Max Ball, the Forrester analyst who has now run this evaluation twice, opened his commentary on the current edition with a warning about how it would be read. The familiar vendor list, he noted, could make it look as though little had changed. Looks can be deceiving.
His summary of what actually shifted in two years is worth taking at face value: generative AI is reshaping customer service, adjacent vendors are working to commoditise CCaaS, and the CCaaS vendors continue to expand their value proposition.
The second of those three is the one that should worry anyone in this market. Commoditise is a strong word for an analyst to use about an established category, and it describes a threat coming from outside the vendor list entirely.
What the platform replaced
The contact centre used to be hardware. Telephony switches, interactive voice response systems, and call distribution equipment sat in a room, bought as capital expenditure on a multi-year refresh cycle, sized for peak volume that occurred a few days a year.
CCaaS moved that to subscription software. The core is still routing, meaning getting a contact to an appropriate agent, but the modern platform bundles considerably more: omnichannel handling across voice, chat, email, messaging and social; self-service and voice bots; the agent desktop where all of it converges; workforce management for forecasting and scheduling; quality management and recording; analytics; and outbound dialling.
The criteria in Forrester's current evaluation describe where the competition actually sits. AI architecture, generative AI and large language model support, AI orchestration, omnichannel routing and experiences, customer self-service, agent assist tools, knowledge management, agent desktop and workflow automation, supporting services, innovation, roadmap, and pricing flexibility and transparency.
Routing appears once. Everything else is about intelligence, assistance, and the surrounding operation.
Inside The Forrester Wave: Contact-Center-As-A-Service Platforms, Q2 2025
Published in April 2025 and authored by Max Ball, the evaluation scored ten providers on current offering, strategy, and customer feedback, sorting them into Leaders, Strong Performers, and Contenders.
NiCE placed as a Leader with the highest score in the strategy category, described by Forrester as a best fit for brands wanting a proven solution with broad capabilities and a strong track record.
Amazon Web Services placed as a Leader with Amazon Connect, taking the highest possible score in AI architecture, generative AI and large language model support, agent desktop and workflow automation, innovation, roadmap, and pricing flexibility and transparency.
That AWS scorecard is the most interesting result in the evaluation, and not because of the AI criteria. Pricing flexibility and transparency is where a consumption-priced cloud service beats a seat-licensed platform structurally rather than competitively, and it points at a tension the whole category is carrying.
Among Strong Performers, Sprinklr took the highest possible score in seven criteria including omnichannel routing and experiences, AI orchestration, generative AI support, customer self-service, agent assist tools, knowledge management, and supporting services, positioned as a fit for digital-first brands. Forrester's characterisation noted that Sprinklr arrived at customer service from marketing, sales, and social listening.
8x8 also placed as a Strong Performer. Genesys and Five9 were also among the ten evaluated.
The predecessor, The Forrester Wave: Contact Center As A Service, Q1 2023, also authored by Max Ball, evaluated eleven providers against thirty four criteria. LiveVox was included in that edition, credited for an easy-to-administer platform with an internal CRM making customer data available for routing and the agent desktop, and positioned for North American brands with small to midsized contact centres.
Eleven vendors in 2023, ten in 2025, and thirty four criteria down to a leaner set. Stable membership, converging criteria, and all the movement happening inside the products rather than across the list.
Who is doing the commoditising
The threat Ball identified comes from two directions, and neither appears in the vendor list.
The cloud platforms are the first. Amazon, Microsoft, and Google are all deepening their presence, either through direct offerings or by distributing CCaaS software through their marketplaces. AWS placing as a Leader with Amazon Connect is the clearest evidence, and its pricing scores explain the mechanism.
Traditional CCaaS is licensed per seat per month. A contact centre with eight hundred agents pays for eight hundred seats regardless of whether volume is seasonal, and regardless of how many contacts the agents actually handle. Consumption pricing charges for what is used.
That difference has always favoured consumption models for volatile workloads. It becomes much sharper when the number of human agents starts falling because automation is handling more contacts, which is precisely what every vendor in this category is now promising to deliver.
The second direction is the CRM vendors, who own the customer record and the surrounding service workflow and can argue that voice is a channel rather than a platform. Functionality is moving into CRM and cloud-native environments, which changes how organisations approach the architecture question rather than just the vendor question.
The contradiction in the pitch
Here is the structural problem underneath this category, and it is worth stating plainly because no vendor will.
Forrester's own framing is that today's contact centre needs AI because increasing the efficiency of hundreds and thousands of human agents delivers compelling return on investment.
That is the pitch, and it is sound. It is also a pitch to reduce the number of agents, sold by vendors who charge per agent seat.
Every CCaaS vendor is therefore selling technology that shrinks its own billable base. The commercial responses are visible across the market: pricing that shifts toward consumption, per-resolution or per-interaction models for automated handling, and bundling that moves value into workforce management, analytics, and orchestration where the seat count matters less.
For a buyer this is leverage, provided you understand the arithmetic. If a vendor projects a thirty percent deflection rate for your volume, ask what your bill looks like at that deflection rate. If the answer is unchanged, the savings are accruing somewhere other than your budget.
Suite breadth versus component maturity
Forrester attached a specific caution to this evaluation that buyers should carry into every demonstration: broad CCaaS systems simplify management, but the newer applications within them may need to mature.
That is a precise description of the trade-off in a consolidated platform. Buying routing, workforce management, quality management, analytics, and AI from one vendor means one contract, one integration surface, one support relationship, and one place where all the data sits. It also means accepting that some of those modules are three years old and competing against specialists who have done nothing else for a decade.
Workforce management is the usual example. It is a genuinely hard discipline involving forecasting, scheduling, adherence, and intraday management, and the specialists are meaningfully better at it than most bundled modules. Whether that gap matters depends entirely on the size and complexity of your operation.
The practical approach is to identify which one or two capabilities actually determine your operating cost, evaluate those against specialists, and accept bundled adequacy on the rest. What you should not do is assume that a Leader placement means every module is a leading module. Forrester is explicitly saying it does not.
The category next door
This sits alongside Forrester's digital customer interaction solutions category, and the boundary is worth understanding.
CCaaS is platform-first and voice-heritage. Its architecture assumes a routing engine at the centre, with digital channels added around it. Digital customer interaction solutions come from the opposite direction, treating asynchronous messaging as the primary mode and voice as one channel among several.
Sprinklr appears in both evaluations, which illustrates the convergence. A vendor from social and digital heritage placing well in a voice-heritage category, while voice-heritage vendors build out digital, means the two categories are approaching each other from opposite ends.
The distinction that survives is architectural. If most of your volume is voice and the operational discipline is queue management, CCaaS is your category. If most of your volume is asynchronous and conversations persist across days, the other one is. Buying from the wrong lineage produces a platform that works while feeling permanently slightly wrong.
What to test
Model your bill under the automation you are being promised. This is the single most valuable exercise in a CCaaS evaluation and almost nobody does it. Take the deflection rate the vendor projects, apply it to your volume, and ask what you pay in year three.
Test the agent desktop with agents handling real concurrency. The desktop is where every capability converges and where the operating cost is actually determined. A manager's opinion of it is not the relevant opinion.
Evaluate the one module that matters most against a specialist. If workforce management drives your costs, benchmark the bundled module against a dedicated vendor and decide the trade-off deliberately.
Ask what happens to conversation history across channels and across time. Voice-heritage platforms treat interactions as discrete events with a beginning and an end, which is exactly wrong for asynchronous messaging.
Probe the AI claims for maturity rather than capability. Given Forrester's caution about newer applications, ask how long each AI feature has been generally available, how many customers run it in production, and at what volume.
And check the integration to the systems your agents actually use. The routing platform is rarely where a contact centre's difficulty lives. The difficulty lives in the four other systems an agent alt-tabs between while a customer waits.
Analyst Source
Forrester Research
Category definition, vendor inclusion, and evaluation findings in this article draw on Forrester's coverage of contact-center-as-a-service platforms, led by principal analyst Max Ball. The Q2 2025 Wave scored 10 providers on current offering, strategy, and customer feedback; the Q1 2023 edition evaluated 11 providers against 34 criteria.
Source research
- CCaaS Vendors Thrive In A Wild Market
- The Forrester Wave: Contact-Center-As-A-Service Platforms, Q2 2025
- The Forrester Wave: Contact Center As A Service, Q1 2023
Forrester does not endorse any vendor named here, and tier placement should not be read as a recommendation to buy.