Get tomorrow's brief in your inbox
Today: Salesforce shipped Headless 360 to give small businesses reusable AI agent access across their entire platform without custom integrations. Google launched Gemini 3.8 Flash with stronger coding and agent performance but warned it might burn more tokens. A two-person event company used ChatGPT Work to cut merchandise inventory from 3 days to 3 hours and automate daily event listing reviews.
Salesforce Headless 360 is now generally available, and it could fundamentally change how small businesses use AI agents across their operations. The platform converts every Salesforce cloud service into reusable capabilities that any authorized AI agent can access securely. The new Model Context Protocol (MCP) lets AI agents dynamically discover and use your Salesforce data and workflows in real time, without building custom integrations for every new task.
What this means for small businesses: you can now deploy AI agents to handle customer inquiries, streamline marketing, and optimize sales while maintaining one consistent governance layer across everything. Engine, an early adopter, launched its AI support agent EVA in just 12 days, and it already resolves about 50% of customer chat interactions autonomously. That is a real, measurable reduction in support workload. The Slackbot MCP Client even brings Salesforce actions directly into Slack conversations, so employees stop toggling between platforms.
The catch: adopting this requires assessing your current infrastructure and training your team. You are giving AI agents access to core business systems, so governance and security planning are non-negotiable. But if you have been avoiding AI because you lack the resources to manage endless custom integrations, Headless 360 is built to solve that problem. Small Business Trends
Google Gemini 3.8 Flash is out, just weeks after 3.7 Flash. Google claims it "works harder" by performing more reasoning steps on complex tasks and calling tools iteratively. The introductory price is the same as 3.7 Flash ($0.75 per million input tokens, $3.75 per million output tokens), but Google warns it might use more tokens to maximize performance, especially at higher effort levels. Early benchmarks show it is the cheapest model at this intelligence level, but actual costs are up roughly 40% from 3.7 Flash due to a 30% increase in output tokens per task. If you are building AI agents or need strong coding performance without paying Opus 5 prices, this is worth testing. Gemini 3.7 Flash remains available if you want to minimize token usage. The Verge, Simon Willison
Amazon Alexa for Shopping can now verify whether an email, text, or phone call actually came from Amazon. You ask Alexa about the message, and it uses AI to compare it against every message Amazon has sent, while analyzing contents, formatting, and sender details. Alexa only confirms a message is real if it is completely certain. If it cannot verify, it directs you to check your orders in the app and contact Amazon support directly. This is a practical fraud defense for small business owners who receive phishing attempts disguised as Amazon messages. The Verge
Adobe acquired Rilo, an Indian marketing intelligence startup that built custom workflow automation for go-to-market teams, including competitor intelligence, content repurposing, sales call analysis, and custom workflows similar to Claude Cowork and ChatGPT Work. Adobe is integrating Rilo's IP and six-member team into its marketing suite. Rilo raised $1 million at a $10 million valuation and shut down post-acquisition. This is Adobe's second India acquisition after Rephrase.ai in 2023. If you use Adobe's marketing tools, expect workflow automation improvements in the next year. TechCrunch
ATV Big Air Tour, a two-person event company running 26 tour dates across the U.S., used ChatGPT Work to turn 3 days of merchandise inventory and reordering into 2-3 hours. Larissa Guetter took photos of merchandise, uploaded them to ChatGPT Work, and it created a spreadsheet, visual inventory website, and reorder recommendations in under 15 minutes. She reviewed, adjusted, and sent the final order to her supplier. The company also automated daily event listing reviews across 30+ publications, cutting fact-checking time from 8 hours to 1 hour per week. ChatGPT Work finds inconsistencies, suggests correction contacts, and drafts emails. This is a perfect example of small teams using AI to operate beyond their human capacity without hiring. OpenAI Blog
LinkedIn is the second-most cited URL by generative AI models, according to SEMrush and Meltwater research. 11% of pages cited by ChatGPT, Perplexity, and Google AI mode come from LinkedIn. 75% of LinkedIn citations are from individual pages, not company pages, and more than half come from profiles with under 10,000 followers. Long-form articles (500-2,000 words) and mid-length posts (50-299 words) get the most AI citations. A solopreneur case study writer ran a LinkedIn AEO experiment using HubSpot AEO to track AI visibility and found that optimizing LinkedIn content for answer engines is a viable strategy for competing against large agencies in B2B search. If you are a solopreneur or consultant, LinkedIn content may now convert better than a traditional blog strategy. HubSpot Marketing
Palo Alto Networks paid $500 million in cash and stock for Console, a two-year-old startup that uses AI agents to automate IT help desk tasks like password resets, app access, and troubleshooting. Console was valued at $157 million before the sale, delivering a rapid return for investors including Thrive Capital, DST Global, and Palo Alto Networks CEO Nikesh Arora (who participated as an angel). Palo Alto will integrate Console into Cortex to give security teams natural language investigation and resolution across the enterprise. This is Palo Alto's seventh acquisition in 2026. TechCrunch
HiddenLayer raised $100 million in a Series B led by Delta-v Capital as enterprises rush to secure their AI deployments. The Austin-based startup makes tools to protect AI models, agents, and workflows from adversarial attacks, vulnerabilities, and malicious code injections. ARR grew more than 10x over the past year, now in the "tens of millions" of dollars, with over 90% of growth driven by new customers. Financial services and large tech companies are the largest verticals, with contracts including the Department of Defense and intelligence community. Runtime security for AI is becoming a priority as AI deployments grow common across businesses. TechCrunch
Microsoft will start disclosing Azure quarterly revenue for the first time, providing a clearer picture of its cloud business competing with AWS and Google Cloud. Azure revenue grew 42% to $29.42 billion in the June quarter, representing almost 33% of Microsoft's total revenue. The company is consolidating from three operating segments to two: Agents and Infra (Azure, M365, productivity/server licensing, industry solutions) and Devices and Consumer (search/ads, Xbox, Windows). Microsoft has over 30 million paid seats for Microsoft 365 Copilot, up from 20 million in April. CNBC
Use ChatGPT Work to automate inventory and ordering. Take photos of your inventory, upload them, and let ChatGPT create a spreadsheet, visual inventory, and reorder recommendations. Review and adjust before sending to your supplier. This is especially useful for merchandise, retail stock, or parts inventory.
Optimize LinkedIn content for answer engines. Write long-form articles (500-2,000 words) or mid-length posts (50-299 words) focused on educational, advice-driven content. Use HubSpot AEO to track which prompts surface your content in ChatGPT, Perplexity, and Google AI mode. This is a viable strategy for solopreneurs competing against large agencies.
AI security spending is exploding (Gartner estimates $2.83 billion this year, up 83% from 2025), driven by enterprises deploying agents at scale and needing runtime protection. Small businesses should prioritize tools with built-in governance and security rather than bolting on security later. Salesforce Headless 360 and Google Gemini 3.8 Flash both reflect a shift toward making AI agents more accessible and affordable for smaller teams, but token costs and infrastructure complexity remain real barriers. Watch for more acquisitions in the marketing automation and AI workflow space as Adobe, Canva, and others race to build end-to-end AI-powered marketing rails.