Practical AI for Business: Epic in Healthcare, Meta Buys Manus, MIT.nano Drug Discovery, US AI Regulation Updates, and Key AI Court Rulings
Happy New Year!
This week, let’s keep it simple. We’re moving past the flashy demos and talking about AI that’s actually “useful on a Wednesday.”
Here are five stories showing where real business value is showing up, plus what the latest rules and court updates mean for teams that are actually deploying this stuff.
Hospitals are pressure-testing AI where it counts
Hospitals are using AI for the unglamorous stuff that actually moves the needle: documentation, claims, patient messages, and speeding up routine workflows. One big example? Epic’s AI is helping draft insurance appeals, which can cut admin time and improve the odds of reversing denied claims.
My take: Healthcare is teaching everyone the playbook on where AI actually works. It’s not about flashy features; it’s about attacking the administrative bottlenecks that drain your team’s energy. If you run a service-based operation, look at your intake, summaries, and follow-ups. That’s your low-hanging fruit. Start there, measure the time saved, and you’ll see immediate ROI without reinventing the wheel.
MIT.nano spotlights AI for drug discovery
MIT researchers are highlighting how generative AI can help design new molecules aimed at difficult diseases. It’s a massive step toward faster drug discovery and more targeted therapies. For businesses in biotech, health, and research services, this is another sign that AI can reduce early-stage R&D cycles.
My take: This is one of the most practical “big impact later” areas of AI. If you support life sciences clients, think about where AI can compress timelines: candidate generation, screening, documentation, and regulatory prep. The companies that build repeatable AI workflows around those steps are going to move faster with the same headcount.
Meta buys Manus as the AI agent race heats up
Meta’s acquisition of Manus (reportedly for over $2 billion) is a clear signal that the next wave is not just chat it’s agents. We are talking about AI that can execute tasks like research, coding, and analysis. Manus will keep selling subscriptions while Meta brings agent capability deeper into its ecosystem.
My take: This is your reminder to start designing “agent-ready” workflows now. Pick one repeatable process you would happily hand to a reliable junior assistant (research brief, first draft, QA checklist, weekly report), then build a version of it with clear inputs, approvals, and auditability. If you want help with how to build an agent, have a look at our YouTube channel where we have some videos on building agents.
New US AI rules are here
A big batch of state-level tech laws is taking effect in 2026, including AI transparency requirements, chatbot disclosures, and rules around higher-risk AI systems. Even if you don’t think of yourself as “an AI company,” these rules can touch you if you deploy AI features or automate decisions that impact customers.
My take: Compliance is quickly becoming a competitive advantage. If your product or business uses AI, tighten the basics now: clear disclosures, documented model purpose, and a simple internal policy for how your team uses AI with customer data. It’s way easier to do this before a complaint, not after. Here’s our sample AI Acceptable Use Policy
Courts are clarifying “access control” in scraping disputes
A recent court development in the Ziff Davis vs. OpenAI case is being closely watched, especially around the argument that ignoring robots.txt could trigger certain anti-circumvention claims. For publishers, marketers, and AI product teams, the practical takeaway is that courts are scrutinizing what truly counts as a technical barrier versus a polite request.
My take: If your business uses web data or builds AI features that rely on scraping, this is a “get your house in order” moment. Make sure your data sourcing is documented, your vendor contracts include clear rights and indemnities, and your engineering team has a playbook for respecting site policies and handling takedown requests.
Tools to try this week (AI Smart recommended)
- Talkadot -Tool for capturing real-time feedback and testimonials. Turn engaged audiences into future clients.
- Jasper –Team-friendly marketing writing tool that helps maintain brand voice and security.
- Gamma– Create presentations, documents, and simple webpages quickly with AI.
What is Fyxer AI?
It’s my must have AI email assistant! Fyxer has transformed my relationship to my inbox!
Fyxer is an “AI executive assistant” that works inside your existing tools. Its two headline jobs are:
- Email assistant: automatically sorts your inbox and creates ready-to-send draft replies in your voice.
- Meeting notetaker: joins calls (Zoom, Google Meet, Teams), then records, transcribes, summarizes, and can generate follow-up notes you can share.
It also supports calendar workflows, team scheduling, and (on higher tiers) CRM-related automation like HubSpot integration.
How to use Fyxer: setup in minutes
1) Create your account, start the trial
Fyxer offers a 7-day free trial on its pricing page.
2) Connect your inbox and calendar
- Connect Gmail or Outlook (and your calendar) so Fyxer can categorize emails and draft replies directly where you already work.
3) (Optional) Connect your meeting tools and CRM
- Link Zoom / Google Meet / Teams so the notetaker can join meetings.
- If you use HubSpot, connect it if your plan includes CRM features.
Your daily “clean inbox” workflow (10 minutes)
Step A: Work from “To Respond” first
Once connected, Fyxer will label and prioritize emails so you can start with what actually needs action.
A simple routine:
- Open To Respond
- Clear the obvious quick wins
- Skim FYI (or your “read only” bucket)
- Leave newsletters and promos for last (or hide them entirely)
Step B: Use the auto-drafts (the big time saver)
For emails that need a reply, Fyxer generates drafts so you mostly do:
- Pick the best draft option
- Edit 1–2 lines for context
- Send
This is the core “I got time back” moment because you are reviewing instead of starting from a blank page.
Step C: Re-label when Fyxer gets it wrong (train your workflow)
If an email is mislabeled (example: a “Marketing” email is actually urgent), re-label it so your inbox stays reliable over time.
How to use the meeting notetaker (and look wildly organized)
1) Let it join the meeting
Fyxer can join meetings to take notes, then produce a summary plus transcript/recording depending on settings and platform.
Best practice: tell participants a notetaker is present before recording/transcribing.
3) After the call, grab:
- Key points and decisions
- Action items
- Transcript for search
- A follow-up message you can send to attendees
5 settings most people overlook (but you should turn on today)
- Hide low-value categories from your main inbox
If newsletters and meeting updates clutter your view, remove them from the main inbox view so you only see what matters. - VIP email rules
Force messages from your boss, top clients, or key leads to always land in To Respond. - Enable drafts to appear in your inbox
Make sure the integration is configured so drafts show up where you actually reply (Gmail/Outlook). - Signature + tone consistency
Add your signature and set your style so drafts match how you naturally write. - Follow-up reminders (“Awaiting reply”)
Turn on reminders so emails you sent do not disappear into the void. This is how you stay on top of threads without mental load.
The last trick (the “did you hire an assistant?” move)
Turn your meeting notes into an instant post-meeting recap email: paste the action items, confirm owners and due dates, and send it within 5 minutes after the call. Fyxer is designed to support this kind of follow-up workflow with meeting summaries and drafting.
Join the Conversation
Have questions, comments, or cool AI stories to share? We’re all ears and always looking to highlight insights from the AI community. Swing by and drop us a line at info@aismartventures.com.
Can’t wait to hear from you!
Cheers,
-Nicole Donnelly
Founder, AI Smart Ventures
