Nothing matches that. Try a different word or tag.
Oct · 9 min read
Banks and insurers can use AI agents to buy advertising, but the firm stays responsible for every promotion. A 12-point checklist of the controls that matter, sourced to FCA and FINRA rules.
Oct · 9 min read
One AI buying agent can trade across the UK, EU and US, but not on one rulebook. How consent, targeting limits and AI law differ by market, with a comparison table and a setup checklist.
Oct · 9 min read
A copyable incident response plan for AI media-buying agents: five incident types, severity levels, six roles, first-hour and first-day steps with suggested timings, a post-incident review template and how to restore autonomy.
Oct · 9 min read
A copyable media buying mandate template for AI agents: ten sections of labelled fields, from objective and budget to approval thresholds and stop conditions, each with a hypothetical example value.
Oct · 9 min read
A numbered checklist for CTV buyers of any size: the twelve deal terms that commit money, define delivery or expose the brand, and the approval rule to set for each before an AI agent can book.
Oct · 9 min read
A printable checklist for agencies: 25 yes or no checks across mandate, testing, counterparty identity, deal records and escalation, to answer before an AI agent commits a client's money.
Oct · 9 min read
A weekly reporting template for agencies running AI buying agents: seven sections and the fields in each, from spend against the mandate to deal-record reconciliation, plus a monthly add-on.
Oct · 9 min read
IAB Tech Lab's OpenProposal 3.0 comment period, now extended to 10 November 2026, had 52 filings from 17 GitHub accounts by its original 22 October deadline. What they raised, grouped by theme, and what happens next.
Oct · 9 min read
How the EU AI Act's risk classes, transparency rules and human oversight principles apply to AI agents that buy and sell media, with the dates obligations apply and a practical checklist for buyers and publishers.
Oct · 9 min read
Agent negotiations stall when mandates do not overlap, rounds run out or nobody has authority to move. A worked hypothetical shows the rounds, the stall, the escalation and what gets recorded.
Oct · 9 min read
The mistakes that recur when brands first hand media budgets to AI agents, from unwritten mandates and skipped benchmarks to unchecked counterparties, with a practical fix for each.
Oct · 7 min read
When a governance agent is configured, AdCP's check_governance validates a buy against a human-approved plan, escalates actions above set thresholds to a person and logs who decided what.
Oct · 7 min read
The questions finance should put to any AI media buying proposal: authority and thresholds, accountability, deal records, audit access, data, exit and pilot sizing, with a hypothetical approval memo.
Oct · 7 min read
Why holdout and geo tests beat attribution for agent-bought media, why the test design belongs in the mandate, how deal records tie spend to test cells, and a hypothetical geo test.
Oct · 7 min read
Caps set inside each deal do not add up to a campaign cap. What the mandate and deal terms need to say about household versus device, unrecognised users and shortfall, and what a buyer can enforce alone.
Oct · 7 min read
DOOH suits agent negotiation, but its impressions are audience estimates and its delivery evidence is a play log. What the deal record must specify on screens, dayparts, audience source and proof of play.
Oct · 7 min read
What triggers a make-good in an agent-negotiated deal, which decisions an agent can take and which need a person, how to record the remedy, and a hypothetical worked example.
Oct · 6 min read
A practical RFP question list for agentic advertising platforms: mandates and approvals, AAMP versions, Agent Registry listing, deal types, the deal record, data handling, audit and exit.
Oct · 7 min read
How to scope a 30-day agentic advertising pilot: the mandate, approval thresholds, what to measure, stop criteria, a week-by-week plan and a hypothetical budget split.
Oct · 7 min read
The seven items a buying agent's decision log should record, why the mandate version matters, what a log cannot show, and why a log is not the same as an agreed record of the deal.
Oct · 7 min read
What changes when price depends on outcomes: measurement definitions agreed up front, attribution windows, who measures, how disputes are settled, and what a mandate should allow an agent to accept.
Oct · 7 min read
Where clean rooms support agent-negotiated deals before, during and after a campaign, what agents can and cannot do with them, and why lawful basis stays with the data controller under UK GDPR.
Oct · 7 min read
An assisted auction, a two-track market or fragmented agent negotiation. Three scenarios for programmatic in 2030, the signals that point to each, and the moves that work in all of them.
Oct · 7 min read
Deal share, standardised proposals, identifiable agents, negotiation over non-price terms and a shared deal record: five indicators a buyer can check to see whether the shift is real.
Oct · 7 min read
A three-way media match of agreed terms, delivery and invoice, why it depends on one shared deal record, the questions to ask vendors, and a hypothetical month-end example.
Oct · 7 min read
Five checks before a publisher accepts an AI buying agent's terms: who the principal is, mandate evidence, Agent Registry listing, negotiation behaviour and a shared record of what was agreed.
Oct · 7 min read
Generative AI produces content such as copy and images. Agentic AI takes actions with consequences, such as committing budget, within a mandate. How they differ, how each is governed and how they combine.
Oct · 7 min read
A hypothetical walkthrough of an agentic media buy: brief, mandate, seller product discovery, negotiation rounds, escalation to a human, approval, a shared deal record and a delivery check.
Oct · 7 min read
Buyer agents compare offers across sellers and never tire of countering. Publishers need floors written as policy: list price, walk-away price, fixed quality terms and clear concession rules.
Oct · 7 min read
Agents change how deal terms are agreed, not how deals are activated. The deal ID still links what two agents agreed to what trades in the auction, and most failures happen at that join.
Oct · 7 min read
Publishers need five things to accept deals from autonomous advertising agents. Four are documents and decisions a sales team can produce; only the agent itself needs engineering.
Oct · 7 min read
AAMP and AdCP standardise how agents communicate and transact. Settling a media deal also needs a shared record of terms, discrepancy rules and an accountable party, which sit in a marketplace or contract.
Oct · 7 min read
SSAI, app and device spoofing all misrepresent where an impression came from. A shared deal record fixes what was agreed and who agreed it, but detection still belongs to verification.
Oct · 5 min read
How does agent-to-agent negotiation apply to CTV inventory? Agent-to-agent negotiation applies to CTV inventory through the same deal-based mechanism it applies to any programmatic inventory, but with…
Sep · 5 min read
When neither party to a deal is human, the question of who verifies and who approves becomes structural rather than operational. A shared deal record with human approval rights is not optional.
Sep · 5 min read
Agents draft and recommend. Humans approve at named decision points. Here is what that approval gate covers, what it does not, and how the deal record makes it real.
Sep · 5 min read
When an agent commits to inventory on a buyer's behalf, accountability is defined by the approval record and deal terms. Here is what buyers need in the deal structure.
Sep · 5 min read
When agents act on behalf of buyers and sellers, the deal record is how disputes are resolved, compliance is demonstrated, and performance is verified.
Sep · 5 min read
Deal structure, approval rights, and audit trail differ significantly between AI-negotiated CTV deals and open exchange buys. Here is how, and why the DealSheet is the governing layer.
Sep · 7 min read
When agents act on behalf of buyers and sellers, both parties need a shared, authoritative record of what was agreed. That is what the DealSheet provides.
Sep · 4 min read
Alkimi is a neutral agentic advertising marketplace where buy-side and sell-side agents negotiate bilateral deals directly, producing a shared DealSheet that both parties hold independently. This piece explains how the connection is made, how proposals ...
Sep · 4 min read
A DealSheet is the bilateral record of agreed deal terms written at the time of negotiation and held by both the buyer and publisher independently. It captures what was agreed before any impression is served: the CPM, the inventory scope, the audience c...
Sep · 5 min read
A tamper-proof deal record has three structural properties: it is written at the moment of agreement rather than reconstructed later; it is held by both parties independently rather than only by the marketplace; and it is held by a neutral marketplace t...
Sep · 4 min read
When a dispute arises over an agent-negotiated deal, the DealSheet is the reference point. The marketplace provides deal record retrieval; both parties compare their delivery data against the agreed DealSheet terms. The neutral marketplace holds the rec...
Sep · 4 min read
Agent-to-agent deal governance is the set of rules, records, and oversight mechanisms that ensure AI agents on both sides of a media transaction operate within authorised boundaries and produce auditable outcomes.
Sep · 4 min read
Bilateral deal ownership means both parties to an advertising deal hold the same document: the agreed terms written to a shared record that neither party can alter unilaterally. In conventional programmatic, each party holds their own version. The bilat...
Sep · 4 min read
Publishers set deal terms for autonomous buying agents through a sell-side mandate: a human-approved governance document that defines the parameters within which the sell-side agent can act. A well-structured mandate covers CPM floors, inventory categor...
Sep · 5 min read
A CTV publisher preparing to receive agentic proposals needs a sell-side mandate that goes further than a standard programmatic floor configuration. It needs CPM floors by content category and daypart, buyer category restrictions, audience data conditio...
Sep · 5 min read
There are three concrete checks for whether a buy-side agent has negotiated fairly on your behalf. First, retrieve the DealSheet and confirm the agreed CPM is within your mandate parameters. Second, review the agent's decision log to confirm it applied ...
Sep · 5 min read
A buyer running agent-negotiated deals should have three audit rights: deal record access, decision log access, and delivery comparison capability. Deal record access means the ability to retrieve the DealSheet for any deal independently of the counterp...
Sep · 5 min read
In conventional programmatic, human campaign controls are platform configuration settings that enforce buying behaviour at the DSP level. In agent mandate governance, controls are expressed as mandate parameters that the agent evaluates before acting. T...
Sep · 4 min read
An agent-to-agent marketplace provides four governance functions: identity verification, mandate enforcement, deal record integrity, and audit access. Together, these functions mean that buyers can demonstrate every deal their agent agreed was within th...
Sep · 4 min read
In conventional programmatic, there is no single document both parties agreed before the campaign ran. In agentic advertising, the bilateral DealSheet is written at negotiation time, timestamped, and held by both parties independently. This piece explai...
Sep · 5 min read
Publishers need three things before they can accept deals from autonomous buying agents: an integration built to an open agentic standard, a sell-side mandate for autonomous acceptance, and deal record infrastructure.
Sep · 5 min read
Publishers evaluating marketplaces for agent-negotiated deals should look for four things: AAMP support, sell-side agent support with mandate governance, a bilateral DealSheet held by both parties, and a neutral marketplace structure.
Sep · 5 min read
Connected television inventory has long been traded through direct insertion orders or programmatic guaranteed arrangements that require significant human involvement at every stage. Agentic buying changes the process, not the principles. A buy-side age...
Sep · 3 min read
Programmatic deal transparency is the ability of buyers and publishers to independently verify the terms, execution, and delivery of a media deal. For US buyers, it is both a commercial imperative and an increasingly regulated requirement.
Sep · 3 min read
A deal sheet in programmatic advertising is a shared, structured record of the terms agreed between a buyer and a publisher before a campaign runs. It is not a deal ID. It is a bilateral document that both parties hold and can audit independently.
Sep · 3 min read
Most US media buyers do not hold a verifiable record of what they agreed in their programmatic deals. This article explains what a verifiable deal record looks like, why it matters, and how to get one.
Sep · 3 min read
Publishers working with autonomous buy-side agents need to define deal parameters their sell-side infrastructure can communicate and enforce. This article explains what that looks like in practice.
Sep · 3 min read
AI agent negotiation in programmatic advertising means structured, mandate-driven deal-making between buy-side and sell-side agents, not simply faster bidding. This guide explains how it works and what US media buyers need to understand before adopting it.
Sep · 4 min read
A machine-readable ad deal for agent-to-agent negotiation expresses every parameter in a format an agent can evaluate: party credentials, inventory spec, CPM terms, audience conditions, delivery commitment, and approval thresholds.
Sep · 3 min read
The two main standards for agent-to-agent advertising: IAB Tech Lab's AAMP, the industry framework for agent communication and deal workflows, and AdCP, a separate protocol published by AgenticAdvertising.org.
Sep · 4 min read
Publishers verify agent-negotiated deals through three layers: agent identity verification, mandate verification against proposed terms, and deal record verification using the shared DealSheet post-campaign.
Sep · 4 min read
Agent-to-agent deal reconciliation works from the shared DealSheet outward. Both parties agreed the terms before any impression was served; the DealSheet is the reference point for post-campaign verification.
Sep · 5 min read
The structural difference between CPMs won in auction and CPMs agreed in bilateral negotiation: how prices are set, what patterns result, and how to calibrate mandate CPM parameters correctly.
Sep · 4 min read
A definition of a bilateral deal record: what it contains, how it differs from a deal ID or an impression log, and how to evaluate whether a marketplace provides genuinely bilateral records.
Sep · 4 min read
The structural difference between programmatic guaranteed (parallel DSP and SSP records) and a bilateral deal record (one shared DealSheet held by both parties), and why this determines post-campaign reconciliation quality.
Sep · 4 min read
How the composition and economics of open exchange change as premium inventory migrates to bilateral channels, and what remains in the exchange for the long term.
Sep · 4 min read
The three requirements for publisher readiness: protocol-compliant integration, a sell-side mandate framework, and deal record infrastructure. The most common gaps and how to prioritise readiness work.
Sep · 4 min read
Why verification vendors are expanding into deal record auditing: the three structural advantages they hold, what deal record auditing involves, and what buyers should ask their current verification vendors.
Sep · 7 min read
Six questions to assess whether a platform gives buyers genuine deal record access or just reporting against a record they cannot independently verify.
Sep · 6 min read
What documentation to request, how to run mandate-to-delivery reconciliation, and the questions that identify whether an agentic platform is actually auditable.
Sep · 6 min read
Under GDPR, data processing requires a lawful basis established by a legal person. A software agent has no legal standing to establish one. First-party data cannot enter agent negotiation without prior explicit authorisation from a human data controller.
Sep · 8 min read
What a DealSheet is, why programmatic needs it, and how it becomes the active negotiation layer in agentic advertising.
Aug · 5 min read
IAB Tech Lab's Shailley Singh: agentic advertising cannot rely on natural-language confirmation. Research across 90,202 simulated deals found 95.3% disagreement between separate agent records, falling to 0.19% with a shared record.
Jun · 41 min read
The three functions that define a working agentic marketplace, and why the order of agree, transact, verify cannot be reversed.
