Durability Ranking: BPO and Customer Operations Firms
Of 171 BPO and customer operations firms assessed in July 2026, only 3 rated Hard to displace while 100 sit losing its protection. Full ranking and watch-items.
Customer operations is one of the most exposed corners of our assessment. Of the 10,525 companies assessed across all sectors in July 2026, 171 fall into the BPO and customer operations vertical, and their numbers are stark: only 3 of the 171 earned the top standing of Hard to displace, while 100 landed losing its protection and another 33 were rated little protection left. Most of the group shows real exposure to AI-driven pricing and volume erosion, and 52 of the 171 companies were flagged among the most exposed in the category. We never declare an outcome for any company; what follows is what the July 2026 evidence shows, plus dated items you can verify yourself.
How this ranking works
Each standing here is built from the same evidence cited throughout this piece: named customer moves, market-share figures, product shutdowns, and financial disclosures. To see how much weight these standings deserve, we checked the approach against outcomes we already know happened: looking only at information available 18 months before each company's failure, it flagged 90% of 116 companies that later shut down as troubled ahead of time, and did not clear any of them as safe. Companies still operating at that same 18-month mark were flagged as troubled only 17% of the time by comparison.
The three that stand apart
Only Genesys, Bandwidth, and RAKUS Co., Ltd. earned Hard to displace among the 171 BPO and customer operations firms assessed in July 2026, and all three face far less pricing and volume pressure than the category as a whole. The common thread is control of the resolution or the infrastructure, not the seat.
Genesys runs the cloud contact center software large enterprises already depend on; the pressure it faces is mostly on new deals, not existing accounts. Bandwidth owns the regulated carrier network that voice and AI-voice applications have to route through, with almost no visible customer loss in the July 2026 evidence. RAKUS serves Japanese SMBs with its Chatplus and Raku Raku Seisan tools, with low competitive pressure and no confirmed customer defections, though a domestic AI rival is named as one to watch.
Just below that tier, seven companies rated defensible for now, including Tanla Platforms, which holds its position on its Karix and Trubloq messaging routes; Tencent Qidian, embedded directly into WeChat and QQ; Zoho Desk; and Transcosmos, on its primary listing in the data.
The squeezed middle: seat-billed and message-billed models
The 100 companies losing its protection share a business model problem: they charge by the human hour, the seat, or the message sent, and AI agents are quietly deleting all three units. This group spans the whole stack.
Helpdesk and contact center software: Zendesk, Freshworks, Five9, Pegasystems, and Intercom all rated losing its protection in the July 2026 assessment, with Salesforce Agentforce and the $3.6B Intercom/Fin acquisition pulling customer-service budget toward a single CRM platform instead of standalone tools. RingCentral, Talkdesk, 8x8, and Zoom Contact Center sit in the same band.
Messaging middlemen: Infobip, Sinch, MessageMedia, and MessageBird face enterprises routing traffic directly through AWS End User Messaging, Azure Communication Services, and Meta's WhatsApp Business Platform rather than paying a CPaaS layer. The WhatsApp inbox tier, Wati, SleekFlow, Omnichat, and respond.io, all run the same Meta approval process with near-identical chatbot and broadcast features, which turns competition into a price war with no real separation.
Human-labor BPO: Transcosmos, Bell24, TaskUs, Sutherland, ResultsCX, Startek, IBEX, and Infosys BPM all rated losing its protection. Transcosmos's own guidance, per the July 2026 assessment, shows profit shrinking even as revenue grows, which is what the seat-billed squeeze looks like on a financial statement.
The little protection left tier and the shutdown pattern
Of the 171 companies in this vertical assessed in July 2026, 33 rated Little protection left, the weakest standing, and the list includes some once-prominent names: Avaya, Mitel, Drift, Kayako, and Delighted, which the July 2026 data flagged as facing the most severe pressure of any company in the category.
Two distinct failure patterns show up here. The first is legacy on-premise erosion. Per the July 2026 assessment, Avaya's on-premise contact center base is shrinking about 20% a year while NICE and Genesys hold roughly 22% and 20% of the cloud contact center market respectively, and Mitel's cloud share fell from 17% to 15.5% as RingCentral absorbed its cloud base. Kayako shrank from 131,000 customers to roughly 300 as Zendesk, Freshdesk, and Intercom took the market it once held.
The second pattern is acquire-and-shut-down. Delighted's owner Qualtrics closed the product entirely, sending every customer to migrate elsewhere. Drift's owners, Clari and Salesloft, are sunsetting the product and pointing customers to a different company, 1mind, with Qualified, Warmly, and Abmatic positioned to pick up leftover accounts. Playvox, Espressive, and Simplr, all rated Little protection left, face the same dynamic: acquirers folding them into larger AI-agent product lines rather than growing the brand.
The dividing line
The sector-level pattern in the July 2026 data is simple. Companies that control the resolution itself, or own regulated infrastructure the AI stack must route through, hold their standing. Companies that route work to humans or to third-party APIs get squeezed from both ends: cloud giants bundle the tooling for free, and AI-native rivals like Sierra and Decagon sell the automated outcome directly. In Asia, the same bundling force applies: Kakao Enterprise, NetEase Qiyu, and RongCloud, all rated protection under test or losing its protection, face Naver Cloud, Alibaba Cloud, Tencent Cloud, and Samsung SDS packing AI customer-service tools cheaply into wider cloud contracts.
Questions this data answers
How many BPO and customer operations firms rated well in July 2026? Only 3 of the 171 companies in the vertical earned Hard to displace: Genesys, Bandwidth, and RAKUS. Another 7 rated defensible for now, meaning fewer than 6% of assessed firms reached the top two standings.
Is Zendesk in trouble? Zendesk rated losing its protection in the July 2026 assessment. It is holding customers for now, but Salesforce's Intercom/Fin acquisition and AI-native rivals like Sierra and Decagon pressure its per-seat pricing model.
Why did Drift and Delighted shut down? Both were acquired and then closed rather than grown. Delighted's owner Qualtrics shut it down completely, and Drift's owners Clari and Salesloft are sunsetting it and directing customers to 1mind, part of a broader pattern where acquirers fold CX brands into larger AI platforms.
Which companies face the most pressure from AI in this sector? Per the July 2026 assessment, seat-billed and message-billed firms face the most: Avaya, Mitel, Kayako, Whispir, and Simplr all combine heavy AI-driven pressure with weak differentiation against AI-native or cloud-bundled rivals.
Is Avaya shutting down? No shutdown has occurred, but the July 2026 assessment records its on-premise contact center base shrinking about 20% a year while NICE and Genesys grow cloud share. That is an erosion pattern to watch, not a verdict.
Watch-items (dated, verify yourself)
- By June 2027: whether Transcosmos reports FY2027 ordinary profit below FY2026 levels despite revenue growth, as its own guidance flagged.
- By December 2027: whether Avaya's on-premise base shrinks another 15-20% and NICE plus Genesys push their combined cloud CCaaS share past 45%.
- By June 2027: whether at least two more standalone helpdesk or CX chat products are shut down or folded into a parent's product line, following the Drift and Delighted pattern.
- Ongoing: whether Salesforce or Microsoft begin publishing AI-agent ticket resolution rates instead of seat counts in earnings, which would mark the industry's official shift from selling human seats to selling automated outcomes.
Assessments reflect evidence available July 2026. Companies change; treat watch-items as things to verify, not conclusions.
Figures are as of each company's assessment date unless dated otherwise.