The Tool Time Index, Q3 2026: Who the AI Engines Actually Name

No vendor paid to be in this Index and no vendor can. No sponsors, no ads, no affiliate links, no client data. Every tool was scored on the same four public signals with the same published formula, and the formula is below so you can argue with it. I get things wrong. Ten errors were caught in this Index before publication and they are all published in a corrections log that ships beside this issue, including four of my own issues that reviewed a company which had already changed hands. Read me skeptically. That is the point of me.

Here is the number that reframes this whole category. Across the twenty-nine ranked positions, the correlation between a tool's Market Momentum score and its AI Visibility score is minus 0.067. That is not a weak relationship. That is no relationship at all.

Market Momentum measures financing recency, capital scale, published shipping cadence and independence. AI Visibility measures how often the engines name a tool when someone asks a real buying question. Those two things move independently. A company can be doing everything an investor rewards and be functionally absent from the surface where buyers now start looking.

The examples are not subtle. Backstory has raised roughly $190M and is named in 1.5 percent of answers about its own category, and zero percent when the engines answer from memory alone. Qualified has raised roughly $163M, holds the strongest buyer sentiment in its category, and is named in 3.9 percent. Attention shipped thirty dated releases in twelve months, closed a $30M Series B in June, holds the third-highest momentum score in the entire Index, and was named zero times across 1,296 answers. Alta closed a $25M Series A twelve weeks before this Index and was also named zero times.

Meanwhile Vidyard, whose last priced round was in January 2016, leads its category. Responsive, which has not raised since 2018, leads its own.

Six of the twenty-nine are invisible to model memory entirely, scoring zero percent when the engines answer without searching: Alta, AthenaHQ, Attention, Backstory, SiftHub and Terret. They exist on the web. They do not exist in the model.

A composite out of 100, per tool. Four inputs, none of which needs data a vendor could give me.

AI Visibility (30). How often the engines name a tool when answering a real buying question in its category, measured in both retrieval modes. Fifty-four prompts, nine per category, put to ChatGPT, Perplexity, Gemini and Claude across six runs. That is 1,296 answers, every one saved so any count can be re-derived rather than trusted.

Buyer Sentiment (25). Ratings and review counts across G2, Gartner Peer Insights, Capterra, TrustRadius and Trustpilot, corrected for how the reviews were collected. Ninety-two cited figures, each with a URL and a retrieval date.

Market Momentum (25). Financing recency, capital scale, published shipping cadence and independence. One hundred and ten fetched source URLs across twenty-eight companies.

Editorial (20). The GTM Tool Time teardown score from the published issue, rescaled. My opinion carries the smallest weight of the four, on purpose.

Scores below are my framework-based read of each tool's public positioning and category archetype, not lab-tested benchmarks. A map, not gospel.

The Index Score is all four signals. It answers how strong a tool's overall position is right now.

The Quality Score is Editorial plus Buyer Sentiment, rescaled to 100. It answers the narrower question you probably came for: is the product any good. It is a derived view of two pillars that already exist. It changes no Index Score and it adds no new signal.

Where the two numbers diverge, that gap is the most useful thing on the page. A high Quality Score with a low Index Score is a good product that buyers are struggling to find. The verdict band reads the INDEX Score, not the Quality Score, so a tool can be rated highly by the people who use it and still read situational or skip or wait here. That is not a contradiction. It is the Index telling you the product is good and the market position is not.

Verdict band: 90+ buy with confidence, 75 to 89 strong with caveats, 60 to 74 situational, under 60 skip or wait. Movement reads Baseline everywhere, because this is the first Index and there is nothing yet to move against. Q4 will grade this one in public.

14 tools reviewed in Issue #003. 4 ranked.

Profound. Index 94.4, Quality 87.9. The only tool in this Index to clear 90. Better known from model memory than findable on the live web, which no other tool here manages.

Peec AI. Index 79.2, Quality 72. The clearest challenger. Strong on the live web, much weaker in model memory, which is rented visibility.

AthenaHQ. Index 60.6, Quality 74.9. Highest editorial score in the category and the smallest balance sheet in the Index. Invisible to model memory.

Scrunch AI. Index 56.1, Quality 75.3. Acquired by Sitecore in June. Public release cadence has been zero since roughly that date.

16 tools reviewed in Issue #004. 5 ranked.

Gong. Index 83.6, Quality 88.3. Still the most complete platform here, and the engines agree. You will pay for it.

Fathom. Index 72.0, Quality 87.6. The best buyer sentiment in the entire Index. Acquired by Superhuman sixteen days before this Index.

Granola. Index 70.8, Quality 76.3. Third-highest momentum in the Index. The engines have barely learned it exists: 1.9 percent from memory.

Avoma. Index 66.8, Quality 84.3. Quietly good on every pillar and first on none. The dedicated release-notes page has been stale since 2020.

Attention. Index 57.6, Quality 76.2. Thirty dated releases and a $30M round in June. Named zero times in 1,296 answers.

14 tools reviewed in Issue #006. 5 ranked.

Quality versus Index. Three of the five here score above 80 on quality and under 60 on the Index. Backstory, Terret and Aviso are all rated well by the buyers who use them and all are close to invisible to the engines. Two of the three renamed themselves inside the last eighteen months.

Gong. Index 79.5, Quality 88.3. Still the most complete platform here, and the engines agree. You will pay for it.

Clari. Index 74.6, Quality 83.7. The engines name it more than anything else in this category. Its own domain now redirects to Salesloft.

Aviso. Index 58.5, Quality 81.8. Last priced round was April 2015. That is the oldest capital in this Index.

Backstory (formerly People.ai). Index 53.4, Quality 82.4. Raised roughly $190M. Named in 1.5 percent of answers, and zero percent from memory.

Terret (formerly BoostUp). Index 51.6, Quality 80.1. Renamed in September 2025. The old name still outscores the new one in model memory.

6 tools reviewed in Issue #002. 5 ranked.

Quality versus Index. Loom has the highest Quality Score in the category and finishes second, because Atlassian owns it and its release feed has been quiet since May. Vidyard wins on the strength of everything except capital age.

Vidyard. Index 78.2, Quality 85.5. The most complete package in the category. Last priced round was January 2016.

Loom. Index 71.5, Quality 86.2. Still the name the engines reach for. Owned by Atlassian since 2023, and its release feed has been quiet since May.

Sendspark. Index 64.4, Quality 82.5. Punches far above its capital. No publishable funding total and a changelog most buyers never find.

HeyGen. Index 60.3, Quality 73.3. Strong momentum, weak recall. Named in 8.4 percent of answers in a category it helped define.

Tavus. Index 56.1, Quality 58.4. Fifty-six dated releases in twelve months, the busiest changelog in the Index. The thinnest review base and the lowest sentiment score.

15 tools reviewed in Issue #005. 5 ranked.

Responsive. Index 77.4, Quality 85.3. Wins on the strength of what the engines say, in both retrieval modes. Has not raised since 2018 and does not publish release notes on its own site.

Loopio. Index 74.7, Quality 81.8. Loses the category by 0.4 points, which is a tie. Its release notes require a login.

DealHub. Index 71.7, Quality 85.1. Best momentum in the category by a distance. The engines name it a fifth as often as the two leaders.

AutoRFP.ai. Index 58.9, Quality 76.2. No financing of any kind we could find, and no public changelog. Still out-scores a better funded rival.

SiftHub. Index 49.0, Quality 73.6. Lowest score in the Index. Invisible to model memory and thin on every public signal.

18 tools reviewed in Issue #001. 5 ranked.

Quality versus Index. The sharpest divergence in the whole Index is here. Qualified holds the highest Quality Score of any AI SDR tool at 88.4, on the strongest buyer sentiment in the category, and finishes third with a skip or wait band. Salesforce closed its acquisition in April and the engines have not caught up: Qualified is named in 3.9 percent of answers about its own category. Meanwhile 11x leads the Index at 67.0 while placing fourth of five on quality. If you are buying on the product, start with Qualified. If you are buying on whether your team will find it, hear about it and get it supported at scale, that is a different question and the Index answers it differently.

11x. Index 67.0, Quality 71.4. Leads on balance, not on strength. Nothing in this category clears 75.

Artisan (Ava). Index 61.1, Quality 62.8. Best AI visibility in the category. Thinnest review base of the five.

Qualified (Piper). Index 59.7, Quality 88.4. Strongest buyer sentiment here. Inside Salesforce since April, and the engines barely name it.

Unify. Index 54.9, Quality 73.7. The highest editorial score we have given anyone, 86. Named in 2.0 percent of answers about its own category.

Alta. Index 52.8, Quality 77.5. Raised $25M twelve weeks before this Index. Named zero times in 1,296 answers.

Issue #003 ranked Ellipsis in a teardown of AI search visibility software. Re-verifying for this Index, no SaaS visibility tracker of that name exists. The only confirmable match is an AI search agency: a services firm doing strategy and implementation for clients. A services business cannot be scored against software vendors on any of the four pillars. It is excluded, AI Search Visibility publishes as a Top 4, and this is the correction.

An index that will not grade its own back catalogue has no business grading anyone else's. Five of the twenty-nine ranked positions have changed hands: Qualified inside Salesforce, Loom inside Atlassian, Scrunch inside Sitecore, Fathom inside Superhuman, and Clari merged into Salesloft, whose domain now redirects away.

Four of those five were already true on the day we published the teardown that reviewed them. Only Fathom changed after publication. So four of our six issues carried a stale ownership fact. A fifth called a company BoostUp when it had been Terret for more than a year.

Ten errors were caught before this Index published, and none of them reached you. Two were killed in the last forty-eight hours. The first was this issue's intended villain, which alleged something about a review platform that turned out to be unsupportable. The second was the statistic I had planned to lead with, a valuation markdown that was arithmetically correct and editorially wrong.

The uncomfortable one: ten companies in this Index were recorded as publishing no release notes, and six of those findings were wrong. Five of the six were reachable from the vendor's own website footer. That error moved six scores and changed a category leader. It is fixed, and the rule that prevents it is now published.

On September 11, 2025, G2 and Profound announced a partnership. Profound's answer-engine-optimization data powers G2's AI Visibility Dashboard inside my.G2. Both companies published it. No financial terms were disclosed.

Profound is separately rated on G2 itself, holding about 4.5 stars on roughly 1,130 reviews in G2's Answer Engine Optimization category, a category G2 also ranks.

I want to be precise about what this is and is not, because the obvious accusation does not survive checking. G2 states publicly that Grid placement is computed algorithmically, is identical across categories, cannot be altered by staff or vendors, and cannot be bought. Incentivized reviews are weighted less, not more. And Profound is not even the highest-rated tool in its own category: several competitors sit at 4.9 while it sits at 4.5. If anyone were tilting the scale, they did a poor job of it.

Here is what survives, and it is enough. A review platform now has revenue exposure to a vendor it rates. The partnership is disclosed in G2's newsroom and its partner hub. It is not disclosed on Profound's G2 product page, and it is not disclosed on the category page. Those are the two places a buyer actually goes to compare. This is a disclosure problem, not a manipulation problem, and it is the kind of thing that gets worse quietly.

Profound is the number one tool in this Index and it is the subject of this section. Both of those things are true and neither changes the other. That is what independence costs and it is the only reason to read a ranking that nobody paid for.

Being fundable and being findable are different skills, and most companies are only good at one. The momentum pillar and the visibility pillar correlate at essentially zero. Every company in this Index has a team working on the first problem. Almost none of them have anyone working on the second, and the six scoring zero in model memory are not small companies. One of them has raised $190M.

Renaming your company costs you visibility you cannot buy back quickly. In the same category, in the same runs, the retired name BoostUp is named in 11.1 percent of memory-mode answers while the current name Terret scores zero. People.ai scores 8.3 percent where Backstory scores zero. RFPIO still carries 28.1 percent against Responsive's own 28.6. The models learned the old name and have not finished learning the new one.

Ownership changes faster than anyone updates their shortlist. Five of twenty-nine have changed hands, and four of those were already true when we reviewed them. If a buyer is working from a comparison article more than a few months old, a fifth of their shortlist is a different company than they think it is.

Three things you would expect in a quarterly index are missing, and saying so is more useful than faking them.

Movement. There is no prior Index to move against. Every row reads Baseline. Q4 2026 is the first quarter with real movement and it will grade this one.

The three standing awards. Climber of the Quarter needs a prior rank, Leader Unseated needs a prior leader, and in a first index every entrant is new. All three begin in Q4.

Scope. Six categories, not the whole market. The rest land as teardowns publish. An index that covered everything on day one would be covering most of it badly.

Q4 publishes in late December with real movement, the three standing awards live, and at least two new categories from the teardowns running between now and then. It will also grade this Index in public: every tool that moves, every call that aged badly, and every number in this issue that turns out to have been wrong.

GTM Tool Time is an independent review. No vendor pays to be here. Reply and tell me which category to tear down next.

Marty

Marty Berman is the founder of Scalient IQ. GTM Tool Time is editorially independent of it. No client data is used in any review or in Index scoring, no vendor pays to appear here, and the Index runs on public signals only.

See the full directory and the scoring rubric at gtmtooltime.beehiiv.com.