What actually happened in 2026
Two forces ran in parallel. Incumbents bought AI-agent capability they could not build fast enough, and AI-native startups raised enormous rounds to attack those same incumbents. The result is a smaller field of much larger players, almost all of whom now meter AI by the resolution.
| Date | Event | Type | Value |
|---|---|---|---|
| Oct 2024 | Crescendo hits $500M valuation, acquires outsourcer PartnerHero | Funding + acquisition | $50M round |
| Jul 28, 2025 | NiCE agrees to acquire Cognigy | Acquisition | ~$955M |
| Jan 15, 2026 | Parloa raises Series D | Funding | $350M at $3B |
| Jan 28, 2026 | Decagon raises Series D | Funding | $250M at $4.5B |
| Mar 11, 2026 | Zendesk acquires Forethought | Acquisition | Undisclosed (its biggest deal in 20 years) |
| May 4, 2026 | Sierra (Bret Taylor) raises Series E | Funding | $950M at ~$15.8B |
| May 12, 2026 | Intercom rebrands the company to Fin | Rebrand | Fin AI agent crossed $100M ARR |
| Jun 15, 2026 | Salesforce agrees to acquire Fin (formerly Intercom) | Acquisition | ~$3.6B |
The Salesforce deal is the clearest signal. Intercom renamed its entire corporate identity to Fin in May 2026, betting the company on its AI agent, and roughly a month later signed to sell to Salesforce for about $3.6 billion. Salesforce framed it as a complement to Agentforce, its agentic AI portfolio, and noted Fin brings more than 30,000 customers. The transaction is expected to close in Q4 of Salesforce's fiscal year 2027. We covered the parallel Zendesk move in detail in our Zendesk and Forethought acquisition breakdown.
There is a human side too. Salesforce eliminated roughly 4,000 customer support roles after AI agents began handling about half of its customer interactions, per CEO Marc Benioff. That is the demand side of why everyone is buying AI agents at once.
Why consolidation is a buyer's problem, not just a headline
Acquisitions are usually pitched as good news for customers. In support software they create four specific risks for SMB buyers.
1. Roadmap risk
When a startup gets absorbed into a platform giant, its roadmap stops serving its old customers and starts serving the acquirer's strategy. Forethought's self-improving agents now exist to accelerate Zendesk's Resolution Platform. Fin's "Apex" model now exists to feed Agentforce. If you bought Forethought or Intercom for a specific workflow, that workflow is now a line item inside someone else's three-year plan.
2. Pricing creep
This is the biggest one. The industry moved almost in unison to outcome-based or usage-based AI billing in 2025 and 2026. That sounds fair ("only pay when AI resolves a ticket") but it makes your bill a function of your ticket volume, which you do not fully control, and it stacks on top of per-seat fees you already pay.
3. Lock-in through integration depth
Once an AI agent is wired into your CRM, your knowledge base, and your billing, switching costs balloon. Acquirers know this. The deeper the integration, the more pricing power they have at renewal.
4. Forced migrations and sunset features
Mergers routinely kill overlapping products. Features get folded, renamed, or sunset. Zendesk reportedly folded its "Advanced AI" capabilities into Suite plans in May 2026 while keeping per-resolution overage. Reorganizations like that change what you pay for without you changing anything.
The pricing math nobody puts on the homepage
Most 2026 AI support pricing splits into three sub-models that vendors deliberately conflate. Precision matters because they bill very differently.
- Per resolution: you pay only when AI actually closes the ticket. Intercom/Fin charges $0.99 per outcome, Help Scout AI Answers $0.75 per resolution, Gorgias $0.90 to $1.00 per resolution.
- Per engaged conversation: you pay whenever AI engages, even if it immediately hands off to a human. Kustomer is about $0.60 per engaged conversation, Gladly's Sidekick about $0.60, Tidio's Lyro roughly $0.50 to $0.65.
- Per session: a time-windowed interaction. Freshdesk's Freddy runs roughly $0.10 to $0.49 per session, sold in prepaid packs that expire at the end of the billing cycle.
Now layer the seat fees back on. Here is what the headline plans look like as of mid-2026.
| Vendor | Base platform | AI billing | The trap |
|---|---|---|---|
| Intercom / Fin | Seats from $29/seat/mo (annual) | $0.99 per outcome, 50-outcome minimum | You pay seats AND per-resolution AI on the same conversation |
| Zendesk | Suite Team $55, Pro $115/agent/mo (annual) | Per automated resolution (third parties estimate ~$1.50 to $2.00); Copilot add-on $50/agent/mo | Per-resolution overage stacks on per-agent plans |
| Freshdesk | Pro $55/agent/mo (annual) | Freddy ~$0.10 to $0.49/session, prepaid packs | Unused sessions expire monthly |
| Help Scout | Plus $45/user/mo (annual) | AI Answers $0.75/resolution | Per-resolution after the free trial period ends |
| Gorgias | Billed by ticket volume, e.g. Pro $360/mo for 2,000 tickets | AI Agent $0.90 (annual)/$1.00 (monthly) per resolution | An AI resolution can also count as a billable ticket (double billing) |
| Zoho Desk | $7 to $40/agent/mo (annual) | Mostly bundled; autonomous resolution gated to $40 Enterprise tier | Best AI locked behind the top tier |
| Corebee | $99/mo flat, or $79/mo billed annually | Included | No per-seat, per-agent, or per-resolution fees |
Two patterns to internalize. First, Intercom and Gorgias stack both a seat or ticket fee and a per-resolution AI fee on the same interaction. Gorgias can bill a single automated resolution twice, once as a ticket and once as an AI resolution. Second, the "engaged conversation" vendors charge you even when the AI fails and dumps the customer on a human. You can read the full side-by-side in our pricing comparison and our breakdown of Freshdesk versus Zendesk pricing.
How Corebee is built differently
Corebee is AI-native customer support priced at $99 per month flat, or $79 per month billed annually ($948 per year). There is no per-seat fee, no per-agent fee, and no per-resolution fee. Your bill does not move when your ticket volume spikes during a launch or a holiday, which is exactly when usage-based vendors get expensive.
The product uses a 3-tier AI engine (Instant, Pattern, and Deep) that reaches up to 86% average auto-resolution across web chat, WhatsApp, and email, grounded in a RAG knowledge base that auto-learns from your docs. Setup is 5 minutes for the widget and 11 minutes for the full platform, with 30+ integrations. Corebee is GDPR compliant with SOC 2 in progress.
The strategic point for a consolidation-wary buyer is simpler than the feature list. A flat price means there is no metered dial for a future owner to turn up. You can compare the model directly on our alternatives page and the Intercom comparison.
A checklist for choosing AI support that survives consolidation
Run any vendor, including Corebee, through these questions before you sign.
- Is the pricing flat or metered? Metered pricing is the lever acquirers pull at renewal. A flat price caps your downside.
- Do you pay seats and AI usage on the same ticket? If yes, your effective cost per resolution is higher than the headline number.
- Are you charged when the AI fails? "Engaged conversation" and "session" models bill you even on escalations. Per-resolution and flat models do not.
- Can you export your knowledge base and conversation history cleanly? If the data does not come out in a standard format, you are locked in regardless of price.
- How deep is the integration coupling? Deeper coupling means higher switching cost and more pricing power for whoever owns the vendor next year.
- What is the contract length and the auto-renew clause? Avoid multi-year commitments to a vendor mid-acquisition. Annual or shorter keeps you mobile.
- Is there a real trial and a money-back guarantee? You want to verify resolution rates on your own tickets, not a demo. Corebee offers a 14-day free trial with no credit card and a 30-day money-back guarantee.
- Who owns the vendor today, and is a deal pending? A signed acquisition that has not closed is a roadmap question mark. Ask directly.
The deflection rate you actually achieve matters more than any list price, because it determines how many tickets a human ever touches. We go deeper on measuring it in our ticket deflection guide, and you can model your own numbers with the SLA calculator.
Where this goes next
Expect more consolidation. Salesforce's Fin purchase was its fifth announced acquisition of 2026 and its third in June alone. The AI-native startups raising at $3 billion to $15 billion valuations either get acquired or grow into acquirers themselves, and both outcomes reshuffle the products underneath your support stack. The defensive move for a buyer is not to predict which vendor wins. It is to choose a tool with predictable pricing, clean data export, and a short contract, so that whoever wins the consolidation game does not get to win at your expense.
If a flat $99 per month with up to 86% auto-resolution and no metered surprises fits how you want to budget, start a free trial and test it on your own tickets.