Sabio Holdings and enterprise AI company Gentoro said on August 26, 2026, from Toronto and San Francisco that they will deploy AI agents inside Sabio's demand-side platform environment, starting with internal campaign management before any natural language self-serve interface reaches advertisers.
The announcement, issued at 07:00 Eastern Time, describes a collaboration rather than an acquisition, an investment or a commercial product release. No financial terms appear. No deployment date appears. No customer is named. What the release does specify is architecture, sequence and scope, and each of those carries more weight for buyers than the framing suggests.
What the two companies agreed
Sabio Holdings, listed on the TSX Venture Exchange under SBIO and on OTCQB under SABOF, describes itself as a Los Angeles-based creator-led, data-driven and AI-powered ad tech company working in ad-supported streaming. Gentoro is described as an enterprise AI company focused on operationalising AI across complex business workflows.
According to Sabio, the two companies will bring agentic AI into the Sabio DSP environment "beginning with internal workflows and execution connected to the launch of a natural language self-serve user interface." Working with Gentoro, Sabio will deploy AI agents across internal operations, initially focusing on campaign management.
The company positions the move as an extension rather than a departure. Sabio states that it has applied AI across its technology platform for years, including within its App Science household graph and natural language processing capabilities, and that agentic AI extends that work from intelligence and analysis into the operational workflows supporting the business.
Aziz Rahimtoola, CEO of Sabio Holdings, framed the arrangement around consolidation of the company's own components. "Our partnership with Gentoro enables agentic AI agents to streamline our workflows, while unifying our full tech stack: our DSP, SSP, SSAI, Creator TV, and App Science analytics capabilities. This will provide new ways for brands, agencies, and small businesses to work with Sabio in a natural language environment," he said.
Pervez Choudhry, CEO of Gentoro, described the target more narrowly. "Operational efficiency is one of the most immediate opportunities for agentic AI to create meaningful business value," he said. "Sabio already has a sophisticated AI strategy and a technology environment uniquely suited to taking the next step. We're helping translate that advantage into operational AI and create an architecture that can scale as new opportunities emerge."
The Campaign Management Agent and the protocol beneath it
The initial implementation uses Gentoro's Campaign Management Agent, referred to in the release as the CM Agent, to streamline cross-system workflows and reduce repetitive manual work. According to Sabio, Gentoro's agentic architecture is built on the Model Context Protocol, which the release says enables AI agents to work securely across Sabio's existing technology environment while providing a foundation for further agentic applications.
That single technical detail places the collaboration inside a well-documented industry trajectory. Model Context Protocolwas introduced by Anthropic in November 2024 and later donated to the Linux Foundation. It functions as a standard interface between large language models and external systems, and it has become the default connective layer across programmatic infrastructure rather than one option among several.
The adoption sequence is dense. Google released an open-source MCP server for its Ads API on October 7, 2025, read-only. Amazon Ads opened a closed beta for its own MCP server on November 13, 2025, then moved it into open beta at the IAB Annual Leadership Meeting on February 2, 2026. Yahoo DSP embedded agents that connect external systems through MCP on January 6, 2026. One day earlier, PubMatic launched AgenticOS with live campaigns already runningon NVIDIA-accelerated infrastructure.
The base specification is not settled ground. A revision dated July 28, 2026 removed sessions from the protocol and forced ad tech vendors to rebuild agent servers, tightening OAuth and OpenID Connect alignment in the same release. Building production tooling on a specification that changed weeks earlier is a live engineering cost, not a theoretical one, and the Sabio release does not address version pinning, authorisation scope or agent permissioning.
Internal first, external later
The most consequential line in the announcement is the ordering. Agents go to internal operations before they reach advertisers.
That sequencing separates this collaboration from most agentic announcements of the past twelve months, which have led with buyer-facing capability. It also sits in tension with Rahimtoola's own quote, which describes new ways for brands, agencies and small businesses to work with Sabio in a natural language environment. The release does not reconcile the two. No date, phase, beta programme or eligibility criterion is attached to the self-serve interface, and the phrase "connected to the launch of" leaves the relationship between internal deployment and external availability undefined.
The caution has company. Fluency disclosed in August 2026 that it blocks AI agents from executing against live budgetsacross roughly three billion dollars in managed spend, a deliberate architectural limit rather than a temporary one. StackAdapt reported more than 15,000 weekly internal AI-assisted workflows on July 28, 2026, a figure describing employee use rather than advertiser autonomy. The same company published research on August 19, 2026 finding that only 6% of marketers act on in-platform AI recommendations.
Survey evidence points the same direction. TripleLift's May 2026 study of 200 advertising professionals recorded 73% using AI for campaign optimisation against 19% fully automating campaign execution, with 67% citing lack of trust in AI output as the reason for keeping manual oversight. Starting inside the building, where errors cost internal hours rather than client budget, is consistent with that data.
What the release does not say
Several gaps are worth logging plainly.
The announcement names no underlying model or model provider. It gives no timeline, no pilot scope and no measurement framework. It provides no baseline for the operational efficiency it targets, which means any subsequent claim of improvement will lack a published starting point. It does not state whether the CM Agent operates with write access to live campaigns or in a recommend-and-approve pattern, a distinction that has become the central design question in agentic advertising.
That distinction matters because the market has split on it. When X shipped its Ads MCP server on August 24, 2026, PPC Land documented 23 tools, ten of which write to production ad accounts, with campaigns arriving paused so that activation requires a separate explicit call. Nexxen took the opposite position on April 6, 2026, embedding recommendations inside human-controlled workflows under a configurable autonomy model. Sabio's release places itself on neither side.
The release also carries a dual dateline of Toronto and San Francisco while describing Sabio as Los Angeles-based, a reflection of the company's TSX Venture listing rather than an inconsistency in the underlying facts.
The owned-stack argument
Sabio's stated rationale rests on ownership. According to the company, it owns and operates its DSP, SSP, App Science household graph and first-party data capabilities through Creator TV, and that unified environment creates an opportunity to extend agentic AI across systems and workflows.
The argument has structural logic. Agentic systems fail most often at the seams between platforms, where one system's output must be interpreted by another without a shared vocabulary. IAB Tech Lab argued precisely this in a technical position paper on April 9, 2026, holding that shared taxonomy identifiers, not natural language prompts, prevent compounding errors as agents pass instructions along a chain. A company operating buy side, sell side, server-side ad insertion and analytics under one roof faces fewer of those handoffs than one stitching together third-party systems.
Whether that translates into commercial advantage is unproven. Cross-platform agent communication has moved faster than single-vendor consolidation: AdRoll and PubMatic connected demand-side and supply-side agents over MCP in April 2026, allowing diagnostics that previously took days across DSP and SSP log files to resolve in minutes.
Where the money actually is
Agentic advertising remains a small revenue line against loud positioning, and the published figures are unflattering.
PubMatic's emerging revenues bundle, which includes AgenticOS, reached roughly 8.8 million dollars in a recent quarter, or 14% of total revenue, against more than 1,000 AI-powered deals and more than 30 fully autonomous campaigns on a platform processing 94 trillion impressions. Magnite chief executive Michael Barrett, in an interview published July 16, 2026, placed 2027 agentic advertising spend forecasts between negligible and 600 to 700 million dollars annually. Gartner has forecast that more than 40% of agentic AI projects could be abandoned by 2027.
There is at least one performance data point on the other side. On the August 25, 2026 edition of AdExchanger Talks, as documented by PPC Land the following morning, Butler/Till chief strategy officer Scott Ensign described an end-to-end CTV campaign run through PubMatic's agentic infrastructure inside an AI assistant, reporting an 80% reduction in supply chain and technology costs, a 40% increase in impressions on the same budget, a 98% video completion rate and waste below 1%. Most of that saving comes from removing fee layers rather than from superior decisioning, which is a different claim than the one vendors usually make.
Demand-side appetite is nonetheless documented. The IAB's 2026 Outlook Study, released January 28, 2026, surveyed more than 200 brands and agency buyers and found two-thirds prioritising agentic systems for autonomous campaign execution, within a forecast of 9.5% United States ad spend growth and 13.8% connected television growth. Snowflake research published March 10, 2026 found advertising and media firms leading all sectors on agentic AI production deployment at 42% against a 31% cross-industry figure.
Standards remain contested
Any agent architecture built now inherits an unresolved governance question. IAB Tech Lab formally named its Agentic Advertising Management Protocols initiative on February 26, 2026, setting out three pillars, seven protocol components mapped to existing standards and a three-level agent hierarchy. The competing Ad Context Protocol launched on October 15, 2025 with six founding companies, offering both MCP and Agent-to-Agent transport.
On August 20, 2026, six days before the Sabio announcement, IAB Tech Lab counted 13 functions where the two frameworks overlap. The organisation's Agent Registry reached 10 registered entries by March 11, 2026, all operating under the MCP standard. Sabio and Gentoro do not appear in the registry figures published to date, and the release does not reference AAMP, AdCP or any compliance posture.
Why this matters for the marketing community
For media buyers, the immediate significance is not a product they can use. It is a signal about where mid-cap independents are placing engineering effort.
The competitive set has been moving for eighteen months. Magnite added a buyer agent and SpringServe AI mediation on April 27, 2026 with Disney Advertising, Spectrum Reach, Kepler and MiQ among first testers. Madhive extended its Maverick platform with agentic capabilities and an MCP server on May 27, 2026, ahead of a partnership covering 117 local television markets. Culture Hive built agentic cultural scoring into CTV activation in June 2026. Independents that do not build in this direction risk being evaluated against those that have.
The second point concerns cost structure. Ari Paparo argued on July 21, 2025 that autonomous systems could automate the campaign setup, targeting and optimisation functions that justify DSP economics. If agents absorb operational labour inside a platform, the headcount that platform fees historically funded becomes negotiable. Sabio's decision to start internally is consistent with capturing that saving before offering the interface externally, and buyers evaluating renewal terms have a reason to ask what the saving is worth.
The third point is access. Rahimtoola's quote names small businesses alongside brands and agencies. Self-serve expansion has been a recurring pattern across the sector, and natural language interfaces lower the configuration burden that has kept smaller advertisers out of connected television buying. Whether Sabio's interface reaches that audience, and at what minimum spend, remains unstated.
The fourth is measurement discipline. Taboola research conducted in March 2026 across 200 marketing leaders found 76% reporting meaningful improvement from agentic campaign tools, with benefits concentrated almost entirely inside search and social walled gardens. Open-web and CTV agentic gains remain thinly evidenced. An operational efficiency claim without a published baseline, which is what this announcement currently offers, belongs in that thin category until numbers follow.
Timeline
- November 2024 - Anthropic publishes the Model Context Protocol as an open standard, later donated to the Linux Foundation
- October 7, 2025 - Google releases an open-source MCP server for its Ads API, read-only
- October 15, 2025 - Ad Context Protocol launches with six founding companies, offering MCP and Agent-to-Agent transport
- November 13, 2025 - Amazon Ads opens a closed beta for its MCP server
- January 5, 2026 - PubMatic launches AgenticOS with live campaigns already running
- January 6, 2026 - Yahoo DSP embeds agentic capabilities connecting external agents through MCP
- January 28, 2026 - IAB forecasts 9.5% US ad growth, with two-thirds of surveyed buyers prioritising agentic systems
- February 2, 2026 - Amazon Ads MCP server enters open beta at the IAB Annual Leadership Meeting
- February 26, 2026 - IAB Tech Lab formally names its AAMP initiative
- March 10, 2026 - Snowflake research finds advertising and media firms lead all sectors on agentic AI production deployment at 42%
- March 11, 2026 - IAB Tech Lab Agent Registry reaches 10 entries, all under the MCP standard
- April 6, 2026 - Nexxen launches an AI-native DSP interface built on configurable autonomy rather than full delegation
- April 2026 - AdRoll and PubMatic connect demand-side and supply-side agents over MCP for deal diagnostics
- April 27, 2026 - Magnite adds a buyer agent and SpringServe AI mediation with Disney, Spectrum Reach, Kepler and MiQ
- May 19, 2026 - TripleLift survey of 200 professionals records 19% fully automating campaign execution
- May 27, 2026 - Madhive extends Maverick with agentic capabilities and an MCP server
- July 16, 2026 - Magnite chief executive Michael Barrett places 2027 agentic spend forecasts between negligible and 700 million dollars
- July 28, 2026 - MCP specification revision removes sessions, forcing agent server rebuilds and tightening authorisation
- July 28, 2026 - StackAdapt discloses more than 15,000 weekly internal AI-assisted workflows
- August 2026 - Fluency discloses that it blocks AI agents from executing against live budgets across roughly three billion dollars in managed spend
- August 19, 2026 - StackAdapt research finds 6% of marketers act on in-platform AI recommendations
- August 20, 2026 - IAB Tech Lab counts 13 functions where AAMP and AdCP overlap
- August 24, 2026 - X ships an Ads MCP server exposing 23 tools, ten of them able to write to production accounts
- August 26, 2026 - Sabio Holdings and Gentoro announce an agentic AI collaboration beginning with internal campaign management workflows
Related PPC Land coverage
- Yahoo DSP embeds AI agents that actually run campaigns today - Documents the January 2026 rollout in which a demand-side platform connected external agents through MCP and allowed execution with human approval.
- PubMatic launches agentic advertising OS with live campaigns running - Covers AgenticOS, the supply-side infrastructure layer positioned for autonomous planning, transacting and optimisation.
- MCP forces ad tech to rebuild agent servers as sessions disappear - Explains the July 2026 specification revision that removed sessions and tightened authorisation for any vendor building on the protocol.
- X Ads MCP gives Grok and Claude Code 23 tools to run ad campaigns - Details the write-capable agent tooling shipped two days before the Sabio announcement, including the paused-on-creation safeguard.
- Nexxen's AI-native DSP UI signals a shift in who controls the machine - Describes the configurable autonomy alternative to fully agentic execution inside a demand-side platform.
- IAB Tech Lab names its agentic ad initiative AAMP to end market confusion - Sets out the three-pillar architecture and seven protocol components that agent implementations are being measured against.
- AAMP and AdCP overlap on 13 functions, IAB Tech Lab COO says - Records the standards contest still unresolved when the Sabio collaboration was announced.
- Agentic AI threatens traditional DSP business models says industry veteran - Lays out the argument that automating setup, targeting and optimisation undermines the economics justifying DSP fees.
- Fluency blocks AI from touching live ad spend across $3 billion in budgets - Presents the opposing architectural position, keeping agents away from live budgets entirely.
- Agencies turn AI tokens into a margin business as agentic spend stalls - Assembles the revenue and forecast figures showing how small agentic spend remains against the volume of announcements.
- Most advertisers use AI for optimization but not creative, TripleLift finds - Quantifies the gap between AI-assisted optimisation and full automation of campaign execution.
- Only 6% of marketers act on in-platform AI recommendations, StackAdapt finds - Measures how rarely practitioners follow through on automated recommendations already available to them.
- Gray Media bets on Madhive's AI DSP to win local TV's programmatic future - Covers a comparable agentic and MCP build by a television-focused demand-side platform.
- Magnite expands AI buyer agent and SpringServe mediation with Disney and MiQ - Documents the CTV supply-side equivalent of the workflow automation Sabio is pursuing.
Summary
Who: Sabio Holdings, the Los Angeles-based ad-supported streaming technology company listed on the TSX Venture Exchange as SBIO and on OTCQB as SABOF, and Gentoro, an enterprise AI company. Aziz Rahimtoola, CEO of Sabio Holdings, and Pervez Choudhry, CEO of Gentoro, provided the only attributed comments.
What: A collaboration to deploy agentic AI inside Sabio's demand-side platform environment, beginning with internal operations and initially focused on campaign management. The implementation uses Gentoro's Campaign Management Agent on an MCP-based agentic architecture, connected to a planned natural language self-serve user interface. No timeline, financial terms, customer names or performance baselines were disclosed.
When: Announced on August 26, 2026, at 07:00 Eastern Time.
Where: Issued under a dual Toronto and San Francisco dateline, covering Sabio's owned DSP, SSP, App Science household graph and Creator TV first-party data capabilities.
Why: Sabio frames agentic AI as an extension of an existing AI strategy from analysis into operations, targeting efficiency in cross-system workflows and reduced repetitive manual work. Gentoro frames operational efficiency as the nearest-term source of business value from agentic systems. The announcement lands amid documented industry-wide MCP adoption, an unresolved standards contest between AAMP and AdCP, and published evidence that agentic advertising revenue and practitioner trust both remain small relative to the volume of vendor activity.
Discussion