Grant Thornton's Network Fixes Problems Before Anyone Knows They Exist

A Problem That Never Became One

When a DHCP service fails, the typical scenario is grim: employees arrive at work to discover devices connected to the network but unable to access the internet, grinding productivity to a halt. At Grant Thornton Advisors, something different unfolded. By the time staff members logged in, the disruption had already vanished. No support tickets were filed. No help desk intervention occurred. No post-mortem meeting was scheduled. The workforce simply proceeded with their day, oblivious to the infrastructure crisis that had transpired.

Laura Lincoln, Director of Infrastructure Operations at Grant Thornton Advisors, frequently references this incident when describing the organization's transformation. The DHCP outage itself was unremarkable—such failures generate routine help desk requests across virtually every enterprise. What sets this story apart is what followed: nothing. HPE Marvis minis, AI-native digital experience twins embedded in HPE Mist, continuously simulate user connections across the network even when no actual users are present. These tools flagged the DHCP failure and immediately surfaced details about the malfunction, its location, and its timing. The result: no company-wide service interruption, no backlog at the service desk, and uninterrupted client delivery. What remained a technical issue never escalated into a business crisis.

Life Before the Fix

The contrast with the past is stark. Prior to implementing HPE Mist, Grant Thornton Advisors' network engineers spent considerable effort manually configuring and fine-tuning network settings. Stability proved fleeting; configurations that functioned smoothly one day would deteriorate within forty-eight hours. Escalations arrived constantly, frequently circumventing the formal ticketing process entirely. Users had discovered that submitting complaints through official channels consumed more time than simply alerting whoever happened to be nearby.

HPE Marvis now possesses the capability to remediate certain problems entirely without human intervention. When network switch ports serving conference-room audiovisual systems become unresponsive, Port Stuck—an automated HPE Marvis self-driving action within HPE Mist—identifies the malfunction and resets the port automatically. According to Lincoln, "We didn't have to get a ticket on it. We didn't have to pull in an AV team or the network team or get a user complaint that the AV rooms didn't work because it fixed itself." Frequently, her team learns of these autonomous repairs only by reviewing system logs after the fact.

The volume of complaints had grown sufficiently severe that Lincoln began documenting them independently, since no existing system was capturing the full scope. When Grant Thornton Advisors announced its transition to a new platform, the incumbent vendor made no effort to retain the business. They essentially conceded that they could not match the direction the technology was heading. This candid acknowledgment made the decision to migrate to HPE Mist considerably simpler than Lincoln had anticipated. Following a pilot phase, the broader deployment proceeded rapidly.

From Proving to Fixing

The most consequential shift wasn't tied to any particular capability. Instead, it reflected a fundamental reorientation of where the team directed its efforts. Lincoln characterizes the previous approach as manually substantiating failures—assembling evidence, locating affected users, and reconstructing events in retrospect. The network now performs much of this investigative work autonomously, enabling her team to spend less time confirming that problems exist and more time actually solving them.

Implementation also became more straightforward. Network switches arrive at locations ready to operate without advance configuration, approaching a plug-and-play experience that allows non-technical on-site personnel to establish connectivity. Firmware updates, previously requiring a designated individual to oversee the process, now execute on automated schedules during off-hours without manual supervision. The advantage extends beyond mere velocity. Consistency matters more: executing the same operation identically each time eliminates a class of errors that previously stemmed from human variability.

Proving the Value of Nothing Happening

Presenting this narrative to budget committees presents a genuine challenge, and Lincoln acknowledges this candidly. The most desirable outcomes in infrastructure management are inherently invisible—a detected and resolved problem registers as a non-occurrence from the business perspective. Rather than highlighting features or uptime statistics, Lincoln presents leadership with process-oriented metrics:

1. How quickly was an issue detected?
2. How quickly was it isolated?
3. Was service restored before users were affected?
4. Did the same failure recur?

These four questions constitute her framework for demonstrating infrastructure resilience gains without relying on a single dramatic intervention. Additional signals reinforce the narrative. Unsolicited user satisfaction surveys now consistently exceed 90% across most office locations.

Prior to deploying HPE Mist, network performance issues were frequently raised directly to me or other channels, bypassing a formal ticketing structure altogether. Those informal reports effectively just went away altogether. Now, our users have actually come back to thank us for everything working well

Laura Lincoln

Scaling Without Building

As Grant Thornton Advisors expands through acquisitions and manages 19+ entities, its current infrastructure delivers a scalable framework for authentication and network access control without requiring dedicated on-premises hardware or other physical components traditionally associated with conventional NAC deployments. By leveraging HPE Mist Assurance, cloud-managed authentication services, and centralized access policies, the team maintains visibility into authentication performance, onboarding metrics, and user experience while accommodating growth without substantially increasing operational overhead.

What Comes Next

Lincoln's vision for the future extends the same principles further. Her network already detects and, in select scenarios, resolves its own problems. She is monitoring for the moment when it can also comprehend intent and propose or execute corrective measures with minimal human involvement. She is also pursuing visibility that transcends her own infrastructure into segments of the internet beyond her control, such as carrier-level latency that presently surfaces only as an isolated complaint from a single office.

It mirrors the challenge she has already addressed: rendering failures visible and correctable before anyone must recognize them. A self-driving network is fundamentally engineered to accomplish precisely this: making failures visible and fixable before anyone notices.