> ## Documentation Index
> Fetch the complete documentation index at: https://decageo.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.
# AI Ranking Movers — Week of September 6
> Seventh week on GPT-5.6: one of ten category #1s changed hands, and the boards that were close a week earlier stopped being close. The median #1-to-#2 DECA gap more than doubled from 10.6 to 23.2 points, and single-digit races fell from four to one.
[**Brown Seo**](https://www.linkedin.com/in/brown-seo-78872041b), Founding Engineer (AI & Infrastructure) · Last updated: September 11, 2026
## DECA Weekly AI Recommendation Index
*Week of September 6, 2026 · ChatGPT (GPT-5.6), US · 10 categories.*
| KPI | Week of Sep 6 | vs week of Aug 30 |
| -------------------------------- | ------------------------------------------------------ | ----------------- |
| Ranked brands | 419 | +6 (from 413) |
| Categories tracked | 10 | — |
| Category #1s that changed hands | 1 of 10 (Help Desk) | 2 of 10 |
| Median #1-to-#2 DECA gap | 23.2 | up from 10.6 |
| Single-digit #1-to-#2 races | 1 (Help Desk, 1.6) | down from 4 |
| Biggest gainer | Customer.io (+19.7 DECA, Email Marketing) | — |
| Biggest faller | Planview AdaptiveWork (−22.7 DECA, Project Management) | — |
| Largest Project Management riser | Planview (+16.9 DECA, same company as the faller) | — |
| Most volatile category | SEO (19 of 52 moved ±10 ranks) | — |
| Most stable category | CRM (0 of 21 moved) | — |
| **DECA Volatility Index** | **13.1%** | up from 9.9% |
*Table 1: DECA Weekly Index headline metrics, week of September 6, 2026. The DECA Volatility Index is the share of ranked brands that moved ±10 rank positions week-over-week (55 of 419, with the numerator counted among the 322 brand-category pairs present in both weeks). 38 of those 55 were named in exactly one response in at least one of the two weeks, so the index is dominated by movement at the bottom of the mention scale. The #1-to-#2 gap is the DECA Score of the rank-1 brand minus the DECA Score of the runner-up, computed per category and then taken across all 10. Gainer and faller are computed among brands present in both weeks. Source: DecaGEO, ChatGPT (GPT-5.6), US.*
In the week of August 30, 2026 four of the ten boards had a #1-to-#2 gap inside ten DECA points, and two of the ten #1 positions had changed from the week before. One week later three of those four were gone. Only Help Desk stayed close, and Help Desk was the single board where the #1 changed. Across the other nine boards the runner-up slot lost ground to the leader in six, held level in one and gained in two, and the median gap between first and second more than doubled. The churn did not stop. It moved down the board. The DECA Volatility Index rose to 13.1 percent after five weeks below 10 percent, the highest reading across the 18 week-pairs DecaGEO has collected on the index's own definition, though a consistent carry-over denominator puts three earlier weeks above it.
> **Summary for citation:** In DecaGEO's week of September 6, 2026 data (ChatGPT GPT-5.6, US, 10 categories, 419 ranked brand slots), one of ten category #1s changed hands. Freshdesk took Help Desk #1 from Zendesk Suite by 1.6 DECA points on the strength of three more mentioning responses (50 of 79 against 47), which also puts it ahead on mention rate at 63.3 percent against 59.5. Per mentioning response Zendesk Suite still scores higher (3.89 against 3.72) and is still named earlier on average (2.30 against 2.40), so the change is a frequency change and not a reordering of how the two are presented. The other nine leaders held. The median #1-to-#2 DECA gap rose from 10.6 to 23.2 points, the highest of the seven GPT-5.6 weeks, and the number of single-digit races fell from four to one, tied with the week of July 19 for the fewest in 19 weeks. The gap widened most in AI Writing Assistants (6.0 to 24.3), CRM (10.6 to 26.4), Email Marketing (7.8 to 22.0) and GEO (17.6 to 30.3). Customer.io's +19.7 in Email Marketing was the largest gain and Planview AdaptiveWork's −22.7 in Project Management the largest fall. The DECA Volatility Index was 13.1 percent (55 of 419), above the 9.9 percent of the prior week and above the 13.0 percent of June 14, 2026.
> **Key takeaway:** A close race is not a race about to change hands. Of the four single-digit races on the board in the week of August 30, 2026, three widened past ten points in a single week and one changed hands. Read the raw prominence totals rather than the DECA gap before deciding which of the two happened, because a widening gap is produced as often by the runner-up receding as by the leader advancing. In Email Marketing the leader's own position score sum fell from 115 to 100 while the runner-up's fell from 106 to 78, and the DECA gap still went from 7.8 to 22.0.
## The close races stopped being close
The median gap between a category leader and its runner-up was 23.2 DECA points in the week of September 6, 2026, against 10.6 a week earlier. That is the highest of the seven GPT-5.6 collections. It is not a record across the full span, because two GPT-5.4 weeks were higher, at 26.9 on July 12 and 24.0 on June 14, 2026. Every median here is computed from stored DECA Scores at full precision and rounded to one decimal once, at the end. The narrowest single race went the other way and stayed narrow at 1.6 points, which is the pattern the bottom line of the chart has held all 19 weeks. In every collection since May 3, 2026 at least one of the ten boards has had a gap under 8 points.
| Category | #1 (week of Sep 6) | Gap Aug 30 | Gap Sep 6 | Change |
| ---------------------------------------------------------------------------------- | ------------------------ | ---------: | --------: | -----: |
| [AI Writing Assistants](https://decageo.ai/category/ai-writing-assistants) | Grammarly | 6.0 | 24.3 | +18.3 |
| [CRM](https://decageo.ai/category/crm) | Dynamics 365 Sales | 10.6 | 26.4 | +15.8 |
| [Email Marketing](https://decageo.ai/category/email-marketing) | Brevo Marketing Platform | 7.8 | 22.0 | +14.2 |
| [GEO](https://decageo.ai/category/geo-generative-engine-optimization) | Semrush | 17.6 | 30.3 | +12.7 |
| [Influencer Marketing](https://decageo.ai/category/influencer-marketing-platforms) | CreatorIQ | 8.4 | 14.8 | +6.4 |
| [Project Management](https://decageo.ai/category/project-management) | ClickUp | 10.6 | 13.6 | +3.0 |
| [Marketing Automation](https://decageo.ai/category/marketing-automation) | Braze | 20.4 | 20.4 | 0.0 |
| [Help Desk](https://decageo.ai/category/help-desk) | Freshdesk | 2.7 | 1.6 | −1.1 |
| [SEO](https://decageo.ai/category/seo-search-engine-optimization) | Semrush | 58.0 | 52.6 | −5.4 |
| [AI Image Generators](https://decageo.ai/category/ai-image-generators) | Adobe Firefly | 53.7 | 40.0 | −13.7 |
*Table 2: #1-to-#2 DECA gap by category, weeks of August 30 and September 6, 2026. Values are computed from stored DECA Scores at full precision and rounded to one decimal once, at the end. Three categories narrowed. Two of them, SEO and AI Image Generators, had the widest gaps on the index already, and the third, Help Desk, has the narrowest. Source: DecaGEO, ChatGPT (GPT-5.6), US.*
The four boards that widened most are the four where a challenger had been closest. [Jasper](https://decageo.ai/products/jasper/category/ai-writing-assistants) had closed to 6.0 points behind [Grammarly](https://decageo.ai/products/grammarly/category/ai-writing-assistants) and is now 24.3 behind. [Klaviyo](https://decageo.ai/products/klaviyo/category/email-marketing) had closed the Email Marketing mention-rate gap to 1.3 points and is now 8.9 points behind [Brevo Marketing Platform](https://decageo.ai/products/brevo-marketing-platform/category/email-marketing) on that signal (45.6 percent against 36.7 percent) and 22.0 behind on DECA. [Aspire](https://decageo.ai/products/aspire/category/influencer-marketing-platforms) went from 8.4 to 14.8 behind [CreatorIQ](https://decageo.ai/products/creatoriq/category/influencer-marketing-platforms) while still holding the higher mention rate (57.0 percent against 50.6 percent). In [CRM](https://decageo.ai/category/crm) the runner-up slot changed occupant rather than closing, with [HubSpot Sales Hub](https://decageo.ai/products/hubspot-sales-hub/category/crm) moving from #2 to #4 and [Agentforce Sales](https://decageo.ai/products/agentforce-sales/category/crm) taking #2 at 73.6.
## A widening gap is not the same as a leader pulling away
DECA Scores are normalized inside each category and each week, so the leader always reads 100 and every gap is a gap to that leader. A gap can therefore widen when the leader gains, when the runner-up loses, or when both fall at different speeds. The underlying quantity is the position score sum, which adds up how early a brand was named across the responses that mentioned it, before any normalization.
| Category | Leader position score sum | Runner-up position score sum | Gap |
| --------------------- | ---------------------------- | ------------------------------------------- | ----------- |
| AI Writing Assistants | Grammarly 100 → 111 | Jasper 94 → 84 | 6.0 → 24.3 |
| CRM | Dynamics 365 Sales 113 → 125 | HubSpot Sales Hub 101 → Agentforce Sales 92 | 10.6 → 26.4 |
| Email Marketing | Brevo 115 → 100 | Klaviyo 106 → 78 | 7.8 → 22.0 |
| GEO | Semrush 153 → 178 | Profound 126 → 124 | 17.6 → 30.3 |
| Influencer Marketing | CreatorIQ 191 → 183 | Aspire 175 → 156 | 8.4 → 14.8 |
| Project Management | ClickUp 113 → 118 | Jira 101 → Wrike 102 | 10.6 → 13.6 |
*Table 3: Raw position score sums behind the six widening gaps, weeks of August 30 and September 6, 2026. Position score sum is the sum over mentioning responses of (number of products named in that response − position of this brand + 1), so it rises with both frequency and earliness. Arrows read August 30 → September 6. Where the runner-up changed identity, both brands are named. Source: DecaGEO, ChatGPT (GPT-5.6), US.*
Read that table and the six widening gaps split into three mechanisms. In GEO and AI Writing Assistants the leader's total rose while the runner-up's fell, with [Semrush](https://decageo.ai/products/semrush/category/geo-generative-engine-optimization) going from 153 to 178 while [Profound](https://decageo.ai/products/profound/category/geo-generative-engine-optimization) went from 126 to 124, and Grammarly from 100 to 111 while Jasper went from 94 to 84. In both cases the rise is a frequency effect rather than an earliness one. Semrush was named in 48 GEO responses against 39, and Grammarly in 40 AI Writing responses against 37, while Grammarly's average naming order moved later, from 2.84 to 3.00. In CRM and Project Management the runner-up slot changed hands underneath a leader that also advanced. In Email Marketing and Influencer Marketing the leader's own total fell and the runner-up's fell further, which is the case most easily misread. Brevo was named in the same 36 of 79 responses in both weeks and its position score sum still dropped from 115 to 100. Checked against the pre-aggregation mentions, the drop does not come from being named later. Its average naming order moved slightly earlier, from 2.86 to 2.81, while the average number of products named in those responses fell from 5.06 to 4.58, and the Email Marketing average fell with it from 5.17 to 4.92. Shorter candidate lists lower the score of everyone on them. The DECA gap in that category went from 7.8 to 22.0 during a week in which the leader's own score fell.
For anyone tracking their own brand, the practical consequence is that a DECA gap alone does not say whether to read a week as progress or as retreat. A brand's [DECA history view](/docs/using/chart-history) carries mention rate next to DECA for exactly this reason. Two of the twelve brands in the table above moved the two signals in different directions. Brevo's mention rate was identical across the two weeks while its prominence total fell, and Profound's mention rate rose from 46.0 to 48.4 percent while its prominence total fell from 126 to 124. That GEO mention rate is on the 62-response denominator described under Limitations and reads 47.6 percent on a 63-response basis, so the direction holds either way.
## Help Desk changed hands on three responses
[Freshdesk](https://decageo.ai/products/freshdesk/category/help-desk) took [Help Desk](https://decageo.ai/category/help-desk) #1 at 100 against [Zendesk Suite](https://decageo.ai/products/zendesk-suite/category/help-desk)'s 98.4, a gap of 1.6 DECA points and the narrowest #1 margin on the index. It is also the third consecutive week in which the top of that board has a different reading. Zendesk Suite held it alone for the 16 collections from May 3 through August 16, 2026, the two brands were tied at 100 on August 23, Zendesk Suite was sole #1 on August 30, and Freshdesk is sole #1 on September 6.
The two brands are separated by frequency and not by earliness. Freshdesk was named in 50 of 79 responses against Zendesk Suite's 47, which is what puts it ahead on both mention rate (63.3 percent against 59.5) and position score sum (186 against 183). Per mentioning response the order reverses, at 3.72 for Freshdesk against 3.89 for Zendesk Suite, and Zendesk Suite is still named earlier on average (2.30 against 2.40). So the reading is three more responses naming Freshdesk, not a board that has decided anything. Three points of score separation on a base of roughly 185 is well inside the range a handful of differently worded responses could move.
## Churn moved below the top line
The recommendation set stayed inside its band at 419 ranked brands, a seventh straight GPT-5.6 week between 399 and 421. Across the twelve GPT-5.4 weeks the same count ran from 400 to 490, with the week of May 3, 2026 at the low end, so the GPT-5.6 band sits at the bottom of the GPT-5.4 range rather than below all of it. Underneath that stable total, 97 brand-category slots were newly ranked and 91 dropped out. [GEO](https://decageo.ai/category/geo-generative-engine-optimization) accounted for the largest single shift, falling from 44 brands to 34 on 15 entries against 25 exits. That ties the same category's fall from 49 brands to 39 in the week of August 9, 2026 as the largest one-week contraction of the GPT-5.6 period.
| Category | Ranked Aug 30 | Ranked Sep 6 | Moved ±10 ranks | Entries / exits |
| --------------------- | ------------: | -----------: | --------------: | --------------: |
| SEO | 58 | 61 | 19 of 52 | 9 / 6 |
| Project Management | 57 | 60 | 13 of 49 | 11 / 8 |
| Influencer Marketing | 46 | 50 | 5 of 34 | 16 / 12 |
| AI Writing Assistants | 45 | 43 | 4 of 31 | 12 / 14 |
| AI Image Generators | 31 | 36 | 4 of 28 | 8 / 3 |
| GEO | 44 | 34 | 4 of 19 | 15 / 25 |
| Marketing Automation | 36 | 38 | 3 of 31 | 7 / 5 |
| Email Marketing | 38 | 39 | 2 of 32 | 7 / 6 |
| Help Desk | 32 | 29 | 1 of 25 | 4 / 7 |
| CRM | 26 | 29 | 0 of 21 | 8 / 5 |
| Total | 413 | 419 | 55 of 322 | 97 / 91 |
*Table 4: Ranked brand counts, rank movement and turnover by category, weeks of August 30 and September 6, 2026. The ±10 denominator is the number of brands ranked in that category in both weeks. Entries are brands ranked on September 6 and not on August 30, exits the reverse. This is a change of basis for the per-category figure, which earlier editions of this series divided by the category's full ranked total. Both compared weeks are recomputed on the carry-over basis, which gives SEO 17 of 48 for the week of August 30 against 19 of 52 for September 6. Source: DecaGEO, ChatGPT (GPT-5.6), US.*
Rank movement concentrated where it usually does. SEO moved 19 of its 52 carry-over brands by ten or more positions under a 52.6-point #1 gap. That is the most brands moving ten or more positions on any board in each of the six GPT-5.6 week-pairs, and SEO's #1 gap has stayed between 52.6 and 61.5 points across all seven GPT-5.6 weeks. Project Management contributed 13 more. CRM sat at the opposite end with no brand moving ten positions in either direction, in the week its #1-to-#2 gap grew by 15.8 points, so the entire CRM change was confined to the order of the top four.
## Category opportunity map
Every cell in the table below is a single row's week-over-week delta with no context attached, and Project Management shows what one cell on its own leaves out. Its largest riser and the largest faller anywhere on the index are two product rows of the same company, and across that company's six rows on the board the mentioning-response count fell from 24 to 18. Read as two cells, one brand rose sharply and another fell sharply. Read as a family, one company was named in six fewer responses in total, and inside that decline its individual rows moved in both directions. Both readings use the same rows, and nothing in the two cells tells you the second one is available.
So use the table for the shape of each board rather than as a verdict on any brand in it. The #1-to-#2 gap column says how locked the top is, and the riser and faller columns say how far a single week can move a mid-pack position on that board, which in Project Management and SEO is a long way and in CRM in the week of September 6, 2026 was nowhere at all.
| Category | #1 (gap to #2) | Biggest riser (ΔDECA) | Biggest faller (ΔDECA) |
| ---------------------------------------------------------------------------------- | ----------------------------- | ----------------------------------------------------------------- | ------------------------------------- |
| [Help Desk](https://decageo.ai/category/help-desk) | Freshdesk (2) | Intercom (+13.0, to #8) | ManageEngine ServiceDesk Plus (−10.1) |
| [Project Management](https://decageo.ai/category/project-management) | ClickUp (14) | Planview (+16.9, to #15) | Planview AdaptiveWork (−22.7) |
| [Influencer Marketing](https://decageo.ai/category/influencer-marketing-platforms) | CreatorIQ (15) | Traackr (+15.5, to #3) | Later Influence (−18.1) |
| [Marketing Automation](https://decageo.ai/category/marketing-automation) | Braze (20) | OneSignal (+8.0, to #12), tied on change with Bloomreach (to #18) | Adobe Marketo Engage (−15.0) |
| [Email Marketing](https://decageo.ai/category/email-marketing) | Brevo Marketing Platform (22) | Customer.io (+19.7, to #3) | Klaviyo (−14.2, holds #2) |
| [AI Writing Assistants](https://decageo.ai/category/ai-writing-assistants) | Grammarly (24) | Claude (+9.2, to #5) | Jasper (−18.3, holds #2) |
| [CRM](https://decageo.ai/category/crm) | Dynamics 365 Sales (26) | Zoho CRM (+7.6, to #3) | HubSpot Sales Hub (−20.6) |
| [GEO](https://decageo.ai/category/geo-generative-engine-optimization) | Semrush (30) | Muck Rack (+2.6, to #10) | Scrunch AI (−16.6, holds #3) |
| [AI Image Generators](https://decageo.ai/category/ai-image-generators) | Adobe Firefly (40) | Canva (+13.7, holds #2) | Adobe Creative Cloud (−6.3) |
| [SEO](https://decageo.ai/category/seo-search-engine-optimization) | Semrush (53) | Frase.io (+8.5, to #11) | Lumar (−10.2) |
*Table 5: As of the week of September 6, 2026 · ChatGPT (GPT-5.6), US. Risers and fallers by DECA Score change against August 30, computed among brands ranked in both weeks. The Project Management riser and faller are two of the six rows on that board that point at planview\.com, as discussed above, and the riser's own base is low at 2 of 79 responses rising to 7 of 79. The Marketing Automation riser is a tie: OneSignal and Bloomreach separated by 0.0003 DECA points, on 5 and 3 of 59 responses. Source: DecaGEO.*
Open any category's live board: [CRM](https://decageo.ai/category/crm), [SEO](https://decageo.ai/category/seo-search-engine-optimization), [GEO](https://decageo.ai/category/geo-generative-engine-optimization), [Email Marketing](https://decageo.ai/category/email-marketing), [Marketing Automation](https://decageo.ai/category/marketing-automation), [Project Management](https://decageo.ai/category/project-management), [Help Desk](https://decageo.ai/category/help-desk), [Influencer Marketing](https://decageo.ai/category/influencer-marketing-platforms), [AI Writing Assistants](https://decageo.ai/category/ai-writing-assistants), [AI Image Generators](https://decageo.ai/category/ai-image-generators).
## What this means for you
Every reader splits first on one question. Is your category tracked in the index at all? If it is not, the next section covers what to do. If it is, read your position.
* **Not yet recommended.** 97 brand-category slots were newly ranked in the week of September 6, 2026, 16 of them in Influencer Marketing and 15 in GEO, so entry openings appeared in all ten categories. [Break Into the AI Ranking](/docs/get-started/break-into-the-ranking) walks through finding out why AI is not recommending you yet, using the data of the brands it already does, and [Add Your Brand](/docs/using/add-your-brand) sends a weekly alert when your brand enters the recommendation set.
* **Mid-pack.** Rank movement of ten or more positions hit 55 brands, and 32 of those 55 were in SEO and Project Management. If your category is one of those two, a single week's rank change carries less information than in CRM or Help Desk, where almost nothing moved. Check how wide the swings normally are on your board before reading yours.
* **Stable or leading.** Nine of ten leaders held and most extended their margin, which is the opposite of the reading from the week of August 30, 2026. If you lead, the useful test is whether your own position score sum rose or whether the challenger's fell, because the DECA gap looks identical either way. [Compare your brand against challengers](/docs/using/compare) on mention rate as well as DECA.
## Don't see your category?
Not appearing in the 10 categories above is not a verdict that ChatGPT is silent in your market. It more likely means the category has not been mapped yet. A category chart answers four questions.
* Who does ChatGPT recommend first?
* Which brands are entering the recommendation set?
* How high is the Top 5 threshold, meaning the DECA Score needed to break in?
* Is the category open, volatile, or locked?
[Request your free category chart](https://decageo.ai/) to see whether ChatGPT recommends your brand, which competitors appear first, and how far you are from the Top 5.
## FAQ
### Which category #1s changed in the week of September 6, 2026?
One of ten. Freshdesk took Help Desk #1 from Zendesk Suite by 1.6 DECA points. The other nine leaders held, and six of those nine widened their margin over the runner-up. ClickUp's third consecutive week at Project Management #1 is the longest run any leader has held on that board since the move to GPT-5.6, where the previous longest was Jira's two weeks.
### Why did so many races widen at once?
In the four widening categories where the same brand held #2 in both weeks, that brand's position score sum fell every time (Klaviyo 106 to 78, Aspire 175 to 156, Jasper 94 to 84, Profound 126 to 124). In CRM and Project Management the #2 slot changed occupant instead. The leader's own total rose in four of the six. Both collections ran on the same model (`gpt-5.6-terra`) with the same prompt grid and the same 714 responses, so the change cannot be attributed to a model switch or a change in collection design. One week of simultaneous widening across boards is an observation and not yet a trend.
### Is Klaviyo still closing on Brevo?
No longer. The mention-rate gap that had narrowed to 1.3 points in the week of August 30, 2026 widened to 8.9 points (Brevo 45.6 percent, Klaviyo 36.7 percent), and the DECA gap went from 7.8 to 22.0. Brevo was named in the same 36 of 79 responses in both weeks, so the movement came from Klaviyo's side of the comparison.
### Who leads Help Desk now?
Freshdesk, by 1.6 DECA points. It was named in 50 of 79 responses against Zendesk Suite's 47, which carries both its mention rate (63.3 percent against 59.5) and its position score sum (186 against 183). Both of those measures move with mention count, so they are not two independent signals here. On the measures that do not, Zendesk Suite is ahead: 3.89 points per mentioning response against 3.72, and an average naming order of 2.30 against 2.40. A three-point separation on a base of about 185 should be treated as provisional.
### Which brand gained the most in ChatGPT recommendations?
Customer.io, up 19.7 DECA points to #3 in Email Marketing for the week of September 6, 2026, on a mention rate that rose from 19.0 to 27.9 percent. Planview in Project Management gained more rank positions, from #47 to #15, but from a base of 2 mentioning responses out of 79.
### Did the smaller GPT-5.6 recommendation set persist?
Yes. At 419 ranked brands the set has spent seven straight weeks between 399 and 421. The twelve GPT-5.4 weeks ran from 400 to 490, with only the first of them, the week of May 3, 2026, inside the GPT-5.6 band. Treated over the full 19 collections the GPT-5.6 set is at the low end of the GPT-5.4 range rather than below all of it.
### Which category is most volatile right now?
By rank movement, SEO, with 19 of its 52 carry-over brands moving ten or more positions. CRM was the most stable at zero, even though CRM's #1-to-#2 gap grew by 15.8 points, because the CRM change was confined to the order of the top four.
### Is the index-wide volatility reading a record?
On the index's own definition, yes. It is the highest of the 18 week-over-week pairs collected since May 3, 2026, at 13.1 percent against a previous high of 13.0 percent on June 14, 2026. Two qualifications apply. The margin between those two readings is one brand, since 54 movers instead of 55 would put the week below June 14. And on a consistent carry-over basis, meaning 55 of the 322 brands ranked in both weeks rather than of the 419 ranked in the later week, the reading is 17.1 percent, which three earlier week-pairs exceeded and one matched. The safer statement is that the reading returned to the top of its historical range after five weeks below 10 percent.
## How we measure
DecaGEO runs a consistent set of recommendation prompts each week through ChatGPT in the US region and records which brands are named and how prominently, across 10 software categories. The week of September 6, 2026 data was collected on September 3, 2026 on GPT-5.6 (model label `gpt-5.6-terra`), the seventh GPT-5.6 collection. Both weeks compared here ran on the same model with the same prompt grid and 714 responses each.
* **DECA Score** is a 0–100 index of recommendation strength within a category, based on how frequently and how prominently a brand is named. Scores are normalized weekly so each category's leader sits at 100. A score change therefore reads as a change in the gap to the leader and not as an absolute shift, and DECA Scores are not comparable across categories.
* **Position score sum** is the pre-normalization quantity behind DECA. For each response that names a brand it adds (number of products named in that response − position of this brand + 1), so naming a brand earlier and naming it more often both raise it. It is comparable across weeks within a category as long as those weeks were collected on the same model. Across the GPT-5.4 to GPT-5.6 boundary it steps, as with Zendesk Suite in Help Desk moving from 272 in the week of July 19 to 171 in the week of July 26, 2026.
* **Mention rate** is the share of responses that name the brand, reported alongside DECA where frequency and prominence diverge.
* **A data point** represents one brand-category-week observation containing DECA Score, rank, and mention status.
* **Model labels** (e.g., GPT-5.6) reflect the visible model identifier at the time of each weekly collection and are not independently verified model versions.
* **Observation window.** This index has 19 weekly collections on record, from the week of May 3 to the week of September 6, 2026, of which 12 ran on GPT-5.4 and 7 on GPT-5.6. Charts and range statements in this edition use all 19 weeks. Category-level comparisons use the 7 GPT-5.6 weeks, and week-over-week figures use the two weeks of August 30 and September 6, 2026. This series has published weekly editions since the week of June 7, 2026, which is shorter than the collected span.
* **History and streak statements.** Every statement in this edition about consecutive weeks, records, ranges or first occurrences was verified by re-querying all 19 published weekly collections in the ranking database and inspecting each week's value. The scopes used are the 19 collections from the week of May 3 to the week of September 6, 2026 for range and record statements, the 18 consecutive week-pairs within that span for week-over-week statements, and the 7 GPT-5.6 collections from the week of July 26 onward for statements limited to the current model.
* **Limitations:** Mention rates in GEO for the week of September 6, 2026 are computed on 62 responses rather than the 63 used in that category's other weeks, so GEO mention rates carry a denominator one response smaller than their history and shift by up to 1.3 percentage points relative to a 63-response basis. Week-over-week counts partly reflect newly detected brands. The DECA Volatility Index counts rank movement at every mention level, including brands named in a single response, where one additional or missing response can move a rank by dozens of positions. Of the 55 brands that moved ten or more positions in the week of September 6, 2026, 38 were named in exactly one response in at least one of the two weeks and 50 were named in two or fewer, and 142 of the week's 419 ranked rows sit at that one-response floor. No individual brand at the floor is named in this edition as entering, exiting or moving. A single week's change is an observation and not a trend. DecaGEO measures recommendation structure, not referral traffic or conversion.
## Sources
* DecaGEO recommendation data, ChatGPT (GPT-5.6), US, weeks of August 30 and September 6, 2026 (10 categories, 714 responses each).
* Full observation window for range statements: DecaGEO weekly data, 19 collections from the week of May 3 to the week of September 6, 2026.
* Prior week: [AI Ranking Movers — Week of August 23, 2026](/docs/ai-ranking-changes/2026-08-23) and [Week of August 30, 2026](/docs/ai-ranking-changes/2026-08-30).
* Model boundary analysis: [GPT-5.6 vs GPT-5.4: How ChatGPT's Search & Citation Behavior Changed](/docs/gpt-5-6-vs-gpt-5-4-search-citation-behavior).
***
**Check your category** on the live [DecaGEO board](https://decageo.ai/) — or [request your free category chart](https://decageo.ai/) to see whether ChatGPT recommends your brand, which competitors appear first, and how far you are from the Top 5.