Perplexity Just Embedded a Research Agent in Enterprise Slack. The Shortlist Gets Built Before Your SDR Emails.

On March 10th, Perplexity moved its Computer AI agent into the enterprise. Not as a standalone product. As a Slack integration.

Enterprise employees can now type @computer inside their company workspace and get multi-model AI research with citations, vendor comparisons, and workflow execution — all within the channel where the team already works. The launch was announced at Perplexity's inaugural Ask 2026 developer conference, with the product positioning aimed directly at Microsoft Copilot and Salesforce Einstein. Most of the coverage treated it as a competitive positioning story.

That framing is accurate but it's the wrong frame for anyone selling into enterprise. The structural change is simpler and more direct: AI vendor research just lost its address.

The friction that was doing work for you

Before Computer's enterprise deployment, AI-assisted research required a deliberate act. A buyer decided to open Perplexity or ChatGPT, form a query, read the results. There was friction — switching context, opening a new tab, figuring out how to query the right thing. That friction had a window in it. A moment before the buyer had formed a view where outbound could still reach someone with an open mind.

Slack removes that window. Not because the AI got smarter, but because the research now happens inside the tool where decisions already get made. When someone has a question in a team channel, the natural response is to ask in the channel. @computer is now in that channel, with SOC 2 Type II compliance, full audit logging, Salesforce and Snowflake connectors, and 20 models running behind a single query.

The question "who are the best options in this category?" gets asked in Slack and answered in Slack — with citations — before anyone thought to loop in the vendor team. Perplexity's enterprise customer base is already above 20,000 organizations.

TechCrunch's coverage of the product noted that Perplexity executives explicitly stepped back from maximizing monthly active users. The focus is a smaller, enterprise-weighted user base making significant decisions. "You don't hear us talk about MAUs ever," one executive said in a background briefing with press, "because we're not actually on a mission to get as many users as possible." The product was built for the accounts where vendor research has pipeline consequences, not casual queries.

It's not acceleration. It's a structural change.

The common interpretation of AI's impact on B2B buying is that it accelerated the research phase. Buyers do more research, faster. That's true, but it undersells what actually changed.

Acceleration implies the same process running quicker. What happened is different: the research phase is no longer a phase.

Phases have beginnings. Someone decides to do research. They open a browser. They build a list of queries. When research required deliberate context-switching out of Slack and into a separate tool, that decision moment was visible and had a time gap attached to it. Outbound, retargeting, and demand gen were all built around capturing people during or just after that moment.

When research lives in Slack, it becomes conversational. The question and the work happen in the same thread. "What vendors should we evaluate for this?" is asked and answered in the same interface where the follow-up question — "who should we reach out to first?" — gets asked two minutes later. The research phase became ambient. It runs continuously, as a side effect of normal team conversation, whenever someone has a question.

Forrester's October 2025 B2B predictions put a number on how far along this was before Perplexity even shipped its enterprise Slack deployment: 30% of B2B buyers used generative AI at the final commit stage of their purchase — not just during early research, but in the moment right before they signed. The same report found that 61% of purchase influencers said their organization has or will use a private generative AI engine to support purchasing. Perplexity Computer, deployed enterprise-wide with IT compliance and audit logging, is that private engine now.

What the shortlist looks like now

McKinsey's February 2026 procurement analysis described the shift from analytical AI ("show me the data") to agentic AI ("do it for me"). They named category copilots — agents that compare vendor offerings before a human has reviewed a single pitch deck — as "no-regret" implementations: available off the shelf, deployable now.

The companies McKinsey described as early adopters were using these tools before they had a Slack integration. The adoption curve on Perplexity Computer starts higher and runs faster, because the deployment barrier is lower. It connects to systems organizations already use, with credentials IT teams already manage.

This is the operational picture: research conversations about your category are happening continuously inside enterprise Slack at any company using Perplexity Enterprise. When someone types "@computer what are the top options for [your category]" and your brand doesn't come back in the cited answer, you are not on the shortlist. When that same person checks their inbox twenty minutes later, the SDR email from your team arrives to someone who already has a list — and you're not on it.

We covered this pattern when McKinsey's procurement data first surfaced the agentic shift — AI agents building vendor shortlists before human buyers get involved. The Perplexity enterprise Slack deployment is the infrastructure version of that pattern. Not experimental, not something forward-thinking companies configured in isolation. Twenty thousand organizations. SOC 2 Type II. Audit logging. This is the deployed layer.

What actually determines whether you appear

When Perplexity Computer researches a vendor category in Slack, it reads from the same editorial ecosystem it always has: TechCrunch, Forbes, Reuters, industry trade publications, and the tier-one outlets that have built their authority over decades. No ad buy gets you into that answer. No SDR email. No retargeting pixel.

MIT Sloan's research on AI-driven search found that large brands with dominant market share discovered smaller local competitors outranking them in AI-generated answers — because those competitors had more current editorial coverage in the right publications. Market share doesn't protect you in an AI-researched shortlist. Coverage does.

This is the operating mechanism of Machine Relations: earned media placements in trusted publications have always been the signal that determines brand authority. That was true when the reader was a human buyer reading Forbes on Saturday morning. It's true now that the reader is an ambient AI agent embedded in your buyer's Slack workspace, answering vendor questions in real time.

PR's core mechanism — getting your brand covered in publications that carry editorial weight — never stopped working. What changed is that the reader is now inside every enterprise workspace, answering research questions continuously, at the moment they get asked rather than waiting for a buyer to deliberately open a separate tab. The brands that show up in that answer didn't get there by optimizing for AI. They got there by earning coverage in the publications Perplexity already trusts.

The enterprise Slack deployment shipped two days ago, to 20,000 organizations. That's not a preview. That's the market your brand is operating in right now.

Related Reading

• AI Visibility for Fintech Companies: How to Get Cited by ChatGPT, Perplexity, and AI Search
• PR Strategy for Series A and B Startups: The Earned Media Playbook for the AI Era

---

If you want to see where your brand currently shows up when Perplexity researches your category — which publications are being cited, where you're absent, and what it would take to appear — the visibility audit runs in about 15 minutes.