See exactly how Altria transformed their legal service delivery - watch the full webinar.
From Pushback to 90% Adoption: Altria’s AI Legal Front Door Transformation
.jpg)

Legal requests scattered across channels left work invisible and untracked, so modernizing the department had to start with fixing intake.
Altria's vision was changing, and the legal department was expected to change with it. The enterprise had committed to a company-wide modernization, and leadership had a clear mandate for every function: be quicker, more efficient, and more effective.
The company had already restructured several functions under a Global Business Services model, pulling routine transactional work out of individual teams and centralizing it. Legal was next. Its goals echoed the rest of the enterprise, with one added constraint. The department committed to modernizing "while reducing outside counsel spend." For a lean team that leaned heavily on external firms, that raised the stakes.
The day-to-day problem was intake. Every legal request, whether from the business or from inside the department, arrived by phone, email, or Microsoft Teams. Requests got buried in inboxes and in some cases the team had to request work from outside counsel again because documents could not be located timely. The department also carried a familiar reputation: that legal was the holdup. And no one could say whether that was fair or not, because there was no way to see who was holding a request or how long anything actually took.
The legal team needed a system it could easily extract information from when required, and the enterprise tool Altria had at the time did not offer that confidence. It also failed to meet people where they were, with no support for the phone, email, or Teams channels the department actually used. Joy Thorpe, Director of Strategy for the Legal Center of Excellence at Altria, needed something that she or her team could customize on their own, without hiring a consultant or waiting in a ticket queue.
Altria centralized repetitive work and rolled out an AI-powered legal front door in two deliberate phases, a web form first and an AI agent second.
Thorpe decided to tackle this challenge by implementing a Legal Front Door. She worked with a consultant and a small project team to map every process in each practice group. They found the same redundant, repetitive tasks repeating across all of them. Centralizing that work into a Center of Excellence promised real efficiency, and it gave the new intake system somewhere to land.
Altria went with Checkbox for their Legal Front Door due to ease of use, self-service customization, responsive support, and commensurate price. One of Thorpe's senior managers summed up the deciding factor, telling her the platform felt like one that would grow with the team.
The rollout came in two phases.
Phase 1: The Legal Front Door
The legal team built a web form as a minimum viable product, gathering typical request types from a representative in each practice group, then testing and tweaking until the workflows made sense. It went live in 52 days to hit the September 1 effective date, scoped deliberately to internal legal use at first so the team could protect privilege while it learned. Thorpe was explicit with the department that this was phase one and that updates would keep coming.
Phase 2: The AI Agent
Once people were comfortable with the web form, the team added an AI agent that collects the same information through email or the Microsoft Teams app. Send it a request mid-meeting, and if anything required is missing, it politely asks for it, a due date, for example, before booking the request into the back-end dashboard. The agent searches through a defined document repository, and when it lacks the information to answer, it declines and points the user to the web form rather than guessing.
Overall, Altria’s legal team built:
- A legal front door web form
- An AI intake agent that captures requests through email and Microsoft Teams
- A document repository inside Checkbox with AI retrieval, so common references like the approved outside-counsel firm list surface on request
Ask Al
The team gave the legal front door a name and a logo: Ask Al, where "Al" stands for Altria. The idea came from a colleague in IT during a working session, and Thorpe ran with it, putting the logo on the front page of the legal front door and on every tips-and-tricks update that followed. The brand took on a life of its own. Now, when people stop Thorpe in the hallway, they barely finish their question before asking, “Should I just ask Al?”
Why Checkbox?
- Easy to Use — A complicated tool would have died on arrival against the existing resistance.
- Self-Service Customization — Thorpe and her team could change a field, a form, or a report themselves, with no consultant and no queue.
- Scalability & Adaptability — Responsive support from the Checkbox team meant Altria could build and iterate in real time.
- Channel Flexibility — Web form, email, and Microsoft Teams meant the tool met people where they already worked.
By managing resistance and keeping the tool easy to use, Altria lifted adoption from a modest 60% target to roughly 90% and finally gained real visibility into workload.
Joy Thorpe set the adoption bar low on purpose. Given how strong the resistance to an intake system had been, her initial KPI was somewhere around 60%. In the end, the system reached roughly 90%.
This number reflects a change-management story more than a software story. Thorpe sorted the department into two groups and worked them differently. The first group was open to adopting it once they understood its value, so she supported them with training, simple “tips and tricks” emails, and clear success stories, like the attorney who got two hours of his day back. The second group would adopt the system only if it was required, so the general counsel set a clear expectation in a department meeting: everyone would use the system, himself included, and he did.
When asked what drove adoption most, Joy came back to one thing: ease of use, plus options.
Users could pick the web form for detailed requests, email when that was faster, or direct message when a request came up mid-meeting. The payoff for the department is visibility. Every request lands in one dashboard, so work no longer sits in a single inbox or a personal queue. Joy can see what is coming in, who is available, and where to route it, which lets her actively manage resources across the compliance professionals, paralegals, and administrative coordinators in the center.
The next step is opening the front door to the wider business, with supervised AI routing and the data foundation to measure turnaround time and outside counsel spend.
The current build is internal, and Thorpe's roadmap points outward. The goal is for the business itself to submit requests through the tool, with AI helping capture them and supervised routing sending each one to the right place, whether that is the center or an attorney directly. Two things have to mature first: privilege protection, including guardrails so people do not over-share in a request, and the data foundation that will finally make turnaround times, and eventuall
At Altria that means an AI governance board, AI policies, contractual clauses preventing vendors from training models on company data, and an internal program that gives employees 20 hours of structured AI training. It also means human review. In a litigation environment, Thorpe notes, the team treats AI as a starting point that still requires fact-checking, never a replacement for judgment.
What Does This Mean For Your Legal Team?
- Start from the pain point, not the tool — Thorpe's first question is always what problem you are trying to solve. "What is the problem you're trying to solve? That is the first question you need to ask."
- Ship and iterate — Waiting for perfect kills momentum. Launch phase one, tell people more is coming, and build as you go.
- Meet people where they already work — Giving users the web form, email, and Teams removed the friction that usually sinks adoption.
- Secure visible executive sponsorship — The general counsel using the system in public moved the holdouts faster than anything else.
- Brand the experience — A name and a logo turned an intake system into something people ask for by name. Branding can improve adoption.
- Keep AI controlled and governed — A constrained agent that admits what it does not know, backed by clear governance, earns trust.
Book a Demo
See the New Era of Intake, Ticketing and Reporting in Action.
