Adthena on September 16, 2026 made a new capability called Decision Intelligence available to its customers. Marketers type a performance problem in plain language, such as a jump in cost per click or a fall in conversions, and the system returns a prioritised list of recommended actions built from two things: the advertiser's own account data and Adthena's record of what other advertisers were doing in the same auctions at the same time.

In Short

Adthena, a company that watches which advertisers show up in search advertising auctions, has released a tool that tries to explain why a campaign suddenly got more expensive or stopped converting. It matters because Google Ads and similar platforms mostly show an advertiser its own numbers, which makes it hard to tell whether a price jump came from a new rival, from a change across the whole market, or from something inside the account. The tool suggests what to change, such as search terms to exclude or bids to adjust, but a person still has to make every change in the account.

Two data sets, one diagnosis

According to Adthena, Decision Intelligence draws on two sources of information. The first is what happened inside a brand's own advertising account. The second is what was happening in the market around it over the same period: which competitors entered the brand's auctions, on which search terms, at what times, with what ad copy, and how that activity shifted over time. The second category, the company states, is information "that cannot be derived from any advertiser's own account."

Joining the two is the product's central claim. By combining account-level diagnostics with a continuous view of competitive activity, Decision Intelligence can show not only that a metric moved but why, according to Adthena, including whether the shift was market-wide or confined to one brand. A cost-per-click increase felt by every advertiser on a term points to a different cause than one confined to a single account.

Marketers interact with the system by describing the problem rather than building a report. Adthena lists four common scenarios supported at launch: climbing CPC, declining conversion rate, increasing cost per acquisition, and volume loss, which the company also describes as fewer clicks on the same terms. More specific scenarios, according to Adthena, are built as custom analyses "within a few days." The product page adds its own examples of starting signals, among them losing ground to a competitor and the effect of AI Overviews on click-through rate.

What the system returns

The output is a prioritised list of actions. Adthena gives three categories as examples: search terms to negate or tighten, bid and budget changes, and ad copy and asset updates. Each action is paired with an estimated impact and a description of the tradeoffs it involves, and the system offers multiple scenarios so an advertiser can pick the one that fits.

Recommendations also account for constraints, according to the company. Three are named: regulatory restrictions, legal review timelines, and strategies a brand has already tried and ruled out. The release does not say how such constraints are entered into the system, or whether they persist between sessions.

Paul Felby, Chief Product and Technology Officer at Adthena, described the starting point in the announcement. "Marketers don't wake up wanting another report. They wake up with a problem. Costs are climbing, conversions are down, budget is under pressure, and they need to know what to do about it," Felby said. "Decision Intelligence starts there. They bring us the problem in their own words, and we investigate both their account and the market around it to tell them not just what changed, but why, and what to do next."

Recommendations only, with an Approve button

On automation, the release is unambiguous. "Decision Intelligence provides recommendations only; changes to a client's account are made by the marketer, not automatically by Adthena," according to the company.

The product page tells a slightly different visual story. Its opening image shows a card labelled "Next move" with the text "A rival lifted bids on your top term, increase impression share target," and two buttons beneath it: Approve and Dismiss. A diagram further down describes the final stage of the system as "Actionable decisions," producing moves "priced and ready to execute."

Neither document explains what pressing Approve does. It could queue a change for a person to make in Google Ads, push it through an account connection, or simply mark a recommendation as accepted inside Adthena. For a product whose release states that account changes are made only by the marketer, the difference has operational consequences, and the materials reviewed for this article leave it open.

The example card raises a second question. Bids in Google's auction are not published to anyone outside Google. A statement that a rival lifted its bids is therefore an inference drawn from what can be observed, such as changes in how often and where a competitor's ads appear. The page does not describe how that inference is made or how confident the system is in it.

Six connections, or two?

At launch, according to Adthena, Decision Intelligence draws on connected data from six sources: Google Ads, Google Search Console, GA4, Search Ads 360, Adobe Analytics and OpenAI Ads. Support for Microsoft Advertising, Meta, Amazon, LinkedIn, TikTok and Reddit is in development, and the company says it plans to extend the capability across further paid search and social platforms over time.

The product page presents the picture differently. An infrastructure diagram headed "Every route in" marks only two inputs as live, Google Ads and ChatGPT Ads, and describes that pair of connections as "the start, not the ceiling." The four remaining sources from the release do not appear in the diagram at all. Further down, an outputs strip names Google Ads, ChatGPT Ads, Google Search Console, Adobe Analytics, Google Analytics and SA360, followed by a "More coming" tag.

Read together, the likeliest interpretation is that Google Ads and OpenAI's platform are the two advertising platforms connected, with the other four supplying analytics, organic search and search management data. Neither document draws that line explicitly. The naming differs as well: the release refers to OpenAI Ads, while the product page uses ChatGPT Ads throughout.

The absence of Microsoft Advertising at launch stands out. Adthena describes its patented Whole Market View as delivering competitor insights across Google, Bing and ChatGPT. The company can therefore observe the Bing auction from the outside, but on the evidence of the release it cannot yet read a Microsoft Advertising account from the inside.

The layer underneath

The product page describes Decision Intelligence as running on the same data layer as the Adthena app, the Adthena API, an MCP server and Ask Arlo. Ask Arlo is Adthena's existing AI assistant, one of several Adthena products recognised at the Search Awards. "Agents work through it in sequence," according to the page, which sets out three stages. The first is ingestion and signal, with "Every route normalised into one callable schema." The second is multi-agent processing, in which "Agents hand context down the chain, step by step." The third produces the actionable decisions. The layer itself is labelled structured, callable and agent-ready.

What the page does not say is whether external AI assistants can call Decision Intelligence through the MCP server, or whether that server only exposes Adthena's underlying data. Its seven frequently asked questions include "Can I trust the recommendations?", but the answers were collapsed in the version reviewed and are not available.

What an account can and cannot see

The release states that Google Ads and other advertising platforms show only the first of the two data sets, what happened inside the account. That is narrower than what Google's interface actually offers. Google Ads has long provided auction insights, a report comparing an advertiser with others competing in the same auctions on impression share, overlap rate and outranking share. Google brought auction insights into the Report Editor at account and manager account level in January 2022.

Access has not moved in one direction only. Google removed auction insights fields from Looker Studio in 2024, with existing data sources losing them on September 23 that year. On August 10, 2026, Google placed a prompt box above the insight cards on the Google Ads homepage, with "how competitors are affecting impression share" given as one of the example queries.

What auction insights does not provide is closer to the gap Adthena describes. The report names competing domains and a set of relative metrics. It does not show the ad copy those competitors ran or the individual search terms on which they appeared. Those are among the fields Adthena says its competitive insights track. The accurate version of the claim is not that platforms show nothing about competitors, but that what they show is thin and aggregated.

Causes that auction tracking cannot see

Separating a market-wide price move from a brand-specific one has value because costs have been moving for reasons unrelated to any single rival. WordStream's 2026 benchmark study, drawn from 13,474 US campaigns, put the average Google Ads CPC at $5.42, up from $2.32 in 2016. Channable's analysis of European ecommerce spend recorded a 15 percent CPC increase between June 2025 and June 2026 across combined Google Ads channels.

Some of those movements originate with the platform. Google began moving budget-limited Target CPA and Target ROAS campaigns closer to their stated targets from August 17, 2026, a change that Greg Finn of Cypress North predicted would lift CPCs in affected accounts. A platform-side change of that kind shows up in an advertiser's numbers, and possibly across a whole category at once, without any new competitor entering the auction.

Adthena's materials do not describe how Decision Intelligence separates competitor-driven movements from changes to Google's bidding systems or auction rules. No accuracy figures, case studies or customer results accompany the release. The estimated impact attached to each recommendation is Adthena's own projection, and the release does not explain the method behind those estimates or whether they have been checked against the results advertisers actually achieved.

A name used before

This is not the first time Adthena has used the Decision Intelligence name. On July 23, 2026, Adthena and dentsu announced a UK partnership built around a product called Decision Intelligence, which mapped competitor entry into ChatGPT prompt categories after UK cost-per-click swings of between 39% and 278%. That version advised on three moves: capping bids on inflated ad groups, redirecting budget toward uncontested prompt categories, and winning placements on creative strength rather than bid size.

The September 16 release describes Decision Intelligence as a new capability and makes no reference to the July product or to dentsu. Because the September version covers Google Ads and several analytics sources as well as OpenAI's channel, the likeliest reading is that the July product has been broadened and presented again under the same name. Adthena's materials do not say so, and the relationship between the two versions remains unconfirmed.

How the product line got here

Adthena was founded in London in 2012. The company became Google's first Trusted Trademark Partner on November 6, 2025. Later that month it detected 13 ads inside Google AI Overviews across 25,000 monitored results pages, a frequency of 0.052%.

The pace picked up in 2026, almost entirely around OpenAI's new channel. In April, Adthena published an analysis drawn from 29.1 million queries and then released AdBridge, a free tool porting Google Ads campaigns into ChatGPT. On May 11 it opened a ChatGPT Ads intelligence platform monitoring more than 300,000 daily prompts. By July it was publishing a weekly index that counted 7,378 ChatGPT advertisers across three markets in a single week, and in August it reported that retail took 39% of ChatGPT ad items on 24% of US queries.

Each of those steps widened what Adthena observes from the outside. Decision Intelligence is presented as the step that joins a customer's own account results to that market record, and it returns the company's focus to Google Ads.

Google's advisor covers similar ground

Google has spent the past year building its own agentic AI tools inside the same interfaces. Ads Advisor, which reached all English-language Google Ads accounts in December 2025, diagnoses problems such as ad disapprovals and can implement some proposed solutions once an advertiser approves them. In April 2026, Google added features that let Ads Advisor resolve certain policy issues, asking for approval before taking action. On May 20, Google merged Ads Advisor, Analytics Advisor and a Merchant Center agent into Ask Advisor, which draws on specialist agents spanning Google Ads, Google Analytics, DV360 and Merchant Center.

The overlap with Decision Intelligence is partial. Google's agent reads Google's own properties and, in some cases, can act inside them. Adthena's case rests on two things Google's agent does not claim to provide: an outside record of competitor ad copy and search terms, and connections to non-Google sources such as Adobe Analytics and OpenAI's advertising platform. Adthena's version, by its own account, does not act at all.

The connection question

Decision Intelligence reads live data from connected advertising accounts, yet the release does not describe how those connections are authenticated.

That detail now carries more weight. On August 31, 2026, Google published updated Google Ads Developer Policiesprohibiting programmatic proxies, defined as third-party interfaces, wrapper services or MCP servers that solely replicate, wrap or re-expose Google Ads capabilities as an intermediate layer. Every integration now has to connect through its own dedicated Google Cloud project. The definition turns on the word "solely," and a product that combines account data with its own competitive dataset is not obviously a wrapper. Nothing in Adthena's materials suggests its integration is affected. Google has said, however, that its Ads API Compliance team is actively reviewing existing integrations.

Why this matters for marketers

The launch marks a shift in what competitive intelligence vendors sell. For most of its history the category has produced reports: share of voice, competitor spend estimates, brand bidding alerts. Decision Intelligence moves Adthena into the recommendation layer, where agencies, in-house teams and Google's own advisor already operate. A tool that says a competitor has entered the auction is descriptive. One that says what to change in response takes on part of a decision that has usually belonged to the account manager.

That raises the question of accountability. The release is clear that the marketer makes every change, which leaves responsibility for outcomes with the advertiser. Yet the estimated impacts attached to each recommendation, and the choice among the scenarios offered, will shape those decisions, and none of the underlying methodology has been published. How much weight can a projected impact carry when no validation data accompanies it?

The inclusion of OpenAI's ad platform in the launch set reflects how quickly that channel has grown. OpenAI's advertising business reached a $1 billion annualised run rate in under 200 days, according to Digiday's reporting on August 31. Advertisers running campaigns on both Google and ChatGPT have been working with separate reporting systems. A diagnostic that reads both at once addresses that split, provided the connections work as described.

Several practical details are absent from the materials. Decision Intelligence is available now to existing Adthena customers, but the release gives no pricing, no statement on whether it carries a separate charge, and no list of markets or languages. The six platforms in development carry no dates.

Timeline

Summary

Who: Adthena, the London-founded search intelligence company, with Paul Felby, Chief Product and Technology Officer, quoted in the announcement. The capability is aimed at paid search teams at brands and agencies that are existing Adthena customers.

What: Decision Intelligence, a capability that takes a performance problem described in plain language, such as higher CPC, higher cost per acquisition, a lower conversion rate or volume loss, and returns a prioritised list of recommended actions with estimated impacts, tradeoffs and alternative scenarios. It combines an advertiser's connected account data with Adthena's tracking of competitor activity in the same auctions. It makes no account changes itself. The release lists six connected sources at launch, while the product page marks only Google Ads and ChatGPT Ads as live connections.

When: Announced on September 16, 2026, and available now to Adthena customers. Support for Microsoft Advertising, Meta, Amazon, LinkedIn, TikTok and Reddit is in development with no dates given. A product under the same name was announced with dentsu in the UK on July 23, 2026.

Where: Inside Adthena's platform, drawing on Google Ads, Google Search Console, GA4, Search Ads 360, Adobe Analytics and OpenAI Ads. The release does not specify markets or languages.

Why: According to Adthena, advertising platforms show an advertiser what happened in its own account but not what competitors were doing at the same time, which makes shifts in cost and conversion difficult to explain. That claim understates the competitive data Google Ads already provides through auction insights, but the ad copy and search-term detail Adthena tracks is not available in advertisers' own accounts.