AI Agents in Abundance: Who Writes the Rules When the Applicant Is a Machine?

November 13, 2026
Julie Sowash
0 Comments

Happy Friday, Job Board Doctor friends!

The 2026-2027 Global Recruitment Site Platforms Survey is still open. Fifteen minutes, anonymous, and respondents who leave an email get the full report before publication. Take the survey at SurveyMonkey.

LAST WEEK’S POLL RESULTS

Last week I asked how the midterms will affect hiring budgets. The results: 43% of you say budgets shrink further, 35% say no change, and 22% say budgets are about to blow up, in a good way. Nearly four in five of you expect flat or down. A little more pessimistic than I expected.

Bar chart of last week’s reader poll on how the midterms will affect hiring budgets. 43% said budgets are going to shrink further, 35% said no change, and 22% said budgets are about to blow up. 78% expect budgets to stay flat or shrink.

Three stories landed this week that on their face seem unrelated, but I see connections. A seat on a standards committee, a layoff report, and an earnings call. Three stories, one intertwining thread. AI in hiring is moving out of the layoff announcement and into the plumbing, and plumbing gets designed in rooms most of us are not in.

The question for 2027 is not whether AI agents will apply to your jobs. It is who wrote the rules they follow, and how those of us not at the table will be impacted by the decisions made by others.

RECRUITICS TAKES THE CHAIR

On November 3, Recruitics joined the Linux Foundation’s Agentic AI Foundation (AAIF) as a Silver member. CTO Austin Anderson was named to lead a new Jobs and Hiring special interest group inside its Applied Industries Committee, chaired by Microsoft’s Andrew Nolan (GlobeNewswire).

AAIF launched in December 2025 with contributions from Anthropic, OpenAI and Block. It is the home of the Model Context Protocol, the standard agents use to connect to tools (Linux Foundation).

For our industry, Recruitics arrived with a draft already written. The Open Job Context Protocol (OJCP), authored by Anderson and launched July 29 (EIN Presswire), asks a job board or career site to publish a small file telling agents what they can do on the site.

The site then exposes up to six tools that run from searching jobs to submitting an application. Agents sign their requests. Results can carry a fit score. Identity checks route through third-party verifiers (OJCP).

Look at the steering committee: Workday, Cross Country Healthcare, Recruitics, HiringCafe, LoopCV, aiApply, scale.jobs and Tink. Three of the eight are auto-apply tools. Indeed and LinkedIn are not on the list.

In July I wrote about Recruitics placing Maki People’s voice AI, Mochi, inside its AdaptiveApply flow, which moved screening ahead of the application (Job Board Doctor).

The release leaned on one safeguard: a human always makes the final call. The series that followed documented what sat underneath that line. Mochi’s bias audit did not test for disability or veteran status. And by Maki’s own answer, the human review override rate had never been measured.

That is the track record that matters here. Recruitics took a screening product to its US market before the basic questions about what it does, and to whom, had answers. The same company is now writing the protocol many of us will be asked to implement.

This is about process, not motive, and I fully support a protocol and a process for the industry. It is something we have failed to do on so many other parts of our advertising standards. A standard for how agents search, score and apply carries fit scores and identity checks into every board that adopts it. The Maki series showed what happens when the impact questions get asked after launch instead of before.

“

“Every industry that sits it out will inherit someone else’s choices.”

Recruitics CEO Adam Stafford

He is right. Job boards are about to inherit his.

The operator question is the one nobody has answered, as far as I have seen. Publishing the file is cheap. What it changes is not. Today a job seeker comes to your site, sees your pages, your sponsored listings and your brand, and clicks apply.

Under OJCP, an AI agent skips the site. It asks your system for the job data, gets it back, and can submit the application without a person ever loading a page. Your board stops being a storefront and becomes a warehouse. The aggregation layer gets routed around.

What does a board get paid for when nobody walks through the door? If you know the answer, tell me and I will share it with the group.

AI IS LEAVING THE LAYOFF NOTICE

Employers announced 43,994 job cuts in October, the lowest October since 2023 and down 71% from a year ago. AI has been cited for 121,401 cuts this year, about 20% of the total, but for only 1,265 in October, its lowest month of 2026. Market and economic conditions have passed AI as the year’s top stated reason (Challenger, Gray & Christmas).

One caveat: Challenger counts announced cuts and the reasons employers choose to give. It is a record of what companies say. My read, and it is inference: AI is fading as a reason to cut and settling in as infrastructure.

ZIPRECRUITER GROWS IN A MARKET THAT IS NOT HIRING

ZipRecruiter’s third quarter is the proof of concept. Revenue rose 8% to $124.4 million, with net income of $5.1 million. Paid employers grew 10% to 73,929 while revenue per paid employer slipped 2% to $1,683. Performance-based revenue grew 13% to $31.1 million, now a quarter of the total. Full-year guidance moved to $468 to $472 million from a flat outlook at the start of the year (SEC).

CEO Ian Siegel credits product, not the market, which he calls “low hire, low fire.” ZIP is selling employer-candidate conversations instead of applications, and it says employer response rate per application tripled.

Traffic from AI platforms nearly doubled after direct integrations with ChatGPT and Claude. Siegel told analysts job seeking “will have agents in abundance” (Motley Fool).

A same-week footnote from DHI Group: ClearanceJobs revenue rose 18% to $16.5 million and passed Dice at $15.7 million (down 13%) for the first time. Before anyone declares the winner, $2 million of that came from an acquisition. Organic growth was 4% (DHI Group).

TIE IT ALL TOGETHER

▪ Job boards: your apply flow is being standardized by a committee you are not on, led by a vendor whose last launch left the impact questions for later. The spec is public. Read it this month.
▪ Employers: you are promised verified agents and a fit score. Ask who computes the score and who tested it for adverse impact. Ask before launch this time.
▪ Vendors: ZIP negotiated directly with the AI platforms. Everyone else gets someone else’s standard.

ONE MORE THING: NVIDIA

Jensen Huang resigned Thursday as Nvidia’s president, CEO and board member, effective immediately, after 33 years. CFO Colette Kress is interim CEO and keeps the CFO job. The board is searching for a permanent successor (SEC).

The why is covered everywhere else. I am interested in the org chart. The company at the center of the AI buildout had no named successor, and one executive now holds two jobs. Most of us run organizations with far less bench than Nvidia.

Do you have a succession plan for your org?