I use Perplexity almost every day. I reach for it when I need to think through a problem quickly, cross-reference a claim, or pull together research from across the web without trawling through a dozen tabs.
It’s one of the few AI tools I keep open all day, and it’s earned that spot because it’s genuinely fast, well-cited, and easy to have a real back-and-forth with.
Where it tends to shine is in that browser-native, open-ended research mode. You ask a layered question, you get a structured answer with sources you can actually verify.
For competitive landscape work, market scanning, or just pressure-testing an idea before you take it to the leadership team, it’s hard to beat.
So when Perplexity announced that its “Computer” agent is now a native add-in inside Microsoft Word, Excel, PowerPoint and Outlook, my first reaction was genuine interest. That research quality, sitting right alongside your documents and data, is a compelling idea.
But like any tool moving into your core work environment, it’s worth understanding how it works before you roll it out across the business.
What Perplexity Computer Actually Does Inside Microsoft 365
The add-in runs as a side panel inside Microsoft 365 apps, so users can draft documents, model data, build presentations and handle email with Perplexity working alongside them, without switching to a browser.
It connects to SharePoint and more than 400 other data sources, supports more than 20 AI models in a single interface, and includes enterprise controls like SAML SSO, audit logs and granular admin settings.
That’s a meaningful upgrade from copy-pasting between a browser tab and a Word document.
The kind of structured, source-backed thinking Perplexity does well can now happen in the same window as your actual work, and for knowledge workers who already rely on it, that’s a genuine productivity step forward.
The question for business leaders isn’t whether the tool is good. It’s whether your environment is set up to use it well, and that starts with understanding where your data goes when the add-in is active.
A Useful Launch and a Timely Prompt
The same week Perplexity Computer arrived in Microsoft 365, CNN filed a copyright lawsuit against Perplexity AI, alleging the company scraped and reused tens of thousands of its news articles, images and videos without authorisation.
CNN claims Perplexity marketed premium access to its content despite having no licensing agreement in place. I’m not sharing that to cast doubt on the product. As I said, I use Perplexity regularly and rate it highly.
The point is that no vendor is above scrutiny, and the timing is a useful prompt to think clearly about what due diligence looks like when a third-party AI tool moves inside your core work environment.
When a tool sits in the browser, the exposure is one thing. When it sits inside Word and Outlook, it’s another. Your documents, your prompts, your client communications are now in the same session as a third-party service running its own infrastructure, its own models and its own data agreements.
Getting clear on that boundary isn’t about being cautious for caution’s sake. It’s about using the tool confidently rather than discovering the gaps later.
Understanding the Security Boundary
Perplexity has positioned Computer as enterprise-ready, and the security credentials it offers are the right ones for a business tool. SAML SSO, audit logs and admin controls are meaningful features, and they show the product is being built with IT and security teams in mind.
But enterprise-grade features on a third-party add-in don’t automatically mean it sits within your existing Microsoft 365 security boundary. When your team uses Microsoft 365 Copilot, data stays within the security and compliance controls your tenant already has in place.
Microsoft operates as the data processor, and businesses with data residency requirements can be confident their content stays in-country.
A third-party add-in works differently. Prompts and responses may travel to that vendor’s own infrastructure, which means your data protection, retention and compliance settings in Microsoft 365 may not automatically extend to those interactions.
That’s not a reason to avoid the tool. It is a reason to ask a few specific questions before your teams start using it at scale:
Where does business data go when a user prompts Computer from inside Word or Outlook?
What data retention and deletion policies apply to those interactions?
Does the vendor’s security posture meet your industry or regulatory requirements?
How does this sit alongside your existing Microsoft 365 data and acceptable-use policies?
If you can answer those clearly, you’re in a strong position to make an informed decision. If the answers aren’t there yet, that’s the gap to close before rollout.
Multi-Tool AI Is Now the Default, Not the Exception
Perplexity Computer is one example of a broader shift that’s already well underway.
The Microsoft 365 add-in ecosystem is growing quickly, and AI tools are a meaningful part of that growth. Most knowledge workers are already using several AI tools, whether or not IT has formally approved each one.
Microsoft has acknowledged this reality in its own product strategy. The Microsoft 365 E7 suite was designed partly because businesses couldn’t keep track of what AI was doing across their environment without a single governance and control layer.
As we covered in our piece on Microsoft 365 E7 and Agent 365, the agent registry surfaced tens of millions of agents in just two months of preview, which tells you something about the pace of AI proliferation inside business environments.
The answer isn’t to block tools like Perplexity Computer. It’s to have a clear picture of what’s running in your environment, what data each tool can access, and what controls are in place if something goes wrong.
Getting that picture right is what separates confident AI adoption from reactive clean-up.
Practical Steps to Get the Most From It
Getting ahead of this doesn’t require a large project. It starts with a few practical steps that most Microsoft 365 environments can move on today.
Know what’s already running. Start with the tools your teams are actively using across Office, browser extensions and integrations. If Perplexity Computer or similar add-ins are already in play, include those in your inventory. You can’t govern what you can’t see.
Check your data classification and permissions. Microsoft Purview gives you the tools to classify data by sensitivity and set policies around what AI tools can access. If that work hasn’t been done yet, this is the right moment to start.
A tool can only do what its permissions allow, and tightening that layer reduces exposure regardless of which AI is in the session. We covered the governance case for this in detail in our blog on unified AI governance for Copilot and beyond.
Set a clear acceptable-use position. Your team will take their cue from you. A short, plain-language policy that tells staff which AI tools are approved, what data they can use those tools with, and when to check with IT before adding a new integration goes a long way. It doesn’t need to be long. It needs to be clear.
Review vendor agreements before scale-up. If a vendor is going to sit inside your email and document workflows, their data processing agreement should be reviewed against your obligations under the Australian Privacy Principles, particularly in industries like healthcare, financial services and legal where client data carries extra protection requirements.
Use your Microsoft 365 admin controls. The admin centre gives you meaningful control over which add-ins are permitted in your environment. If you want to allow Perplexity Computer for a specific team or role while you assess it, that’s possible.
Broad access before a proper review isn’t the right approach in most business environments, but a controlled pilot absolutely is.
The Opportunity Is Real, and It’s Worth Pursuing Properly
I genuinely think tools like Perplexity Computer will improve the way knowledge workers operate inside Microsoft 365. The productivity case for in-app AI that can research, draft and model alongside your actual documents is strong, and it’s only going to grow as more vendors follow the same path.
As we discussed in our piece on Claude AI inside Microsoft 365, the model-diverse future of enterprise AI is already here. Business leaders aren’t choosing between one AI and another.
They’re managing several AI tools working across the same environment, and the businesses that get their governance in order now will be the ones using these tools with confidence rather than catching up later.
The conversation I’m having with most Australian business leaders right now isn’t “should we use AI?” It’s “how do we know the AI we’re using is working safely?”
For Microsoft 365 environments, the answer lives in your governance settings, your data classification, your vendor review process and your people’s understanding of where the boundaries are.
If you’d like to work through what that looks like for your business, our team at CG TECH is happy to help.
About the Author
Carlos Garcia is the Founder and Managing Director of CG TECH, where he leads enterprise digital transformation projects across Australia.
With deep experience in business process automation, Microsoft 365, and AI-powered workplace solutions, Carlos has helped businesses in government, healthcare, and enterprise sectors streamline workflows and improve efficiency.
He holds Microsoft certifications in Power Platform and Azure and regularly shares practical guidance on Copilot readiness, data strategy, and AI adoption.
I use Perplexity almost every day. I reach for it when I need to think through a problem quickly, cross-reference a claim, or pull together research from across the web without trawling through a dozen tabs.
It’s one of the few AI tools I keep open all day, and it’s earned that spot because it’s genuinely fast, well-cited, and easy to have a real back-and-forth with.
Where it tends to shine is in that browser-native, open-ended research mode. You ask a layered question, you get a structured answer with sources you can actually verify.
For competitive landscape work, market scanning, or just pressure-testing an idea before you take it to the leadership team, it’s hard to beat.
So when Perplexity announced that its “Computer” agent is now a native add-in inside Microsoft Word, Excel, PowerPoint and Outlook, my first reaction was genuine interest. That research quality, sitting right alongside your documents and data, is a compelling idea.
But like any tool moving into your core work environment, it’s worth understanding how it works before you roll it out across the business.
What Perplexity Computer Actually Does Inside Microsoft 365
The add-in runs as a side panel inside Microsoft 365 apps, so users can draft documents, model data, build presentations and handle email with Perplexity working alongside them, without switching to a browser.
It connects to SharePoint and more than 400 other data sources, supports more than 20 AI models in a single interface, and includes enterprise controls like SAML SSO, audit logs and granular admin settings.
That’s a meaningful upgrade from copy-pasting between a browser tab and a Word document.
The kind of structured, source-backed thinking Perplexity does well can now happen in the same window as your actual work, and for knowledge workers who already rely on it, that’s a genuine productivity step forward.
The question for business leaders isn’t whether the tool is good. It’s whether your environment is set up to use it well, and that starts with understanding where your data goes when the add-in is active.
A Useful Launch and a Timely Prompt
The same week Perplexity Computer arrived in Microsoft 365, CNN filed a copyright lawsuit against Perplexity AI, alleging the company scraped and reused tens of thousands of its news articles, images and videos without authorisation.
CNN claims Perplexity marketed premium access to its content despite having no licensing agreement in place. I’m not sharing that to cast doubt on the product. As I said, I use Perplexity regularly and rate it highly.
The point is that no vendor is above scrutiny, and the timing is a useful prompt to think clearly about what due diligence looks like when a third-party AI tool moves inside your core work environment.
When a tool sits in the browser, the exposure is one thing. When it sits inside Word and Outlook, it’s another. Your documents, your prompts, your client communications are now in the same session as a third-party service running its own infrastructure, its own models and its own data agreements.
Getting clear on that boundary isn’t about being cautious for caution’s sake. It’s about using the tool confidently rather than discovering the gaps later.
Understanding the Security Boundary
Perplexity has positioned Computer as enterprise-ready, and the security credentials it offers are the right ones for a business tool. SAML SSO, audit logs and admin controls are meaningful features, and they show the product is being built with IT and security teams in mind.
But enterprise-grade features on a third-party add-in don’t automatically mean it sits within your existing Microsoft 365 security boundary. When your team uses Microsoft 365 Copilot, data stays within the security and compliance controls your tenant already has in place.
Microsoft operates as the data processor, and businesses with data residency requirements can be confident their content stays in-country.
A third-party add-in works differently. Prompts and responses may travel to that vendor’s own infrastructure, which means your data protection, retention and compliance settings in Microsoft 365 may not automatically extend to those interactions.
That’s not a reason to avoid the tool. It is a reason to ask a few specific questions before your teams start using it at scale:
If you can answer those clearly, you’re in a strong position to make an informed decision. If the answers aren’t there yet, that’s the gap to close before rollout.
Multi-Tool AI Is Now the Default, Not the Exception
Perplexity Computer is one example of a broader shift that’s already well underway.
The Microsoft 365 add-in ecosystem is growing quickly, and AI tools are a meaningful part of that growth. Most knowledge workers are already using several AI tools, whether or not IT has formally approved each one.
Microsoft has acknowledged this reality in its own product strategy. The Microsoft 365 E7 suite was designed partly because businesses couldn’t keep track of what AI was doing across their environment without a single governance and control layer.
As we covered in our piece on Microsoft 365 E7 and Agent 365, the agent registry surfaced tens of millions of agents in just two months of preview, which tells you something about the pace of AI proliferation inside business environments.
The answer isn’t to block tools like Perplexity Computer. It’s to have a clear picture of what’s running in your environment, what data each tool can access, and what controls are in place if something goes wrong.
Getting that picture right is what separates confident AI adoption from reactive clean-up.
Practical Steps to Get the Most From It
Getting ahead of this doesn’t require a large project. It starts with a few practical steps that most Microsoft 365 environments can move on today.
Know what’s already running. Start with the tools your teams are actively using across Office, browser extensions and integrations. If Perplexity Computer or similar add-ins are already in play, include those in your inventory. You can’t govern what you can’t see.
Check your data classification and permissions. Microsoft Purview gives you the tools to classify data by sensitivity and set policies around what AI tools can access. If that work hasn’t been done yet, this is the right moment to start.
A tool can only do what its permissions allow, and tightening that layer reduces exposure regardless of which AI is in the session. We covered the governance case for this in detail in our blog on unified AI governance for Copilot and beyond.
Set a clear acceptable-use position. Your team will take their cue from you. A short, plain-language policy that tells staff which AI tools are approved, what data they can use those tools with, and when to check with IT before adding a new integration goes a long way. It doesn’t need to be long. It needs to be clear.
Review vendor agreements before scale-up. If a vendor is going to sit inside your email and document workflows, their data processing agreement should be reviewed against your obligations under the Australian Privacy Principles, particularly in industries like healthcare, financial services and legal where client data carries extra protection requirements.
Use your Microsoft 365 admin controls. The admin centre gives you meaningful control over which add-ins are permitted in your environment. If you want to allow Perplexity Computer for a specific team or role while you assess it, that’s possible.
Broad access before a proper review isn’t the right approach in most business environments, but a controlled pilot absolutely is.
The Opportunity Is Real, and It’s Worth Pursuing Properly
I genuinely think tools like Perplexity Computer will improve the way knowledge workers operate inside Microsoft 365. The productivity case for in-app AI that can research, draft and model alongside your actual documents is strong, and it’s only going to grow as more vendors follow the same path.
As we discussed in our piece on Claude AI inside Microsoft 365, the model-diverse future of enterprise AI is already here. Business leaders aren’t choosing between one AI and another.
They’re managing several AI tools working across the same environment, and the businesses that get their governance in order now will be the ones using these tools with confidence rather than catching up later.
The conversation I’m having with most Australian business leaders right now isn’t “should we use AI?” It’s “how do we know the AI we’re using is working safely?”
For Microsoft 365 environments, the answer lives in your governance settings, your data classification, your vendor review process and your people’s understanding of where the boundaries are.
If you’d like to work through what that looks like for your business, our team at CG TECH is happy to help.
About the Author
Carlos Garcia is the Founder and Managing Director of CG TECH, where he leads enterprise digital transformation projects across Australia.
With deep experience in business process automation, Microsoft 365, and AI-powered workplace solutions, Carlos has helped businesses in government, healthcare, and enterprise sectors streamline workflows and improve efficiency.
He holds Microsoft certifications in Power Platform and Azure and regularly shares practical guidance on Copilot readiness, data strategy, and AI adoption.
Sources
Recent Posts
Popular Categories
Archives