Kustomer Pricing: Concierge and Envoy Rose 25% After the Prices Vanished

Kustomer now charges $0.75 per engaged conversation for its customer-facing AI agent, Concierge, and $50 per user per month for its rep-side agent, Envoy. Both are exactly 25% above the $0.60 and $40 it published until October 2026. Kustomer pricing did not move in one step, though. For about a week in between, the AI agent rows on its pricing page carried a dollar sign and no number.

Short answer

Did Kustomer raise its AI agent prices? Yes. Its pricing details page, last updated 12 October 2026, lists Concierge at $0.75 per engaged conversation and Envoy at $50 per user per month. The previous published rates were $0.60 and $40. Seat prices for Enterprise and Ultimate remain unpublished and require a sales call.

Kustomer pricing before and after, line by line

The table below rebuilds the page in three states. The first column is the version Kustomer dated 19 December 2025, as a pricing tracker captured it on 5 October 2026. The middle column is the version dated 9 October. The last column is the page as it reads today, dated 12 October, which we opened directly on 4 November 2026.

Every figure in the first two columns comes from UsagePricing's dated captures. Every figure in the last column comes from Kustomer's live page. Where a row did not exist in a version, the cell says so rather than guessing.

Product names are quoted as the page printed them in each version. The tracker also reports that the attachment storage allowance, "20 GB + 5 GB per full time user", left the plan matrix in the October versions. Data storage beyond the allowance stayed at $50 per GB per month.

What moved and what did not

Two prices rose. One price disappeared and came back unchanged. Three new usage meters arrived. The seat price, which is the largest part of most Kustomer contracts, was never on the page in any version.

That last point frames everything else. Kustomer's main pricing page says plainly that "we don't post one-size-fits-all pricing." The AI agent rates were the only part of a typical Kustomer bill a buyer could check without a sales call. In October those were the rows that went blank.

The HIPAA row is the useful control

HIPAA compliance was blanked alongside the two AI rows and returned at the same $25. That matters for reading intent. If the blank had been a deliberate move to stop publishing AI prices, HIPAA would have no reason to vanish with them.

The simpler reading is a page rebuild in progress. All three rows were rewritten together, and only two of them came back at a new number. That is less dramatic than "Kustomer hid its prices", and I think it is the more likely story.

Three versions of one Kustomer pricing page in ten months

The order of events is what turns a routine price rise into a question about disclosure. Laid out in sequence, the rename, the blank and the new rates arrive months apart and then days apart.

The date on the page does not match the captures

The current page says "Last Updated: October 12, 2026". The tracker's capture from 15 October still showed the older 9 October version with blank rows. Both cannot describe the same live page on the same day.

There are innocent explanations. A cached copy, a staged publish, or a date typed when the edit was drafted rather than shipped would all produce this. The Usage Credits PDF linked from the page carries the document title "Kustomer Usage Credit Details Page - 10.6.26" and a creation date of 12 October. That suggests the change was being drafted from 6 October.

The practical lesson for a buyer is narrow. A vendor's "Last Updated" stamp tells you when someone edited a field. It does not tell you what you were quoted, or when the price you were quoted stopped being the public one.

Checking the 25% Kustomer price increase claim

The tracker's headline was that Kustomer "re-prices AI agents at $0.75 per conversation and $50 per rep," and that both rates are 25% higher. We checked the arithmetic and the source rather than the summary.

The arithmetic is exact. $0.75 divided by $0.60 is 1.25. $50 divided by $40 is 1.25. There is no rounding in either direction, and the same rates apply on both the Enterprise and Ultimate plans.

The source holds up on the current side. We opened Kustomer's comprehensive pricing details page on 4 November 2026 and it reads "$0.75 per engaged conversation" for Concierge and "$50 per user / month" for Envoy. The prior side rests on the tracker's 5 October capture, which we could not independently reopen. That is stated again in the limitations section below.

One correction to the tracker's framing

The tracker summarised the change as "AI agents re-priced 25% higher as Concierge and Envoy." That wording implies the rename and the repricing were one event. They were not.

The price is verified. The implied causation is not. A buyer reading "re-priced as Concierge and Envoy" might assume new products arrived at a new price. The products were renamed in June and repriced in October.

The Kustomer Concierge rename came four months before the price

Kustomer's own blog dated 1 June 2026 states it directly: "Envoy is the new name for AI for Reps. Concierge is the new name for AI for Customers". Kustomer's product updates log repeats it under a June 2026 entry titled "AI Products Renamed: Envoy, Concierge & Architect".

So from June to early October, the products had new names and the pricing page still used the old ones at the old rates. The October rewrite brought the page into line with the June names, and the higher prices arrived in the same edit.

Bundling a rename with a rise is a known pattern

A rename gives a vendor a natural moment to reprint a price. Buyers compare a new name against nothing, so the increase reads less like a change. This is not a Kustomer invention. It is one of the oldest moves in software packaging, and it works because comparison requires memory.

My view is that this case is mild. Kustomer kept the old name in brackets on Envoy, "Envoy (AI Agents for Reps)", which makes the comparison easier, not harder. A vendor trying to obscure a rise would not leave the trail visible on its own page.

The Envoy inclusion claim needs reconciling

Kustomer's blog of 5 June 2026 says "Kustomer Envoy is now part of all Kustomer AI pricing plans". The pricing page lists Envoy as an add-on at $50 per user per month on both current plans.

Those two statements can both be true if "part of" means "available on" rather than "included in". They cannot both be true if it means included at no extra charge. If you were sold Envoy on the strength of that June post, get the inclusion written into your order form. It is the cheapest clause you will ever negotiate.

What a bare dollar sign on a pricing page signals

A blank price is information. It is just not the information the buyer wanted. There are three readings of a "$" with no number, and they lead to different actions.

The first reading is a rebuild in flight. Rows get cleared while new copy is approved, and someone publishes the template early. The HIPAA row returning unchanged supports this reading for Kustomer.

The second reading is a move to quote-only pricing. The vendor stops publishing so that each deal can be priced against the buyer's alternatives. Google's split between listed seats and sales-priced agent work in Gemini Enterprise is the clearest current example of that path, and Kustomer was already most of the way there with seats.

The third reading is a price change waiting for a date. The number is decided but cannot go live until existing customers are told. A blank row is the safest placeholder because it cannot be quoted back at the vendor.

Why the blank matters more than the rise

A 25% rise on a published rate is a fact you can budget around. A blank rate is uncertainty, and uncertainty is the expensive part of any procurement. Your finance team can absorb a known increase. It cannot forecast against an empty cell.

I would rather a vendor publish a higher price than publish no price. Kustomer ended up doing the first, which is the better outcome. The week in between is the part worth remembering at renewal, because it shows the vendor can take the number down when it chooses.

This also fits a wider pattern. When we checked flagged price moves across the market, none of eight tracker-reported AI price changes held up exactly as first reported. Kustomer's is unusual in that the headline number is exact. The narrative around it is where the error crept in.

What the Kustomer price increase costs a support team

The rise is a flat 25% on both meters, so the percentage is never in doubt. What changes is the absolute amount, and that depends on volume. The table uses three illustrative team sizes. They are scenarios, not customer data.

Zan Digital calculation at list rates. Excludes Enterprise or Ultimate seat licences, which are not published, and any negotiated discount. Negotiated contracts may not move at all until renewal.

Concierge dominates the bill

At every volume above, the per-conversation meter is roughly 88% of the AI line. Envoy seats are the smaller number. That is the right place for your negotiation effort. A cap or a fixed rate on Concierge conversations is worth several times a discount on Envoy seats.

The meter also grows with success. Every conversation Concierge engages is billable, so the better it gets at engaging customers, the faster the bill climbs. That is the general trap described in the breakdown of how credits and tokens confuse AI billing, applied to a support desk.

Kustomer Concierge pricing against Intercom Fin's $0.99

Intercom's pricing page lists Fin at "$0.99 per outcome," charged "once per conversation". An outcome counts when a customer confirms the issue is resolved, when they do not ask for more help after Fin responds, or when Fin completes a workflow, including handoffs.

On the headline numbers, Concierge at $0.75 is 24% below Fin's list rate. Before October, at $0.60, it was 39% below. Kustomer's per-unit discount to the best-known competitor fell from $0.39 to $0.24. The full side-by-side of Fin against Zendesk and HubSpot is in the cost-per-resolution comparison.

The break-even rate moved from 61% to 76%

The units are the trap. Kustomer bills engaged conversations. Fin bills outcomes. An engaged conversation that ends with an angry customer still costs you $0.75 on Kustomer. On Fin, a conversation that produces no outcome is not billed.

That gives a simple test. Divide Kustomer's price by Fin's. If fewer engaged conversations than that ratio would have counted as a Fin outcome, Kustomer is the more expensive option per outcome despite the lower sticker.

At $0.60, Kustomer matched Fin per outcome if 60.6% of engaged conversations would have been Fin outcomes. At $0.75, the bar is 75.8%. Fin's own definition is generous, since a customer simply not asking for more help counts. Many support desks will clear 76% on that definition. Some will not, and they now pay more per outcome on Kustomer than the sticker suggests.

This is a calculation from list prices, not a measurement of anyone's deployment. Run it on your own conversation logs before you treat it as a verdict. The broader case for outcome meters over activity meters is made in the analysis of Fin's resolution pricing.

Where this analysis is weakest

Three parts of this argument rest on less than we would like, and one competing reading is stronger than the headline suggests.

First, the "before" column is a tracker's capture, not one we reopened. We tried to retrieve archived copies of Kustomer's page directly and could not reach the archive from our research environment. The $0.60 and $40 rates are corroborated by several earlier 2026 buyer guides, but those are Tier 3 sources that may themselves trace back to the same page. Treat the prior rates as well-supported, not independently verified by us.

Second, we do not know what an engaged conversation is. Kustomer's help article defining billable units did not render when we opened it. The break-even calculation assumes the unit is comparable enough to Fin's to divide one by the other. If Kustomer counts only conversations where the AI meaningfully participates, the comparison flatters Fin.

Third, list prices are not contract prices. Every Kustomer customer has a sales-negotiated seat price, and many will have negotiated AI rates too. A customer on a multi-year contract may see no change until renewal. The 25% applies to the public rate card and to new buyers, which is where it is most visible and possibly least material.

The competing reading is that this is simply catching up. Kustomer's AI rates had been flat since at least December 2025 while it added features through 2026. On that view, a 25% list rise after ten months, still landing below Fin, is ordinary pricing housekeeping. I find that reading reasonable. I still think the blank week deserves attention, because the useful signal is not the rise but the fact that the public number can disappear.

Frequently asked questions

Where to start before your Kustomer renewal

Pull one month of conversation logs and count two numbers: conversations your AI agent engaged, and conversations that ended without a further request for help. Divide the second by the first. If the answer is below 76%, you are paying more per useful outcome than the Fin list price, and you have a concrete number to bring to the renewal call.

Then ask for one clause, not a discount. Request that your per-conversation and per-seat AI rates be fixed for the contract term, with any change requiring 90 days' written notice. The October blank shows the public number can vanish for a week. A rate written into your order form cannot. The renewal tactics that work against this kind of rise are covered in the guide to grandfathered pricing at renewal.

Related reading