buying AI tools adds the obligation; it does not satisfy it.