Quick answer: The Adecco Group, one of the world's largest staffing companies, is deploying Salesforce's Agentforce Coworker to 27,000 employees across more than 40 countries. The AI assistant, powered by Anthropic's Claude, helps sales staff and recruiters find prospects, prepare briefs and launch AI agents for candidate screening and onboarding. The rollout is a live test of how AI coworkers change daily work.

Most companies talk about AI at work in the future tense. Adecco has moved into the present tense, and at scale.

On September 15, 2026, the group announced it is rolling out Agentforce Coworker following a pilot in the UK and France. It is one of the largest workforce-facing deployments of an AI assistant announced so far, and it comes from a company whose whole business is matching people to jobs. That makes it worth reading closely, both for what it promises and for what it leaves out.

What Is Agentforce Coworker?

Agentforce Coworker is Salesforce's AI "teammate" for enterprise staff. According to Adecco, it is built directly into the Salesforce platform and runs on Anthropic's Claude. Instead of an employee opening a dozen tools to find a client note, a candidate's history and a lead's status, the Coworker sits in one place and pulls it together.

Adecco says the tool supports three areas of work:

• Sales: finding priority prospects, preparing sales briefs, enriching prospect data and checking lead status across teams.
• Recruitment: selecting candidates and launching the pre-screening, recruiter or onboarding agents.
• Client and candidate engagement: using shared context so the person on the call already knows the history.

The tool will also draw on context from more than 2.5 million agent-candidate interactions recorded since April 2025.

The Adecco Rollout at a Glance

• Announcement date: September 15, 2026
• Scope: 40+ countries, 27,000 employees
• Pilot markets: UK and France
• Underlying model: Anthropic's Claude
• Existing agentic AI: Recruitment workflows in ten countries, which Adecco says represent 50% of Adecco business revenues
• Contract backdrop: Multi-year agreement with Salesforce signed in March 2026, running through 2027, for unlimited access to Agentforce 360 under its license terms
• Adecco's stated goal: More than half of Adecco revenues powered by agentic AI by the end of 2026

What the Numbers Do and Don't Tell Us

This is where careful readers should slow down. Adecco has shared impressive figures, but almost all of them come from the company itself.

What was reported:

• Adecco's CEO Denis Machuel said at Dreamforce that recruiters have saved 35% to 40% of their time through agentic AI, and that the business has placed 20,000 more people this year, which he attributed to better conversations. These were relayed by CX Today and noted by AI News as company-provided and not independently verified.
• In March, Adecco reported that its UK agents had helped generate 15% time savings, plus shorter time-to-fill, higher fill rates and lower cost-to-serve.
• Salesforce said on X that Adecco handles 2 million service calls with 30% resolved by AI agents.

What is missing or unclear:

• The September statement gives no usage or productivity data from the UK and France pilot. "Strong adoption in a matter of days" is a claim, not a metric.
• The 27,000 figure was described as recruiters in March and as employees in September.
• The "50% of Adecco business revenues" line does not make clear whether it refers to group revenue or the Adecco unit, which sits alongside Akkodis and LHH.
• Interaction counts differ (2.5 million since April 2025 versus 2.6 million this year), and neither source explains how an "interaction" is counted.
• There is no completion timeline for the 40-country rollout.

None of this means the results are wrong. It means the honest reading is "promising, self-reported, not yet proven." A time saving is also not the same as a business outcome: freed hours only matter if they turn into better placements, better candidate experience or better work.

Benefits: Who Gains From an AI Coworker?

The points below combine what Adecco has claimed with analysis of where the value is most plausible.

For recruiters and sales teams

The most concrete promise is less tool-hopping. Recruiters spend a large share of their day gathering context before they can have a useful conversation. If one interface holds candidate history and client context, that prep time shrinks. Adecco's reported 35–40% time saving, if it holds up, would be substantial, but it remains unverified.

For candidates

Faster screening and updates are the visible upside. In a demo reported by CX Today, a voice agent named Ada told a caller why their application was paused, took a new certification number, verified it and updated the application during the same call. Candidates who currently wait days for a status update would notice that difference.

For clients

Clients hiring through a staffing firm care about time-to-fill and fill rate. Adecco says its UK agents improved both, alongside lower cost-to-serve, though the figures were self-reported.

For the business itself

A single, unified interface across 40+ countries gives a global company consistent workflows, cleaner data and a way to scale without adding headcount at the same rate. This is the commercial logic behind Adecco's target of more than half of revenues powered by agentic AI.

Concerns: What People Are Right to Worry About

This section is analysis, not reporting. Neither source describes employee sentiment, and Adecco has not published any.

1. Job security and role drift

Adecco frames the tool as freeing people for "what only humans can do." But when 35–40% of a recruiter's time is reported as saved, the natural question is what happens to that capacity: redeployment, higher targets, or fewer roles? Adecco has not said. Employees will ask, and silence tends to be read pessimistically.

2. Accuracy and over-reliance

AI systems can be confidently wrong. In recruitment, a wrong summary of a candidate or a mis-scored screen has real consequences for a real person's livelihood. The risk is not just errors; it is people trusting the AI's output and checking it less over time.

3. Bias and fairness in hiring

Recruitment AI is one of the most scrutinised uses of the technology. Under the EU AI Act, AI systems used for recruitment and worker management are generally treated as high-risk, which brings obligations around transparency, oversight and data quality. With Adecco operating across dozens of countries, compliance will vary by market. Candidates will reasonably want to know when an AI agent is screening them and how decisions are reviewed.

4. Data privacy across borders

A tool that knows candidate history, client context and millions of past interactions is only as safe as its data controls. Different countries have different rules on candidate data, retention and consent. Putting everything behind one interface raises the stakes on getting access controls right.

5. Surveillance and performance pressure

An AI layer that sits inside every workflow can also measure every workflow. Employees may worry, fairly, that "AI-assisted productivity" becomes a new set of metrics used to rank them. How the data is used internally matters as much as how the tool is used externally.

6. Skill erosion, especially for newer recruiters

If agents handle screening and prep, early-career staff may get less practice at the fundamentals: reading a candidate, probing a CV, building a client relationship. Organisations that automate the training ground can find themselves short of experienced people later.

7. Uneven adoption and change fatigue

"Strong adoption in days" is a pilot result in two markets. A rollout across 27,000 people in different countries, languages and team cultures will not land evenly. Some teams will embrace it; others will route around it.

8. Vendor dependence

Adecco's agentic strategy is built on a multi-year, unlimited-access Salesforce agreement. That buys speed and integration, but it also concentrates dependency on one platform and one model provider.

How AI Coworkers Are Changing Work Culture

Whatever the final numbers show, the shape of the change is already visible.

From tools to teammates. The word "coworker" is deliberate. When software is framed as a colleague that prepares briefs, chases leads and launches other agents, people start to delegate to it the way they would to a junior teammate. That changes how work is assigned, reviewed and credited.

From searching to asking. The old pattern was to open several systems and assemble context by hand. The new one is to ask a single interface and refine the answer. That rewards employees who are good at framing questions and judging outputs, and it rewards less the ability to know where every piece of data lives.

Human judgment as the premium skill. Adecco's CEO described the goal as deploying AI alongside people, saying it "had to happen with people and not to people," as reported by AI News from the Dreamforce keynote. Whether or not you take the framing at face value, it points to where value is shifting: empathy, negotiation, discernment and accountability become the parts of the job that stay human.

Speed becomes the default expectation. If a status update happens during a phone call instead of three days later, candidates and clients will expect that everywhere. Faster cycles raise the baseline for every team.

Managers manage humans and agents. Team leads will increasingly be responsible for workflows that mix people and AI agents, including deciding what is delegated, what is reviewed and who owns the outcome.

Trust becomes a cultural asset. The organisations that get this right will be transparent about what the AI does, honest about what happens to freed capacity and clear about where humans make the final call. The ones that don't will face quiet resistance regardless of how good the technology is.

What Leaders Should Take From Adecco's Move

• Ask for independent evidence. Treat vendor and self-reported time-saving numbers as hypotheses to test in your own environment.
• Decide in advance what freed time is for. Say it out loud to employees.
• Keep humans accountable for decisions that affect people, especially in hiring.
• Invest in training for judgment, not just tool usage.
• Plan for regional compliance, especially where AI is used in recruitment.

Frequently Asked Questions

What is Agentforce Coworker?

Agentforce Coworker is Salesforce's enterprise AI assistant, embedded in the Salesforce platform and, according to Adecco, powered by Anthropic's Claude. It helps employees with tasks such as finding prospects, preparing sales briefs, enriching data and launching recruitment agents.

How many Adecco employees will use it?

Adecco says the rollout covers 27,000 employees in more than 40 countries, following a pilot in the UK and France.

Will Agentforce Coworker replace recruiters?

Adecco says the aim is to free recruiters and sales staff for richer engagement with candidates and clients. The company has not published headcount plans, so it is not possible to say from public information what the long-term effect on roles will be.

What time savings has Adecco reported?

Adecco's CEO said recruiters have saved 35% to 40% of their time through agentic AI, and the company reported 15% time savings in UK recruitment in March. Both are company-provided figures that have not been independently verified.

What are the main concerns about AI coworkers in recruitment?

Common concerns include job security, accuracy, bias and fairness in hiring, cross-border data privacy, performance surveillance, erosion of junior skills and vendor dependence.

How does this change workplace culture?

It shifts work from searching across tools to asking a single AI interface, raises speed expectations, and places more value on human judgment, empathy and accountability. It also makes trust and transparency central to how AI is introduced.

Conclusion

Adecco's rollout is a meaningful signal: agentic AI is moving from pilot slides to the daily workflows of tens of thousands of employees. The benefits it describes are plausible, and the direction of travel toward AI coworkers handling context and preparation while humans handle judgment is clear.

But the evidence is still mostly the company's own. Until independent data on productivity, candidate outcomes and employee experience emerges, the fair verdict is: a serious, large-scale experiment worth watching, with real benefits on offer and real questions still open. The businesses that learn the most from it will be the ones that ask those questions early.