The CFO AI Digest: January 14, 2026

Happy 2026, we’re back with the biggest stories in AI!

OpenAI and Anthropic launch healthcare updates, while Microsoft and Google push deeper into AI-powered retail. Anthropic publicly announces a new team to accelerate experimentation, and Slack adds agentic capabilities.

Here’s the rundown:

OpenAI

1. OpenAI launches ChatGPT Health, a secure AI tool built with physicians

OpenAI introduced ChatGPT Health, a dedicated tool for health and wellness queries. The service allows users to securely connect medical records and wellness apps such as Apple Health, MyFitnessPal, and Function to get personalized responses like interpretation of lab results. The tool is focused on data protection and adds safeguards such as purpose-built encryption, on top of ChatGPT’s existing security architecture, to keep Health conversations separate from other chats. Data shared in ChatGPT Health is also not used to train OpenAI’s foundation models. The product was developed with input from more than 260 physicians across 60 countries and is initially rolling out to a limited group of users with Free, Go, Plus, and Pro plans. The company will roll out broader access in the coming weeks.

CFO takeaway: ChatGPT Health formalizes demand that already exists — OpenAI reports that over 230 million people globally ask health and wellness related questions on ChatGPT weekly. OpenAI is now however anchoring that usage in consumer health data for added context through secure integrations. Its acquisition of Torch, a company that builds “medical memory for AI” also speaks to this goal. By collaborating with physicians, adding special data protections, and stating that ChatGPT Health seeks to support and not replace medical care, OpenAI is also building trust with consumers. This consumer-focused update counters Anthropic’s more enterprise-oriented approach with launches like Claude for Healthcare this week. The question is what platform becomes the default entry point for health queries — and whether that's decided by patient adoption or enterprise purchasing power.

Read more

Google

2. Google introduces new commerce protocol and agentic AI tools for retailers

Google rolled out new tools and infrastructure in its push toward AI-powered shopping. The company introduced the Universal Commerce Protocol (UCP), aimed at enabling AI agents to work together across the shopping journey from discovery to checkout. Co-developed with partners like Shopify, Etsy, Walmart, and Target, the UCP creates a shared language for agents, retailers, and payment systems, avoiding the need to build unique connections for each individual agent. The UCP will soon support AI-powered checkout directly in Search and the Gemini app, Google said in a blog post. Google also launched Business Agent, a chatbot that allows select brands to answer user questions in Search like a virtual sales associate, and Direct Offers, a Google Ads pilot for surfacing exclusive discounts to shoppers in AI Mode.

CFO takeaway: Google’s UCP is a bid for infrastructure control in agentic commerce. Unlike Microsoft’s recently-launched Copilot Checkout or OpenAI’s ChatGPT checkout, which centralizes transactions within platforms, Google is positioning UCP as an open standard that works across platforms. If UCP becomes the default language for AI shopping, Google gains influence over how commerce flows without being the merchant of record, shaping economics, data access, and bargaining power across the retail stack.

Read more

3. Google adds Gemini-powered features to Gmail

Google is rolling out new Gemini-powered features in Gmail to turn the inbox into what the company is calling a “personal, proactive assistant.” Features include AI Overviews, which summarize long threads and answer natural language queries by pulling key information from across the inbox. "Help Me Write" can draft or refine emails, while updated “Suggested Replies" generate context-aware responses in the user’s tone. A new Proofread tool checks grammar, tone, and style before sending emails. And an AI Inbox feature, now in testing, highlights important messages to help users prioritize across tasks.

CFO takeaway: As AI detects what is important in a sea of information and high-volume inboxes, a few questions arise. For enterprises, how do AI’s judgment calls on what is important get audited to ensure they’re working as expected and what controls do companies need to set up to influence prioritization? Longer term, this raises a second-order effect worth watching: how will writing change as email writers, and AI itself, learn to optimize not for human readers, but for what AI systems are most likely to surface?

Read more

Anthropic

4. Anthropic announces Labs for experimental product development

Anthropic announced Labs, an internal team focused on developing experimental products using Claude’s most advanced capabilities. In a blog post, the company noted that Instagram co-founder Mike Krieger will join Labs to build alongside Anthropic co-founder Ben Mann. Labs will extend Anthropic’s approach of experimentation, testing versions with early users and scaling what works into reliable products. Anthropic reported that this process has already led to launches like Claude Code and the Model Context Protocol (MCP), and that Labs is intended to accelerate this work as Claude continues to evolve.

CFO takeaway: Anthropic's public launch of Labs, along with its decision to bring on Krieger to help lead the team, is a strong signal of its commitment to upleveling product experimentation and development. Krieger said on LinkedIn that he's putting his "product founder hat back on," with the goal of building AI products to tackle some of the world's hardest problems. Labs formalizes Anthropic’s push to scale its application layer, raising the likelihood that the company increasingly competes with partners building on Claude.

Read more

5. Anthropic launches Claude for Healthcare and expands life sciences capabilities

Anthropic rolled out Claude for Healthcare and expanded Claude for Life Sciences to support clinical and scientific workflows. Claude for Healthcare includes connectors to medical policy databases, coding systems, and provider directories to help payers and providers save time finding information and generating reports. Use cases for healthcare startups and large enterprises include reviews of prior authorization requests, claims appeals, and tool development for clinical documentation and decision support. Claude for Healthcare also allows individuals to connect health records through HealthEx and Function to receive personalized insights. Claude for Life Sciences has added connectors to platforms like Medidata and ClinicalTrials.gov to support clinical trial planning and early-stage drug discovery.

CFO takeaway: Back in October, we reported that Anthropic’s GTM team flagged Claude Sonnet 4.5’s launch as important to healthcare and biotech. The company is now moving deeper to capitalize on that opportunity. Dave Nolan on Anthropic’s health team named orchestration the core bottleneck in U.S. healthcare, claiming that the country spends more on healthcare administration than many developed countries spend combined on healthcare. Addressing this bottleneck is a core priority of this launch, with Claude integrating fragmented workflows that burn billions annually in administrative overhead.

Read more

Microsoft

6. Microsoft debuts Copilot Checkout and Brand Agents for AI-powered retail

Microsoft introduced Copilot Checkout and Brand Agents, AI tools to help retailers convert high-intent shoppers. Copilot Checkout enables customers to complete purchases directly within AI conversations, while keeping the merchant as the seller of record. The feature is launching with partners like PayPal, Shopify, and Stripe. Brand Agents are AI assistants that live on a brand’s website, guiding shoppers through product discovery and purchase decisions in the brand’s voice. Microsoft said sessions assisted by Brand Agents have shown higher conversion and engagement rates compared to unassisted sessions.

CFO takeaway: Microsoft's Copilot Checkout enables the same capability as OpenAI's ChatGPT checkout and PayPal's Agentic Commerce Protocol: purchases completed directly inside AI conversations. The key difference is in positioning: Microsoft embeds commerce into brand-owned environments (your website, your checkout flow), while OpenAI centralizes transactions within its platform. As AI becomes a new commerce interface, the question becomes whether brands retain control over customer data, conversion, and pricing, or AI platforms increasingly intermediate the transaction and take a share of each transaction.

Read more

Salesforce

7. Salesforce updates Slackbot with Claude-powered AI

Salesforce released an upgraded version of Slackbot, its in-app assistant for the team messaging app, to its business and enterprise customers. Powered by Claude, the updated Slackbot can now surface information from Slack messages, files, and integrated apps like Salesforce and Google Drive, draft emails, and schedule meetings, according to TechCrunch. Salesforce CTO Parker Harris told CNBC that Slackbot is the most quickly adopted feature in the company’s history. He also noted that internal usage has replaced some reliance on external tools like ChatGPT.

CFO takeaway: In its launch post, Slack emphasized that “Slackbot understands how you and your team speak and how decisions actually get made.” That’s the enterprise selling point of context-aware AI, and exactly what platforms like ChatGPT have been trying to accomplish with company knowledge. Giving Slackbot agentic capabilities takes that capability a step further, from company-embedded knowledge sharing to acting on a user’s behalf in accordance with company standards.

Read more

In other news

1. Microsoft pledges to absorb utility cost increase for U.S. AI data centers (Microsoft)
2. Meta names banking exec Dina Powell McCormick president and vice chairman (Meta)

2 new AI tools to try

1. Claude Cowork — Waitlist here (Anthropic)
2. 11ai — sign up here (ElevenLabs)

Recommended reading

1. AI Global: Global Sector Trends on Generative AI (Similarweb)
2. Verbalized sampling: How to mitigate mode collapse and unlock LLM diversity (Jiayi Zhang, Simon Yu, Derek Chong, Anthony Sicilia, Michael R. Tomz, Christopher D. Manning, Weiyan Shi)