Camp Tech: 1)CTV features tips to keep your devices safe while you travel this summer. 2)What is agentic AI? 3) We’re Only Starting to Grasp the Pitfalls of Using A.I. at Work 

1) Travelling abroad? These tech tips will keep your devices safe this summer

ByElianna Lev, Updated: June 21, 2026
Courtesy of Sci-Tech News

Tech expert Avery Swartz has tips on how to keep your tech safe before you travel abroad this summer.

Summertime travel can be stressful, but there are ways to minimize any potential woes, particularly when it comes to tech. Tech expert Avery Swartz spoke with CTV Your Morning on Friday to share how eSim cards, downloading content and tracking your devices can help make your travel adventure a lot less stressful.

Invest in an eSim card
While Canadian phone carriers provide international roaming rates for people travelling abroad, they can charge up to $20 a day, which quickly adds up. Swartz recommends setting up an eSim card, which is similar to a SIM card, except it’s downloaded and installed onto your device before you leave.

“When you get to your destination, you just turn your phone on, and it works,” she said, adding that some plans go for as low as $40 for an entire week.

Taking photos of all your important documents, like passports, flight information and hotel bookings can help you if you’re in a pinch.

Photograph important documents
Swartz strongly recommends taking photos of all your important documents, like passports, flight information and hotel bookings. While it helps to keep those photos on your phone, she also suggests to keep it somewhere else you can access off your phone, as well as sharing the photos with someone.
She also recommends securing the photos somewhere like DropBox or WhatsApp, which has end to end encryption.

“If you’re at the consulate, it’s the middle of the night, you don’t have your passport, at least you could bring up a picture of it,” she said.

Track your phone
It’s not unheard of to lose or have your phone stolen when travelling abroad. Swartz strongly suggests setting up Find My app on Apple phones, which tracks the location of the device.
If the phone has been stolen, or dropped into the ocean, logging into iCloud, or Android.com, will lock it remotely.

Download everything
If you want to keep entertained during your travels, or want to know exactly where you are, Swartz recommends downloading shows, podcasts and maps onto your devices, before you leave.
“Even in Google Maps, type in any destination, and then there’s a little button that says, download offline Some airlines won’t allow passengers to pack their portable charging blocks in their overhead storage.

Don’t pack batteries in checked luggage
Swartz stressed that if you are travelling with lithium batteries, which include power bank portable chargers, be sure to pack it in your carry-on luggage.
“A lot of airlines also are not even allowing you to put them in the overhead bins,” she said. “You have to ‘have them at your seat.”

Swartz added that airlines might not allow you to use a power bank on the flight, so be sure to charge your device beforehand.

2) Important note: Agentic AI is available in all kinds of tools, not just Zapier’s proprietary stuff.

NOTE: Read for the explanation, not for the product plug.

What is agentic AI? And how you can start using it By Jessica Lau · June 29, 2026 The idea of AI tools that can be trusted to operate independently has always been exciting—and, for a long time, just out of reach. That’s changing: 84% of enterprise leaders now say they’re likely or certain to increase AI agent investments over the next 12 months. Not everything marketed as agentic AI actually clears the bar, though. Here, we’ll explore what agentic AI really is, how it works, and some real-world examples of agentic AI workflows you can start experimenting with today. Table of contents: What is agentic AI? How agentic AI works Examples of agentic AI workflows Challenges of agentic AI Use Zapier to get started with agentic AI

What is agentic AI?
Agentic AI is an AI system that can accomplish a goal autonomously: taking actions, making decisions, and course-correcting along the way without the need for human intervention.
These goals may be narrow in focus—like automatically replying to customer emails—or wider, requiring multiple, coordinated steps—like researching a prospect, drafting a personalized outreach email, and logging the interaction in your CRM without you touching a thing. 

Agentic AI vs. AI agents
The terms are related but not quite the same. Here are the key differences.  
An AI agent is a single autonomous “worker” that can perceive its environment, make decisions, and take action toward a goal.  
Agentic AI is the bigger picture: the overall system—memory, tools, decision-making logic, and feedback loops—that lets the whole thing run without constant human input. In practice, though, the line between the two is blurring. As the technology has gone mainstream, people tend to use the terms interchangeably, so you’ll frequently see “AI agent” used to mean the whole system and vice versa.

How agentic AI works
Agentic AI systems tackle problems in a four-step process: perceiving, reasoning, acting, and learning.
Step 1: Perceive
Before an agentic AI system can do anything, it needs to understand what’s going on. It takes in information from its environment via emails, documents, APIs, databases, user-entered AI prompts, or whatever sources are relevant to the task at hand. Think of it as the system getting its bearings: it reads the room before deciding what to do next.
Step 2: Reason
Once the AI system has a picture of the situation, it needs to figure out what to do with it. This is where an AI model, typically an LLM, looks at the goal, takes stock of the tools and information available to it, and comes up with a plan. That might mean pulling in additional context using processes like retrieval-augmented generation (RAG), or breaking a big goal into a sequence of smaller steps it can tackle one at a time.
For enterprises, this is often where businesses bake in logic and guardrails, like routing issues differently for VIP customers, enforcing refund policies, or flagging certain decisions for human review.
Step 3: Act
With a plan in hand, the system gets to work. It takes action using the tools available to it—either directly through integrations or through protocols like Model Context Protocol (MCP). Each action changes the environment, which the system can then perceive again, closing the loop.
Step 4: Learn
This is what separates agentic AI from a one-off script. Rather than running the same fixed sequence every time, the system adjusts based on what worked and what didn’t. If a strategy succeeds, it leans into it. 
In practice, learning isn’t always a clean, discrete step. Instead, it happens continuously across the whole perceive–reason–act loop, as the system accumulates experience and refines its approach. Over time, especially with explicit feedback loops built in, the system gets better at reaching its goals.

Examples of agentic AI workflows
To show you what agentic AI workflows look like in the real world, here are a few examples. Some are hypothetical (but completely possible), and some are based on actual workflows created with an agentic AI system on Zapier.

Agentic AI workflows in support ticket triage
ClickUp‘s support team was handling about 5,000 tickets a month, each requiring 15 minutes of manual research before a rep could type a reply. Here’s how they used Zapier to connect their support stack via Zapier MCP and build an agentic AI system that handles that upfront work automatically.
Perceive: When a ticket lands in Zendesk, Zapier automatically pulls the full ticket, including customer history, issue description, and any relevant metadata, and cross-references it against ClickUp’s internal knowledge base and past tickets.
Reason: AI by Zapier interprets the unstructured ticket data, classifies the issue, and maps it to relevant documentation, runbooks, and a recommended response path.
Act: Before the rep has typed a single word, the AI delivers a structured summary: relevant docs surfaced, history pulled, and next steps laid out. 
Learn: The agentic AI system tracks which recommended response paths lead to successful resolutions and which get overridden by reps, and uses that signal to improve its classifications over time.

Agentic AI workflows for bug fixing
Now let’s consider something a little different: an AI coding agent that’s tasked to fix bugs.
Perceive: The agentic AI has access to the codebase, the server logs, and the bug report database. If the server logs show an error or it receives a bug report from a customer or internal user, it takes action. 
Reason: The agentic system uses an LLM to consider the bug report, searches through the codebase to find the problem, and comes up with a solution. It may also need to pull in information from internal databases, external help documents, and even ask a software developer for more information.
Act: Once the agentic system has identified a potential fix, it tests it in a dedicated local environment. If it works, it submits a merge request to GitHub for a software developer to review. If it fails, it considers the error message and tries to create a new fix.
Learn: The agent sees whether its merge requests are approved or rejected and also learns from the error messages from testing its own code fixes.

Agentic AI workflows for automatic sales follow-ups
Say you want to take the manual work out of post-sales call follow-ups. Here’s what that workflow could look like. 
Perceive: The agentic AI receives call recordings from the sales team and transcribes them.
Reason: The agentic system uses an LLM to scan the transcript and pull out the prospect’s details and any action items. 
Act: Once the agentic system has identified the action items, it generates a draft email attaching any documents, prospectuses, and files necessary. It also logs all the relevant details in the CRM.
Learn: The agent sees whether the deal closes in the CRM and also learns whether the email it drafted gets sent or not.

Agentic AI workflows for assessing potential hires
If you’re a recruiter drowning in job applications, an agentic AI system can handle the initial screening for you. Here’s what that looks like in practice.
Perceive: The agentic AI receives job applications. 
Reason: The agentic system uses an LLM to parse the job application and see if they meet the job criteria. 
Act: If the AI determines the potential applicant meets the job criteria, it connects to a professional registry to check licensing details, pulls in a job history from LinkedIn, searches Google for any relevant personal or professional details, and creates a summary of its findings that it emails to a recruiter—along with the attached resume. It also calculates a hireability score.
Learn: The agent learns which candidates are hired or not and uses its hireability assessments as a measure of how well it assessed each candidate. If it scores candidates incorrectly, it updates its criteria.

Challenges of agentic AI
For all the excitement, agentic AI also adds new challenges around control, safety, and operations that companies are still figuring out.
Oversight and control: Agents that get more and more autonomous might also be harder to understand when it comes to their decision-making. It could also get tough to guarantee they’ll always stay within policy. This is why explicit approval steps are a crucial part of any agentic workflow. Zapier’s Human in the Loop (HITL) tool makes it easy to build in these checkpoints so you can review, intervene, or course-correct at key moments before the AI moves on.
Security and data privacy: Agentic systems often touch sensitive customer records, financial data, or proprietary code. Zapier found that 18% of enterprise leaders say security and data privacy limit their AI agent adoption, which makes enterprise-grade controls like role-based access and audit logs non-negotiable.
Complexity and maintenance: Coordinating multiple agents across apps and APIs makes systems harder to debug and maintain. A small change to one tool or prompt can have ripple effects downstream. Many teams are responding by starting with narrow, deterministic workflows that include AI steps before gradually layering in more autonomy. Zapier is built for exactly this approach. With AI by Zapier, you can add AI-powered steps into existing workflows without committing to full autonomy from the start.
Unpredictable outputs: The more autonomous an agentic system gets, the harder it is to guarantee it’ll behave consistently, especially when it encounters edge cases or ambiguous inputs it wasn’t designed for. Left unchecked, a single bad output can cascade into bigger problems downstream. AI Guardrails by Zapier lets you define boundaries for what your agents can and can’t do, so the system stays within the parameters you set even as it operates independently.

Use Zapier to get started with agentic AI
The perceive–reason–act–learn agentic AI loop is only as reliable as the infrastructure underneath it. When an agentic system is touching your CRM or customer data, you need to trust that it can do it safely.
Zapier gives your agentic AI systems governed access to the apps your business already runs on, with OAuth-managed connections, scoped permissions, and a single place to control what your agents can and can’t touch. The agent sees results, not credentials. And if you need to revoke access, you do it once and it’s done everywhere.
You can build incrementally: start with a narrow workflow, add AI steps where they make sense, and layer in more autonomy as your team’s confidence grows. When you’re ready to go further, Zapier MCP lets you connect your AI assistant directly to 9,000+ apps. This means you can ask Claude or ChatGPT to update a deal in your CRM, draft and send a follow-up, or kick off an onboarding sequence, without leaving your chat window.
Try Zapier
Related reading:
AI in the workplace: What it looks like now and where we’re headed
AI integration: How to bring AI into your workflows
Which AI models can you automate on Zapier?
The best AI agents for enterprises
This article was originally published in June 2025 with contributions from Harry Guinness and Mike Floeck. The most recent update was in June 2026.

3) We’re Only Starting to Grasp the Pitfalls of Using A.I. at Work

This article in The New York Times is long, but it’s well worth your time to read it. Scholars say the “unknown unknowns” of using artificial intelligence in the workplace may be undermining the technology’s advertised benefits, in ways you probably didn’t consider.

We’re Only Starting to Grasp the Pitfalls of Using A.I. at Work

Scholars say the “unknown unknowns” of using artificial intelligence in the workplace may be undermining the technology’s advertised benefits.

By Noam Scheiber, June 29, 2026

Over the past year or two, companies have started using so-called artificial intelligence agents as bona fide “employees,” even including them in their organizational charts.

Emma Wiles, a Boston University professor who studies how A.I. affects workers, stumbled onto this phenomenon in October, at a conference where two human resources executives said that treating A.I. agents like real employees was a way to increase productivity and to put their companies on the cutting edge.

But when Dr. Wiles and three collaborators from Boston Consulting Group investigated further, they discovered a pitfall. In an experiment involving dozens of companies with A.I. employees, the researchers found that managers tended to vet documents less carefully when told an A.I. employee had produced them. The managers missed errors that other managers caught when told they were vetting the work of a human.

Dr. Wiles speculated that managers didn’t think sussing out mistakes made by A.I. employees was their responsibility. If something went wrong, they could dismiss it as the fault of the tech team, or of the executives who wanted A.I. employees in the first place. “But it’s not your problem,” she said, channeling the managers’ mind-set about their own roles.

In the years since A.I. burst onto the scene, many companies have become aware of flaws produced by the technology and, at times, taken steps to offset them. They know that A.I. models can be biased against certain groups of people, like nonwhites. They know that chatbots can provide confident but incorrect answers to queries. They know that the bots sometimes spill the beans on information that should remain private.

But as companies race to bring A.I. into their day-to-day operations, researchers are discovering more subtle defects. In principle, these flaws could be corrected, too. For example, companies could hold managers directly responsible for the mistakes of A.I. subordinates.

But in practice, most corporate users appear to be blissfully unaware of these issues, raising the possibility that A.I.’s promise of increased productivity and vast cost savings could be undermined.