Joshua Glover

Joshua Glover

IT leadership · Ottawa, Ontario

I build technology functions that work for the business they serve.

Over 25 years at Bell, MD Financial Management, Smart & Biggar and IPH, I've taken on environments that were complex, under-structured or mid-change and left them with capability, process and systems the organization could rely on. Most recently, as Regional CTO for IPH in Canada, I led a 45-person IT organization through three acquisitions while taking AI from pilot into production.

I lead from inside the business: start with the outcome the organization needs, then decide whether technology is the right lever.

Case studies

What I've led, with the numbers

Three acquisitions, two integration models

IPH Limited · Regional CTO, Canada · 2023–2026
  • $850Kannualized IT synergies, Bereskin & Parr, above target
  • $500Kplanned infrastructure spend avoided
  • 20%lower IT operating cost at Robic, no staffing cut
  • 0turnover on the IT team through three acquisitions

Situation

IPH acquired three Canadian IP firms while I ran IT for the Canadian group: about 800 users, eight offices, two independent legal entities and an operating budget of about $8M. The firms couldn't all be integrated the same way. Ridout & Maybee and Bereskin & Parr were merging fully. Robic had to stay independent.

What I did

I assessed each firm's infrastructure and systems and built an integration roadmap for each. We ran two models in parallel. For the full mergers we decommissioned data centres, collapsed networks and moved all data and endpoints onto group standards. For Robic we kept the firm independent but merged its IT team, Microsoft tenant, security tooling, service desk and endpoint management.

We consolidated the IP management systems across firms, migrating millions of client records on M&A timelines, and put every Canadian firm on one Microsoft 365 tenant.

Three integrations in a row are hard on an IT team. Systems change, reporting lines change, and people worry about their own jobs. I kept the team inside the change instead of on the receiving end of it. Whenever significant change was announced, I held an all-team IT call, and people knew they could ask me anything and get a straight answer. My weekly meetings with my managers deliberately left out project status. They covered the team's questions and concerns, and how the function itself was running: process changes, prototypes, internal tooling.

Outcome

The Bereskin & Parr IT workstream delivered $850K in annualized synergies, beating its target, and avoided $500K in planned infrastructure spend. That was part of $4.5M in firm-wide synergies against a $3.7M target. Robic's IT operating costs fell 20% with no cut to staffing or client support, while security and monitoring coverage grew. One IT team now supports two legally independent firms.

The IT team came through all three acquisitions and the restructuring around them with no turnover. When engagement surveys started late in my tenure, second-year scores were up 5% over the first, during the heaviest integration work.

AI from pilot to production, inside a law firm's constraints

IPH Limited · Regional CTO, Canada
  • 50%less legal assistant time on trademark filing
  • ~100practitioners: Harvey AI rolled out to the full target group
  • 4vendors in the Canadian AI RFP

Situation

IPH wanted AI in its legal work. An IP firm holds confidential client data under Canadian privacy law, so every tool had to meet data governance and sovereignty requirements before it reached production.

What I did

I ran adoption through AI champions and a community of practice. Champions brought ideas forward and tested whether they'd extend past their own work. We completed a 50-user pilot of Harvey AI for legal drafting, then rolled it out to the full target population of about 100 practitioners.

I wrote a Canadian AI vendor RFP covering four vendors. When global reprioritization cancelled it, we moved to building in-house on Microsoft Copilot Studio. I directed a trademark filing classification tool from a proof of concept for one champion, to a test with two more users against different work styles, to the production build.

I led the architecture review for an agentic email-triage solution on a local Kubernetes cluster, keeping data inside our Microsoft environment, with an API integration framework so core systems could change without breaking the AI layer. I worked with the Canadian privacy officers on jurisdiction-specific input to IPH's enterprise AI governance framework.

Outcome

The classification tool cut legal assistant time on the trademark filing process by 50%, removed manual review steps and a legacy software dependency, and is targeted at up to 200 users. The email-triage solution was designed to fully automate inbound triage, freeing about 15 FTE of capacity for higher-value work.

IT that went looking for the work

Smart & Biggar · IT Director, then CIO · 2018–2023
  • 2×delivery function, overall headcount flat
  • Weeks → hourson processes we automated

Situation

Smart & Biggar is Canada's leading IP firm, with five offices. When I joined as IT Director, two newly implemented enterprise systems, case management and billing and finance, weren't yet stable. The IT team was built around support and spent its time on the problems the business brought to it.

What I did

I stabilized the core systems and restructured IT to move capacity from support into delivery.

Then I took IT to the business instead of waiting for requests. I coached the Operations Director and worked with his managers to find where work was slow or stuck, and to break down old barriers between their teams and IT. What followed was a long run of small wins: automations that turned weeks of work into hours, and processes to clear and prioritize backlogs, using RPA and custom development where they fit. That partnership is what moved me from IT Director to CIO.

Alongside it, we modernized the firm's tools and devices and built disaster recovery with site-to-site failover. I built an ISO/NIST-aligned security policy framework the team could actually maintain, and reported quarterly to the Executive Committee and annually to the full partnership.

Outcome

The delivery function doubled with overall headcount flat. The improvement work was many small wins rather than one big number, and together they changed how the business saw IT.

The foundation paid off in a way we hadn't planned for. When the firm moved to remote work in March 2020, day-to-day productivity held, because the tools, devices and failover were already in place. I then led IT through the firm's acquisition into IPH.

Building delivery discipline, one role at a time

MD Financial Management · 2009–2015 · $30B+ AUM
  • 20–50%post-launch opex overruns, eliminated
  • $800Kannual outsourcing spend removed
  • 300%increase in CRM delivery frequency

2010–2014

Release management. Release management was outsourced to Deloitte. I brought it in-house and rebuilt it with one redeployed, retrained person, removing about $800K a year in spend.

Separately, I introduced agile delivery on the CRM platform. Within six months, delivery frequency rose 300% and minor enhancement completion doubled, with no slip in project commitments.

2014–2015

Enterprise architecture. Before this role existed, I'd spotted a pattern: projects landed on budget, then added 20–50% to the ongoing IT operating budget once they went live. When MD restructured and moved the PMO out of IT into the Strategy team, I pitched leadership on that gap in budget and project sprawl. They created a role to lead enterprise architecture and IT project delivery together, and promoted me to build and lead it as AVP Enterprise Architecture.

I built an EA team that assessed every new project against documented principles, and a Technology Stream Lead practice that carried each design into delivery. I wrote the target architecture from scratch and reviewed it every year. The overruns stopped and the operating budget stabilized. One year I advised the CIO not to spend the $2M enhancement budget because the backlog wasn't worth it, and held that position when an operational VP challenged it.

Widening the scope to win the business case

Bell Canada · Sympatico · 2005–2006
  • 25 → 5active platforms
  • 80%simpler agent desktop tooling
  • $25Mprogram, 100+ people, three vendor teams

Situation

Sympatico needed to replace its order management system with modern tooling, as the foundation for IP services beyond internet, such as VoIP. It was an expensive program whose benefits would arrive with future products, not right away, which made it hard to get the business unit behind it.

What I did

I broadened the scope. Ordering, fulfillment and client care ran on 25 systems plus a layer of shadow IT tools, and agents worked across a cluttered desktop. Bringing the agent experience and that clutter into scope gave the business unit benefits it would feel right away, and that won its support for a major program.

I was senior project manager and lead architect on the resulting $25M, 18-month program, directing more than 100 people across four workstreams and three vendor teams. I chaired the order management RFP and owned the selection. I designed an integration translation layer so dependent Bell systems kept working while the platforms underneath were replaced. When business stakeholders pushed back mid-program, I kept delivery moving while working toward agreement, and we got it.

Outcome

25 platforms went down to 5, and agent desktop tooling became 80% simpler. The new order management foundation sped up later product launches. This is where my approach to business partnership started: find the version of the IT need that the business wins from too.

Recent work · since March 2026

Building with AI myself

Since leaving IPH I've done advisory work and built with AI tools myself, to see adoption from the user's side.

Management System

  • AI-assisted development
  • Agentic workflows
  • MCP
  • Human-in-the-loop
  • API design
  • Build vs. buy

Problem

My work was spread across Notion, ClickUp and spreadsheets, and planning happened inside AI chat sessions with nothing durable behind it.

Decision

I ran the build like an enterprise program: requirements, detailed design, a security review, then a dependency-ordered plan of 2–3 hour build blocks. AI-assisted development throughout: Claude was the build partner with read-only access to the code, and I reviewed and committed every block.

The app exposes a REST API that my AI agents use as a tool, with a scoped token for each project and every call audit-logged; elsewhere they reach Notion and Gmail through MCP connectors. Scheduled AI automations write only to the app's own data, and nothing happens downstream without my approval: human-in-the-loop by design. Nothing touched a live external system until a staged cutover in October 2026.

Once the app could do the job, I ran a keep, reduce or exit review on ClickUp, the SaaS tool the whole system had run on, and exited it. Every open item moved over with dry-run-first scripts, and old links still resolve. It's a build-versus-buy call I expect more organizations to make as AI makes targeted in-house tools cheap to build.

Lesson

AI made the build fast. The program discipline is what made the result trustworthy, and it mattered more with AI, not less.

Retrospective loop for AI work

  • AI governance
  • Context engineering
  • Continuous improvement
  • Human-in-the-loop

Problem

Working with AI, the same mistakes kept coming back. The lesson from Monday's session wasn't there on Wednesday. It's the lessons-learned gap every organization has, playing out one session at a time.

Decision

I built the retrospective into the operating process instead of relying on memory. Any session can raise an issue, and it goes into a review queue. A weekly review decides whether the fix applies to all work or one project, then writes it into the standing instructions the AI loads at the start of every session, and the item leaves the queue. Each rule lives in one place, so a fix reaches every project at once. An issue that recurs is the signal to fix it structurally rather than log it again.

I wrote about the organizational version of this problem in The Retrospective Gap.

Lesson

Getting the lesson is easy; the AI will tell you how to do better. Keeping it is a governance problem: AI behaviour needs a change process, the same way systems do.

Keep It Local

  • Agentic research
  • Claude skills
  • Human-in-the-loop
  • Data governance

Problem

People who want to buy Canadian can't easily tell who owns or controls a brand, and existing sites mix fact with opinion.

Decision

I treated ownership and control as separate facts. The score is built only from sourced facts, and any editorial adjustment is disclosed and must cite a source.

Research runs as an agentic workflow: AI agents source and assess companies against a written rubric, packaged as reusable Claude skills so every run follows the same method, and a person reviews before anything is published. Least privilege by design: AI never has database credentials or push access, and every change goes through files I commit. It launched in September 2026 with 366 brands in the first push. keepitlocaleh.ca

Lesson

AI research at scale is only as good as the rubric and the review step. The human review is where the credibility comes from.

Guided AI setup tool

  • AI enablement
  • AI onboarding
  • Rapid prototyping

Problem

Non-technical people around me didn't know where to start with AI, and generic advice didn't stick.

Decision

I built a guided, interview-based walkthrough for setting up Claude with Projects and connected tools like Notion. It starts from the tools a person already has and gets them an early win before anything more complex. The working version came out of a single one-hour session that had been booked for scoping, and it now has its first outside user.

Lesson

Adoption starts with the tools people already use and one early win, not a tour of capabilities. It's the same pattern I saw with the AI champions at IPH.

Leadership

Where I've been useful

Most of my roles started with something that wasn't working: financial reports with errors in them, newly implemented enterprise systems that hadn't stabilized, an IT budget that kept growing after projects closed, acquired firms running on their own infrastructure. The work was to find the real cause, fix it, and leave behind a structure the team could keep running.

  • IPH LimitedRegional CTO, Canada2023–2026
  • Smart & BiggarChief Information Officer; IT Director2018–2023
  • Abilis SolutionsClient Solutions Director, Telecom, Media and Technology2015–2018
  • MD Financial ManagementAVP Enterprise Architecture; Manager, Release Management; Project Manager2009–2015
  • Bell CanadaAssociate Director, IT Strategy; program, architecture and project leadership2000–2009

How I work

The habits behind the results

I start with the outcome. Whether technology is the right lever is the first decision, not an assumption.

I ask why until the cause is clear. A lot of IT problems are process or ownership problems wearing a system's clothes.

Benefits get measured after launch, not only in the business case. I helped set up MD's PMO on that basis and have carried it into every role since.

Change management is part of delivery, not a phase after it. I've worked that way since Bell and got Prosci certified in 2010.

Governance should be sized to what the team can maintain. A framework nobody keeps up is worse than a lighter one people use.

I'll recommend not spending money when the value isn't there.

My team hears about change from me and can ask me anything. Credit goes to them when it works.