Conversation Intelligence For Customer Service

One Leader. Eleven vendors. Thirty two criteria.

The Forrester Wave: Conversation Intelligence For Customer Service, Q3 2023, is the rarest artifact an analyst firm can produce: a first-time evaluation of a new market with a single winner. Published in August 2023 and authored by principal analyst Christina McAllister, it remains, as of August 2026, the only edition ever published under this name. The successor Wave arrived under a different one, for contact centers, which is its own category and its own story.

This article is about the one-and-only edition: what its entry rules admitted, what its sole Leader proves about first editions, and what the scorecard is still worth now that the category has moved on.

The one-and-only edition: The Forrester Wave: Conversation Intelligence For Customer Service, Q3 2023

The evaluation scored the top eleven vendors against thirty two criteria built on three foundational pillars: conversation analytics, automated and hybrid quality scoring, and agent augmentation. McAllister's framing in the announcement blog is worth keeping verbatim: the pillars are where vendors "declared their major," and while every evaluated vendor is building toward all three, not all have made it there yet.

CallMiner was named the only Leader, with the top score in both the current offering and strategy categories. Forrester's explanation: "CallMiner's leadership stems from its mastery of operational workflows. Its highly configurable and flexible taxonomy management experience, coupled with a vast library of pre-built categories, delivers value to customers. Its quality scoring product enables teams of varying maturity levels to go beyond compliance tracking, translating system scores into meaningful coaching moments." The current offering citation included the top score in conversation analytics itself, with the highest marks in automated and hybrid quality scoring, compliance, and unified platform experience, and strategy maximums in vision, roadmap, and pricing flexibility and transparency.

Cresta placed as a Strong Performer in its first participation, with maximum scores in coaching, agent augmentation, unified platform experience, delivery model, and integrations, plus the top strategy marks in innovation and roadmap. Forrester's line is worth reading twice: "Cresta's strength in coaching is unparalleled." The report credited market-leading capabilities for supervisor-led coaching, strong agent self-coaching engagement, the strongest score in agent augmentation, real-time guidance, and a comprehensive approach to generative AI. A Strong Performer citation that reads like a Leader citation, which turns out to be the foreshadowing.

The remaining nine placements sit behind Forrester's paywall. The two that are public are enough to carry the analysis.

The entry rule that defined the field

The most honest line in the whole evaluation is not a score. It is the inclusion rule.

To be evaluated at all, a vendor had to sell a standalone conversation intelligence solution, independent of interaction channels and independent of workforce optimization suites. Forrester stated the reason plainly: the contact center platforms carry similar features, but many fail to meet the level of analytics sophistication of these specialists, and few companies run a single platform for all interactions, so a specialized insights layer still has a job to do.

That rule is a definition of the market wearing the costume of a formality. It said the category existed to be bought separately from the channel and the workforce tools. It said specialization was the product. And it quietly excluded the suites that would later absorb the category anyway.

The blog's other section title does the same work with more humor: This Ain't Your Grandpappy's Speech Analytics. The old tool counted keywords in call recordings. The new category was pitched across three buyer teams at once: CX teams using it to lean less on surveys, marketing teams mining in-call outcomes for campaign optimization, digital teams finding web journey fixes. Inside the contact center, the sharpest line is the coaching one: "Hey you, improve your FCR!" is not actionable guidance. A score that converts into a coaching moment is.

What the sole Leader tells you about first editions

Now look at the winner's citation again, because it explains the entire result.

Taxonomy management. Pre-built categories. Quality scoring that becomes coaching. Every one of the criteria CallMiner topped maps to the playbook of a vendor that has spent two decades inside operational workflows. The report says exactly this: the leadership stems from mastering operational workflows and translating that into a flexible toolkit.

That is not an indictment of the scorecard. It is the mechanics of first editions. An analyst firm writes the criteria for a new category from what the market's incumbents have already built, because that is what a category is at first: the accumulated practice of its oldest vendors. The first scorecard therefore measures proximity to the incumbent playbook. CallMiner won the conversation analytics pillar, and conversation analytics was, in 2023, largely CallMiner-shaped.

First editions are definitions wearing scorecards. The ranks tell you who the definition was written around. The criteria tell you what the definition was. A buyer reading a first edition should read the criteria first and the ranks never.

The Strong Performer that outgrew the scorecard

Cresta's placement in this edition is the clearest evidence that first-edition verdicts have short half-lives.

In 2023: Strong Performer, strongest score in agent augmentation, real-time guidance singled out, coaching called unparalleled. In the successor Wave for contact centers, Q2 2025: Leader, highest current offering score, sixteen maximum criteria, and Forrester's "a force to be reckoned with."

The vendor did not get rewritten between editions. The category did. The 2023 criteria weighted the incumbent pillars, analytics and quality scoring, where CallMiner was strongest. The successor category weighted what happened next: real-time guidance, genAI safety and controls, revenue generation inside the live call. Cresta's playbook, in other words. The same vendor, the same product, two different categories, and the score moved with the definition, not with the vendor.

That is the lesson this old scorecard is uniquely placed to teach: a Wave score is a measure of fit to a specific definition at a specific moment. Definitions move. Scores do not follow vendors across the move.

What a retired scorecard is still worth

The honest limitation of this edition is its solitude.

There is no second edition under this name to correct against, no trend to read, no movement to chart. The category's successor took the Contact Centers name, and Forrester's December 2025 announcement folded separate conversation intelligence coverage into the new Customer Feedback Management And Analytics Solutions market. As a standalone category, this one existed for exactly one scorecard.

What survives retirement is the reasoning, not the ranks. Three pieces of it still pay rent in 2026. The entry rule: buy a specialist because the suite's analytics are not yet sophisticated enough, and re-test that assumption annually, because the suites are absorbing specialists as fast as they can. The pillars: whatever you buy must score quality without adding headcount and must turn scores into coaching moments, not dashboards. And the Cresta proof: before you weigh any score, read the criteria the score was built on, and ask whether your operation looks more like the criteria or more like your problem.

Three questions the old scorecard still answers

Does the vendor's strength line up with your operation's weakest pillar? The report split the market into analytics, quality scoring, and agent augmentation. Most vendors declared a major in one. Your contact center has a weakest pillar too. Buy against your weakness, not the vendor's trophy case.

Can a score become a coaching moment in your shop? The report's sharpest line is that FCR feedback is not actionable guidance. Ask any vendor to show one of its scores traveling from the dashboard into a coaching session with a named agent. Most will show you the dashboard.

Are you buying a specialist or a suite feature, and have you priced the difference? The 2023 entry rule made specialization the product. Two years of consolidation later, the honest conversation is about what a standalone bet costs you when the category itself is being folded. Get the roadmap in writing, including what happens to the product if the vendor is acquired.

The category existed for one scorecard, and the scorecard was better than its shelf life. It defined the market precisely, named its winner fairly, and then watched the market redefine itself out from under both. That is not a failure of analysis. It is what happens when the analysis is good.

Forrester does not endorse any vendor named here, and tier placement should not be read as a recommendation to buy.