The per-seat number on a pricing page is the least useful figure in the decision, and it is the only one most teams compare. It tells you what one seat costs on one tier. It does not tell you which tier your requirements will put you on, or how the AI is billed on top. It does not tell you how many of your seats will still be in daily use a year from now.
We build Sektra, so weigh our own section accordingly. Everything else here is arithmetic, and arithmetic is checkable. Prices are the published list rates at the time of writing, on annual billing. Every vendor on this list has changed them before and will change them again.
What do shared inbox tools cost in 2026?
Published per-seat pricing, billed annually, cheapest paid tier to most expensive standard tier. Enterprise tiers are quote-only almost everywhere and are left out.
- Drag: about $12 to $24 per seat per month. A Gmail extension, board-based, the lowest entry price in the category.
- Keeping: about $12 to $20 per seat per month. Support-shaped, lives inside Gmail, the lower tier caps agent count.
- Missive: about $14 to $36 per seat per month. Its own app, email and chat combined, free for very small teams.
- Hiver: a free plan, then about $25 to $85 per seat per month. Ticket workflow inside the Gmail interface, with SLAs and analytics from the middle tier up.
- Sektra: $29 per seat annually on Standard and $59 on Pro. Every AI feature, relationship memory included, is in both tiers at no extra cost.
- Gmelius: about $19 to $40 per seat per month. Shared inbox plus sequences and automation, aimed partly at outreach teams.
- Help Scout: about $20 to $65 per seat per month, with contact volume affecting the bill on some plans.
- Front: $25 per seat on Starter, $65 on Professional, $105 on Enterprise. Starter is single-channel and caps at 10 seats.
Two things to notice before reading anything into the order. Monthly billing runs roughly 20 to 25 percent above these figures nearly everywhere. A team that avoids an annual commitment is choosing to pay a quarter more for the privilege. And the spread within a single vendor is often wider than the spread between vendors. Front's lineup ranges from $25 to $105, so which plan you land on matters more than which logo you pick.
The four costs that do not appear on the pricing page
1. The tier you land on
The left column is the number that gets quoted in the meeting, and it is often not the column you end up in. Entry tiers are scoped on purpose. They are built for smaller teams with simpler needs, which is why they carry seat caps, a single channel, or a shorter integration list.
That is how the tiers are drawn. It just means the feature that made you start shopping is often one tier to the right of the price you quoted.
So price the tier that contains your requirement rather than the tier that contains your budget. If routing rules are why you are here, find the row where routing rules appear and read the number above that column. The difference is rarely small. On Front it is the step from Starter at $25 to Professional at $65, which on twelve seats is about $5,800 a year.
Professional carries more, including omnichannel, analytics, and a higher seat ceiling. The number is only a problem if you are not going to use what it buys. The Front pricing breakdown walks through every tier and add-on with worked examples.
2. AI priced as a separate appetite
This is the newest line item in the category and the one most likely to move between quote and renewal. Across the market, AI arrives in three shapes. It is bundled into an upper tier, sold as a per-seat add-on, or metered by use. Metering ties spend to usage, so the bill moves with your volume.
The tradeoff is that a metered line is a forecast rather than a fixed number, and forecasting your own inbound is harder than it sounds.
Our own model, stated plainly since we are on the list. Every AI feature is in every Sektra plan, because a tool whose point is memory should not sell memory as an upgrade. The same reasoning applies to summaries, priority and tagging.
AI credits are pooled across the team rather than metered per person. That makes the bill a fixed number you choose in advance rather than a variable you discover at renewal. One heavy assistant user never creates a line item their colleagues have to explain.
3. The seats that never adopt
The most expensive line in a shared inbox budget is usually a seat that is paid for in full and used lightly. It does not appear as waste in any report. It appears as a full-price seat, invoiced annually, belonging to someone who kept working in Gmail and forwards things to a colleague instead.
This cost tracks one variable more than any other, which is whether the tool asks people to work somewhere new. Tools with their own workspace, Front and Missive among them, put email, coordination, and in Missive's case chat in one window. That only pays off once the whole team has moved in.
The question is whether your particular team will move, and that is answerable from evidence rather than optimism. Ask how many tools this team has been rolled onto before and how many of those rollouts fully took. That number predicts your effective per-seat cost better than anything on the pricing page. Sektra syncs with Gmail, so a teammate who has not moved yet still works from the inbox they already open.
4. The rollout
Workflow engines have to be configured before they pay off, and configuration is somebody's week. A tool with deep routing, SLA policies, and multi-team escalation carries a setup cost on top of the license. Someone senior tunes it and someone keeps maintaining it afterward.
Teams whose inbound moves between departments are the ones that investment is scoped for. For a team of eight working a single queue, the same investment mostly sits idle. Sektra's assignments, notes and client record work from the first sign-in, with no workflow to design.
The same team, priced three ways
Take a ten-person client-facing team, annual billing, and price the plan they would need rather than the cheapest one on offer.
- Front Professional at $65 a seat: $650 a month, $7,800 a year. Starter would be $3,000, but it is single-channel and caps at exactly 10 seats, so the eleventh hire forces the jump.
- Help Scout mid-tier around $40 a seat: roughly $4,800 a year, with contact volume a variable to watch on some plans.
- Hiver mid-tier around $55 a seat: roughly $6,600 a year, which is the tier where SLAs and analytics begin.
- Sektra Pro at $59 a seat: $590 a month, $7,080 a year. Standard at $29 caps at 5 seats, so a team of ten is on Pro.
- Drag or Keeping near the bottom of their ranges: roughly $1,440 to $2,400 a year, for shared labels and assignment inside a Gmail inbox.
Sektra is not the lowest number on that table. Drag and Keeping list lower, and their price covers shared labels and assignment inside Gmail. Sektra covers those same essentials and adds the client record, the assistant and the quiet-account flags in the plan price. Sektra Pro lands within a few hundred dollars a year of Front Professional.
What separates the two is what the seat price includes. Front's Professional price covers routing rules, SLA policies and cross-team escalation, with AI sold as an add-on below Enterprise. Every Sektra plan carries the full AI layer at no extra charge rather than reserving it for an upper tier or an add-on. Each price on that list buys a different job, and the expensive mistake is paying for a job you do not have.
Where Sektra sits, and where it does not
Sektra is a shared inbox that syncs two-way with your team's Gmail, so nothing migrates and the mail keeps living in Gmail. Setup is a Google sign-in of about 90 seconds. You get the collaboration essentials, including assignments, mentions, internal notes, and collision detection so two people never answer the same client.
What justifies paying above the floor of this market is the layer underneath, so here is what that layer does. Sektra reads the Gmail history you already have and maintains a live record for every client. It holds each commitment made in either direction and when it comes due. It holds each client question still waiting on an answer, including the ones buried inside threads that already got a reply.
Ask the assistant to watch an account and it alerts you the day that account has been quiet longer than you allowed. The assistant answers any question about any client from that history, so where do we stand with this account becomes a question for the system instead of a colleague. Because the record is read backward through existing mail, it is populated on day one, including relationships years older than the tool.
Sektra is not a workflow engine. Routing across five departments with SLA escalation policies is what Front's Professional tier is scoped for, and it needs someone to configure and maintain it.
Sektra is built for the team whose inbound is named clients, where the costly miss is a promise nobody tracked. Full row-by-row breakdowns live on the comparison pages, and the pricing page has the current plan details in case this article has aged since publication.
How to price your own decision
- Find the feature you are buying, then find the tier that contains it. Price that tier. Ignore the entry column.
- Multiply by the seat count you will have in eighteen months rather than today, and check the seat cap on that tier before you do.
- Ask how AI is billed, in writing: bundled, per seat, or per use. Metered pricing is not automatically worse, but it is a forecast rather than a number.
- Estimate the seats that will be lightly used. If the tool asks people to work somewhere new, weigh your team's track record. Multiply your per-seat rate by that count and treat the result as part of the price.
- Add the rollout. If someone has to design workflows before the tool earns its keep, that is a week of salary in the first year and a recurring tax after.
- Then run a two-week trial on real mail before signing anything annual. On Gmail this costs you nothing, because your mail is not moving either way.
The cheapest tool is the one that solves the job you have. Price the job first, and the column you should be reading stops being a judgment call.
