8DIGITS news

News & Opinions.

Timely takes on technology and systems that matter for small and mid-sized businesses.

Google's new Gemini agent is in early access with SMB customers and is designed to act as a persistent coworker across work apps. Broader availability is expected, but there is no reason to rebuild existing workflows before access and pricing are confirmed.

Agents · Brief · 4 min · SoloWatch

Google opens early-access Gemini agent that works across Workspace and SMB tools

Join early access if offered; do not rebuild workflows yet.

Read →

Agents · Brief · 4 min · SoloTest

HeyGen Voice tops independent TTS tests; free instant clones land this week

Clone your voice for free in seconds if you already make video content.

Read →

The OpenAI Decisions API replaces a free-form classification reply with a typed probability, choice or score. Powered by GPT-6 Luna, it can simplify lead scoring, ticket routing and agent decision steps for teams already using the API.

Agents · Brief · 4 min · SoloTest

OpenAI Decisions API enters public beta for typed yes/no and routing calls

Swap one classification or routing step to Decisions this week if you already call the API.

Read →

Google's SynthID Detector is now public at synthid.com, in English, worldwide. It flags watermarks from Google, OpenAI, Nvidia, and Kakao, with Apple expected later. Use it before a client asset or testimonial goes live. A miss does not mean the file is real.

Measurement · Brief · 3 min · AgencyAdopt

Check client images for a SynthID watermark before you publish them

Run client UGC and testimonials through synthid.com before they go live. A clean result is not proof.

Read →

Anthropic put live dashboards and short animated explainers into Claude on 8 October. Dashboards are beta on paid plans and need a data-platform or Salesforce connection. Motion is beta on Team and Enterprise, written as editable code and exported as MP4. Docs, Slides, and Design are now on every plan.

Agents · Brief · 4 min · AgencyTest

Claude Motion is a code animation, not a video model. Test it on one explainer

If you are on Team or Enterprise, time one explainer in Motion. Do not buy a plan for it.

Read →

ChatGPT is rolling out GPT-6 Sol to paid plans and GPT-6 Luna to Free and Go, with Intelligent UI that can return charts, buttons, and forms instead of plain text. Everyday GPT-6 is no longer a paid-only surface. One client job this week is enough to see if the new reply format saves a round of editing.

Agents · Brief · 3 min · SoloTest

ChatGPT now answers with charts, buttons, and forms, and free accounts get GPT-6 Luna

Try Intelligent UI on one client job this week. Free accounts get GPT-6 Luna from 8 October.

Read →

Claude Haiku 5.5 is generally available with effort control, a 1 million token window, and list prices of $0.10 input and $0.50 output per million tokens on prompts up to 100K. Anthropic says typical runs cost about 75% less than Haiku 4.5. Teams already on Claude can shift classification, extraction, and subagent traffic this week and leave the expensive model for the step that needs it.

Development · Brief · 4 min · SoloAdopt

Claude Haiku 5.5 puts effort control and a 1M window on a $0.10 routing model

Move high-volume routing and extraction to Haiku 5.5. Keep Sonnet for the hard steps.

Read →

Mistral Large 4 is live as a preview API, with open weights planned for the end of October after red-team work with security partners and authorities. It is a real European frontier-scale candidate, multimodal, and aimed at coding, cyber, finance, and law. Independent scores are not the full picture yet, so the sane move is a preview trial, not a production cutover.

Development · Brief · 3 min · SoloWatch

Mistral Large 4 is in preview, and the open weights are the part worth waiting for

Try the preview API on one European workload. Do not migrate until the weights ship.

Read →

Google put Nano Banana 2.1 into Gemini, Search AI Mode, Google Ads, and the API on 6 October. Standard API image output is about half the old price at 1K and 2K. Input and thinking tokens went up, so a volume batch is the only fair test.

Agents · Brief · 3 min · AdsTest

Google cut Nano Banana image prices in half. Rerun one ad batch before you lock the month.

Rerun one paid image batch on Nano Banana 2.1. Do not switch the whole creative stack.

Read →

OpenAI will test a visual ad format during ChatGPT image generation for Free and Go users in the US later this month. Ads are labeled, sit separately from the generated image, and do not change ChatGPT's answers. Early partner numbers are not a media plan, so check your own measurement stack before you buy.

Measurement · Brief · 4 min · AdsTest

ChatGPT will test product ads next to image generation

Test ChatGPT image ads only if you can measure incrementality. Do not treat partner CPA as your forecast.

Read →

Zapier Agents no longer sits outside the Zap. On 6 October 2026, the same prompt, tools, and permission to act became an AI by Zapier step you can drop into a normal workflow. Approval is set per tool, you choose the model, and every run appears in Zap History.

Agents · Brief · 4 min · AgencyAdopt

Zapier Agents now runs as a step inside the Zap

If you already pay for Zapier, move Agents into an AI step and approve only the risky tools.

Read →

On 5 October 2026 TikTok unveiled Buy Direct, a one-click in-feed checkout, and Shopping Assistant, a conversational agent that answers size, shipping, stock, and purchase questions inside the app. Both sit outside TikTok Shop and are starting with US users and brands. If you sell on Shopify and buy TikTok traffic, bad product data now fails the agent, not just the ad.

Agents · Brief · 4 min · ShopifyTest

TikTok can close the sale in the feed. Messy product data gets skipped.

Test one SKU in Buy Direct. Fix stock, size, and returns before you raise spend.

Read →

Book a 20-minute diagnosis