🤖 AI Agents

ASAPP Deploys a 5-Agent AI Workforce for Enterprise Customer Service

Customer experience platform ASAPP yesterday unveiled five purpose-built AI agents that work together inside its enterprise CXP platform to run customer service end-to-end — without a human in the loop for routine interactions.

Here’s how the team divides the work:

ASAPP’s CEO put it plainly: “An AI agent proves its value not in a single brilliant response, but in its consistency across interactions — remaining accurate, safe, responsive, and protective of privacy at every turn.”

What this means for your business: The “one chatbot” era is over. The companies winning with AI aren’t deploying a single tool — they’re deploying coordinated AI workforces with defined roles, handoffs, and feedback loops. If your customer service still runs on a single bot or a static FAQ page, you’re already two generations behind.

OpenAI Breaks Free from Microsoft’s Exclusive Grip — Models Now Available Anywhere

Source: Ars Technica  |  Business Insider

After seven years as Microsoft Azure’s exclusive AI engine, OpenAI has rewritten the deal. Starting now, OpenAI models can run on Amazon Web Services Bedrock, Google Cloud, and any other provider — not just Azure. Microsoft keeps its non-exclusive license through 2032 and Azure remains the “primary” partner, but the monopoly is gone.

Why it matters for you: If your team uses AWS or Google Cloud and felt locked out of GPT-class models, that wall just came down. Expect Amazon Bedrock to add OpenAI models “in the coming weeks,” per Amazon CEO Andy Jassy. More competition = better pricing and more choice for builders.

Xiaomi Open-Sources MiMo-V2.5: A Lean, Affordable Model Built for Agentic Tasks

Source: VentureBeat

Xiaomi released MiMo-V2.5 and V2.5-Pro under the MIT License — free to use commercially, fine-tune, and deploy. The models use a Mixture-of-Experts (MoE) architecture that delivers strong agentic performance at a fraction of the compute cost of larger models. Xiaomi also launched companion voice tools (TTS + ASR) for a complete, open-source voice pipeline.

Why it matters for you: MIT-licensed means zero royalties, zero usage restrictions. Small businesses and developers can now run capable agentic AI — including voice — on their own infrastructure without paying per-token API fees. The open-source stack for building AI employees just got a lot more accessible.

📰 AI News

Accenture Deploys Copilot to All 743,000 Employees — The Biggest Enterprise AI Rollout Ever

Source: Reuters / MSN  |  SiliconAngle

Microsoft just landed its largest-ever enterprise deal: Microsoft 365 Copilot is rolling out to all 743,000 Accenture employees worldwide. Early internal data is striking — 97% of Accenture employees using Copilot report completing routine tasks up to 15x faster. The deal is a major push by Microsoft to convert its 450 million+ Microsoft 365 users into Copilot subscribers (currently only ~3% pay the $30/month add-on).

The takeaway: When the world’s largest consulting firm bets its entire workforce on a single AI tool, it sends a signal to every enterprise buyer sitting on the fence. If your organization runs Microsoft 365, the business case for Copilot just got a very large case study behind it.

DeepSeek Slashes V4-Pro Prices 75% — And the AI Price War Heats Up Again

Source: Blockonomi

China’s DeepSeek dropped a promotional bombshell: its new V4-Pro model — which outperforms GPT-5.4 and Gemini on coding benchmarks and is optimized for Huawei chips — is now 75% cheaper through May 5. API cache hit pricing was cut by 90% across its entire model portfolio. Input costs dropped from $1.74 to $0.44 per million tokens; output from $3.48 to $0.87.

The takeaway: DeepSeek’s R1 triggered a global price war last year. V4-Pro could do it again. If you’re building AI-powered workflows or products and paying premium API rates, now is the time to benchmark DeepSeek V4-Pro against your current stack — especially for coding, data analysis, and agent tasks.

Australia Launches Its First AI Workplace Forum — Governments Are Starting to Set the Rules

Source: HCA Magazine

Australia’s Albanese government launched a formal tripartite AI Employment and Workplaces Forum today — bringing together government, business councils, and labor unions to shape how AI gets used at work. The forum will meet at least three times a year. The government’s stated focus: preventing “work intensification” (AI-driven surveillance and longer hours) over outright job loss, and building AI adoption on five principles: trust, capability, transparency, safety, and productivity.

The takeaway: Australia joins the EU, UK, and US states in formalizing AI-at-work policy. If you’re deploying AI tools that touch your employees’ workflows — monitoring, task automation, performance tracking — expect compliance requirements to arrive sooner than you think. Start documenting your AI use cases and human oversight protocols now, before regulators require it.

🚀 Want AI working for YOUR business? Most companies are experimenting with AI chatbots. We deploy AI workforces — AI Employees that follow up on leads, resolve support tickets, publish content, chase invoices, and screen 200 job applicants overnight so your hiring manager starts Monday with the top 10. Each role has a cost profile and human oversight, managed through one platform. This newsletter? Written by an AI Employee, approved by a human — so our team stays focused on what only humans can do.

AIToken Labs helps businesses design their AI Workforce Operating Model — starting with the 2-3 roles that deliver ROI in the first 60 days.

Book a free 40-minute AI Workforce Blueprint Session.
→ https://schedule.aitokenlabs.com/blueprint/40min