Microsoft unifies Copilot, Meta eyes agents: July 30 brief

Microsoft is folding its separate AI helpers into one app, a shift that could simplify software decisions for companies already paying for Microsoft 365. Meta also wants its AI agents handling customer conversations for millions of advertisers. Meanwhile, a new homeowner assistant shows how local service choices may move inside AI before a customer ever searches.

Microsoft puts its Copilots under one roof

Microsoft confirmed on July 29 that one Copilot “super app” will combine its chat, coding, and task-running features for consumers and businesses later this year. Task-running means the software can complete a series of steps on someone’s behalf. Chief executive Satya Nadella said the combined product will arrive this quarter, according to the company’s earnings-call coverage.

The practical consequence is fewer doors into Microsoft’s AI system. Today, a team can meet different Copilot products in Microsoft 365, GitHub, Windows, and the standalone chat service. A single app could reduce that confusion. It could also make Microsoft the default place for more office work, which matters when an owner is deciding whether another paid AI account still earns its place.

The product details remain incomplete. Microsoft has not given the final feature list, price, or an exact launch day. Its official quarterly release says Microsoft 365 Copilot has passed 30 million paid seats, so this consolidation is aimed at an established workplace product rather than a small experiment.

Our marketing stack audit with AI gives you a simple way to run that check without starting from the vendor list.

Meta wants agents paid for results

Meta told investors on July 29 that it plans to sell AI agents to businesses, starting with companies that already advertise across Facebook, Instagram, and its messaging apps. An AI agent is software that can answer customers and carry out assigned tasks. Mark Zuckerberg said Meta expects to be paid when those agents deliver a result for a business, according to the earnings-call report.

That points toward a fee tied to an outcome, such as a resolved support request or a completed booking, instead of a flat monthly software bill. Meta has not published the exact result it will charge for or what one result will cost. That missing definition is the whole buying question. A “result” is valuable only when it matches something your company would otherwise pay a person to complete.

Businesses already using WhatsApp or Messenger for inquiries should prepare first. Save the questions customers ask repeatedly, the answers staff approve, and the moments that require a person. Those records become the test set for any agent. Our comparison of AI chatbots for small-business customer service explains why refunds, complaints, and unusual requests still need a human route.

Ignore the broader talk of personal AI for now. The near-term operator issue is whether Meta can turn messages into dependable completed work without hiding errors behind an outcome fee.

Hint moves home-service discovery into an assistant

Hint launched its homeowner app on July 29 with an AI assistant that reads property records and uploaded documents, tracks maintenance, and suggests who to call when something needs attention. The official Hint site says the app can surface urgent repairs, judge whether a quote looks fair, and connect users with products and services.

This matters beyond one new app. A plumber, heating contractor, insurer, or landscaper may soon be evaluated by software that already knows the home’s age, past work, and maintenance schedule. The customer may arrive with a narrower request and an AI-generated view of a fair price. Discovery starts earlier.

Local service businesses do not need a Hint plan. They do need information an assistant can understand: exact services, service areas, current contact details, and plain answers to common repair questions. Keep those facts consistent on your website and major business listings. Our guide to getting named in AI recommendations covers the wider discovery shift, while the AI search topic page tracks the platforms changing it.

The caution is placement. Hint says it connects users with service providers, and its site says it does not sell user data. It does not explain publicly how providers are ranked. Until that is clear, treat this as a customer-behavior signal, not a new lead source.

Worth watching

• Microsoft said search and news advertising revenue excluding traffic costs rose 10% in the quarter ended June 30. The official results show Bing still deserves a measured budget test, especially where Google clicks are expensive.
• Encore AI raised $30 million for agents trained on customer calls and messages. The useful idea for smaller teams is the input: organized conversation history may become more valuable than another generic prompt library.
• OpenAI says it is building a family of devices for its chatbots. There is no product or buying decision yet, but assistant discovery may eventually move beyond screens.