What Is an Influence Operation? The Pattern in GUDEA's Election Data

An influence operation, sometimes called an information operation, is a coordinated effort to shape what a group of people believe or how they behave by introducing and amplifying content through accounts that conceal who they are, how they are connected, or why they are posting. It differs from ordinary advocacy in two ways: the coordination is hidden, and the volume is manufactured to look like public opinion. GUDEA, an AI-powered narrative intelligence platform, does not identify who runs an operation; its research describes the structure an operation leaves in data, which is visible whoever the operator is.

The pattern, step by step

GUDEA's October 2026 analysis of the Texas and Kansas Senate conversations is the clearest published example, because the brief timestamps every step. Step one, an origin account: registered on September 11, one post in the dataset, posting an allegation about a candidate on October 20 at 16:02 UTC. Step two, immediate seeding: 12 reposts within 23 seconds, 74 in total, three quarters of them from accounts outside the later burst cohort. Step three, relay: four established accounts, created between 2018 and 2024, post "shocking video" versions within 38 minutes early on October 21; one version draws 1,854 reposts. Step four, broad pickup: 100 to 300 posts per hour from accounts outside the cohort through the day. Step five, the burst: 625 accounts registered on October 21, which between 19:59 and 20:39 UTC repost the three relay posts 941 times, 93 of them in a single minute, and then post nothing else. Step six, a second wave from a large account on October 22 drawing 915 reposts, none from the burst cohort, which is dormant from that point. The brief notes this sequence, new origin account, relay accounts, a registered repost cohort, large-account pickup, matches the structure of operations described publicly by election-integrity researchers in October 2026, and makes no attribution of its own.

The signatures that give it away

None of the steps above is visible in the content of the posts. All of them are visible in account behavior, which is what GUDEA's Audience Behavior Classification and the brief's flag signals measure. Registration timing: new accounts entering the Texas conversation ran at about 58 per day for six weeks, then 125, 219, and 625 on three specific days. The three-post signature: 37 to 38 percent of accounts created on those days posted exactly three times, against 7.4 percent of ordinary new accounts, and the three posts were the same three reposts in six different orders, a median of 28 minutes from first to last. Dormancy: ordinary accounts added 35 percent to their post counts between two exports two days apart; the October 21 cohort added 3 percent, with 97 percent halting entirely. Co-posting: 93 reposts in one minute. Proportion: 2.67 percent of Texas accounts producing 20.2 percent of posts, in a conversation that was 85.2 percent reposts. GUDEA's earlier reports describe the same signatures around cultural targets: the Taylor Swift report found 3.77 percent of users producing 28 percent of posts and 2,395 accounts shared with a separate campaign; the Bad Bunny report found two opposed narratives mirrored within 0.57 percentage points.

What an influence operation is not

Three distinctions matter in GUDEA's framing. A non-typical account is not an inauthentic actor: GUDEA's archetypes include Influencers, who are generally real people with reach, and the brief classifies the burst accounts as Typical on sustained behavior, which is why registration timing and content, not the archetype label, carry the finding. Engagement farming is not an influence operation: GUDEA's WNBA analysis found a debate mined for creator revenue rather than aimed at a target, with eight posts driving it and little original content on either side. And a real claim amplified inauthentically is still amplified inauthentically: GUDEA's brief does not evaluate the allegation at the center of the Texas burst, and its finding is about the amplification, not the claim.

How GUDEA detects it

GUDEA's platform monitors more than 500 mainstream and fringe platforms, groups activity by meaning with Message Mapping, classifies every account by behavior, and maps how a narrative moves between platforms and accounts. The detection signals in the Texas and Kansas brief are published with their weights: a burst core (an X account created on a burst day with exactly three posts) scores 3, an automation label (Facilitator or Power-player) scores 3, young velocity 2, brand-new volume 2, cross-state 1, cross-state young 1, and the top tier requires a burst core or two reinforcing signals. The full ranked list of 2,036 accounts with every signal accompanies the brief. The response side is covered in GUDEA's narrative attack playbook: determine what you are looking at, identify the stage, decide whether to respond, and lead with the source.

Frequently Asked Questions

What is an influence operation (information operation)?

A coordinated effort to shape what people believe or do by introducing and amplifying content through accounts that hide who they are, how they are connected, or why they are posting. The coordination is concealed and the volume is manufactured to read as public opinion. GUDEA describes the structure such operations leave in account behavior rather than attributing them to an operator.

What does an influence operation look like in data?

In GUDEA's Texas and Kansas brief: an origin account with one post, reposted 12 times in 23 seconds; four relay accounts within 38 minutes; broad pickup; then 625 accounts registered the same day reposting the same three posts 941 times in one hour, 93 in a single minute, after which 97 percent went dormant. Registration bursts, a three-post signature, co-posting, and dormancy are the signals.

How is an influence operation different from coordinated inauthentic behavior?

Coordinated inauthentic behavior is the behavior: accounts working together while concealing that they are connected. An influence operation is the campaign that uses it toward a goal. GUDEA's research measures the behavior; its brief explicitly makes no attribution about who is running a campaign or why.

Can an influence operation amplify a true claim?

Yes, and GUDEA's method does not depend on the claim's truth. The Texas and Kansas brief states that it describes how an allegation was amplified and does not assess the allegation. Presley has said GUDEA does not serve as an arbiter of truth; the object is how content is repurposed and disseminated in a coordinated way.

How does GUDEA detect influence operations?

By classifying account behavior across more than 500 platforms rather than reading content: registration timing decoded from X user IDs, posting signatures, co-posting at minute resolution, dormancy between exports, proportion of volume from a small cohort, and cross-narrative overlap. The Texas and Kansas brief publishes the signal weights and the full ranked account list.