Planning an AI pilot for claims? Check first: - One claim type, not all - A baseline for cycle time and errors - Direct access to claims data - A named owner - An agreed definition of success Full checklist 👇 hubs.ly/Q04BJvBR0 #AIinInsurance #InsurTech

Should AI approve a claim payout? No. But it can open the file, read estimates, flag what’s missing and suggest a reserve for review. AI prepares the work. People make the decisions. What to automate vs. keep human 👇 hubs.ly/Q04Bg-pG0 #AIinInsurance #InsurTech

Think about the last time you called your insurer about a claim. You probably waited, explained everything twice, and then heard “we’ll get back to you.” Most of that delay isn’t anyone’s fault. It’s files moving between systems, documents being checked by hand, and agents searching for the right policy details. That’s where AI is starting to make a real difference: - Claims files arrive complete and go to the right adjuster - Agents get draft answers based on the actual policy and claim status - Simple billing and status questions get answered even after hours And the important part: people still approve settlements, reserves and anything binding. We wrote about how insurers are doing this in practice 👇 hubs.ly/Q04zVsMz0 What’s the most frustrating step in your claims process right now? #AIinInsurance #InsurTech #CustomerExperience #ClaimsAutomation #InsuranceInnovation #EnterpriseAI

Ten years is a long time for a product to stay the same. Geofacets, Elsevier's research tool for geoscientists, kept growing its library to more than 1.75 million maps, figures and tables. The interface around it barely changed, and over time it became harder and harder to use. On top of that, it had to work for two very different groups. Academic researchers wanted every detail. Corporate engineers and executives wanted a quick answer. So we started by talking to the people who use it every day. We ran one-on-one interviews, looked at what competitors did well, and asked users to help us rank which features actually mattered. What we did: - Simplified the layout and added views tailored to each type of user - Added saved lists and personalized search results - Made filters easy to add and remove - Moved map search front and center - Made results pages quicker to scan and easy to share Sometimes the most useful thing you can do for a product is take things away. Full case study: hubs.ly/Q04zzM0c0 What our clients say about working with us: hubs.ly/Q04zzLNr0 #UXDesign #ProductDesign #UserResearch #DigitalTransformation

Some of the data was in an ERP, some in email threads, and some in a system older than half the team. Most of the problems we see in AI implementation aren't really about the model. They're about the organization around it: - Data lives in too many places - Legacy systems block integration - Nobody owns the outcome - Success was never defined - The pilot is too broad - Staff don't trust the tool - Governance shows up too late Each of these has a practical fix, and none of them require replacing your whole tech stack. We wrote down what works for each one 👇 hubs.ly/Q04zsK-90 Which one sounds familiar? Tell us in the comments. #AIImplementation #EnterpriseAI #AIStrategy #AIAdoption #AIGovernance #DigitalTransformation #AIAutomation

Which workflow should you hand to AI first? Claims intake, contract review, invoice processing… Picking where to start is often the hardest part of AI implementation. So we turned our framework into a simple checklist: → How to score and pick your first workflow → 4 levels of AI autonomy, from human-in-the-loop to AI agents acting within limits → A one-page owner sheet for every workflow → How to run a shadow test before going live → A 90-day plan from AI pilot to production Save it for your next planning meeting. The full guide, with gate criteria and examples, is here 👇 hubs.ly/Q04z8wKw0 #AIAutomation #WorkflowAutomation #AIAgents #EnterpriseAI #AIImplementation #AIGovernance #DigitalTransformation

Don't start your AI rollout by replacing your core system. Run AI in shadow mode on one workflow, compare it to your team's decisions, and set rollback rules before the pilot starts. How to integrate AI without disrupting operations: hubs.ly/Q04yS_Vy0 #AIIntegration #EnterpriseAI

An AI readiness scorecard shows what your team believes. An assessment shows what's verifiable. Did your last pilot stall, or is the workflow regulated or running on legacy systems? Skip straight to an in-depth assessment. 5 questions to decide: hubs.ly/Q04yzKrl0 #AIReadiness #EnterpriseAI

Two days at AI Summit Barcelona, and easily one of the most useful events we've attended this year. Strong talks, and even better conversations with people actually building with AI, not just talking about it. Most companies no longer need convincing that AI works. What they need is a way to make it work across the whole organization: connected to their existing systems, trusted by their teams, and governed well enough to scale. That part isn't about picking the right model. It's about the workflows, integrations and guardrails underneath it. It's slower, less flashy work, and it's where the real results come from. Great to see so many teams at the summit asking exactly these questions. Thanks, Tom, for representing us. Were you there too? We'd love to compare notes. #AISB26 #AISummitBarcelona #EnterpriseAI #AIGovernance

Most people know WWE for the spectacle. Fewer know the work it does off screen. When WWE wanted to bring its social-responsibility work forward, the challenge was not telling people what it does. It was showing the softer side of a brand built on spectacle, in a way that felt genuine. We designed and built the digital home for WWE Community, the property that carries its commitments across four pillars: Diversity, Education, Military Support, and Hope. What we did differently: we led with UX, not a corporate-CSR template. The experience had to be as considered as the brand itself, and it had to land on a tight timeline, built in close partnership with WWE's team. The result is a purpose-driven experience that reaches beyond WWE's core fans and makes its social impact something people actually want to explore. That is the work we care about at DOOR3: enterprise UX and digital product work that makes a brand's harder-to-tell stories land. Case study: hubs.ly/Q04yktvj0 What our clients say: hubs.ly/Q04yktFD0 #UX #DigitalProduct #CSR #ProductDesign #EnterpriseUX

A high AI readiness score is not the same as being ready to build. Most quizzes measure what you say, not whether your data, workflows, and governance can actually handle AI. How to tell a real assessment from a checkbox: hubs.ly/Q04ycgYs0

The enterprise AI programs that create real value share one discipline the rest skip: they treat AI as a business strategy, not a model selection. In our new piece, we break down what separates the companies reaching production-grade results, with examples from Rio Tinto, HCA Healthcare, DBS Bank, and Dentsu Creative. The patterns that define the leaders: - Business goals set the direction before any tool is chosen - Governance is built to enable progress, not only to block risk - Use cases are sequenced by ROI and feasibility, not by hype - Organizational readiness matters more than the model The through-line is clear: a formal enterprise AI strategy is what turns a pilot into production and a budget into an outcome. Read the full breakdown: hubs.ly/Q04xXLFf0 #EnterpriseAI #AIStrategy #DigitalTransformation #AILeadership

AI has already transformed individual productivity. The harder question is what it changes across a whole organisation. Next week DOOR3 is on the ground at AI Summit Barcelona, September 22 and 23, partnering with Miro as part of Barcelona AI Week. We will be there to talk about one thing: rethinking AI ROI beyond individual productivity. Modern teams do not struggle because they move too slowly. They struggle because they move fast without shared context, so individual AI gains never compound into team-wide impact. The real return sits in shared context and collective decision-making, and in how leaders structure teams, workflows and technology around it. If you are attending, come find us. Tom McClean from DOOR3 and Rory Keddie from Miro will be there both days and would welcome the conversation. And the night before, September 21, we are hosting an intimate dinner on the same theme, Rethinking AI ROI, with a curated group of leaders. A few seats remain: RSVP at hubs.ly/Q04xRd070 See you in Barcelona.

How do you redesign a higher education website that has grown into hundreds of thousands of pages across dozens of disconnected microsites? That was the challenge at the New Jersey Institute of Technology (NJIT). A sprawling digital presence that even experienced users struggled to navigate, and that no longer reflected the quality of the institution's research or the programs it wanted to fill. DOOR3 rebuilt it around the prospective student, not the org chart: - A topic-based information architecture and mega-menu that replaced the microsite sprawl with one coherent path. - A flexible modular content system, 50+ content blocks across 26 core pages, so every program from art to engineering fits one design language. - An interactive Major Finder that lets prospective students explore programs and the career outcomes behind them. All grounded in real UX research: usability testing, card sorting, and personas for students and families, plus a prioritized roadmap to keep improving after launch. The result is a cohesive digital presence that finally does the university justice. Read the full case study: hubs.ly/Q04xRq390 We are a 2025 Clutch Global Award winner for UX Strategy and User Research. See our reviews: hubs.ly/Q04xRq1z0 #HigherEducation #HigherEdMarketing #UXDesign #WebsiteRedesign #DigitalTransformation #CaseStudy #DOOR3

What does a practical AI roadmap look like when it comes from someone who actually builds the software, not just advises on it? Our CEO Alex Asianov sat down with a room of CEOs for a candid session: "Ask Alex, A Practical AI Roadmap, No Matter Where You Are Starting From." No tech lecture. A business-level conversation about making AI pay off: the four stages of AI adoption, where to start, how to cut through the noise and judge who can really help, and how to fund the program so it pays its own way. If you are a leader weighing where AI actually pays off, this one is worth an hour. Watch the full session: hubs.ly/Q04xJY4J0 #AI #Leadership #DigitalTransformation #EnterpriseSoftware

81% of in-house legal teams already use AI. Only 55% of outside firms do. The gap is not adoption. It is implementation. Buying a legal AI tool is easy. Deploying one that clears ABA ethics rules, protects privilege, and actually earns its keep is where most firms stall. The difference is a disciplined rollout, not a bigger budget. Our new Legal AI Assistant Implementation Checklist maps the 10 steps from data audit to firm-wide launch: how to sequence workflows by ROI, build an attorney review protocol, and pilot one high-value use case before you scale. Teams that follow it see up to 70% less document review time and 60% faster due diligence. For general counsel, managing partners, and legal operations leaders who would rather lead the shift than catch up to it. Read the checklist: hubs.ly/Q04xDBQq0 #LegalAI #LegalTech #LegalOperations #GeneralCounsel #AIGovernance #LawFirmInnovation #DocumentReview

Your associate just pasted a privileged deposition summary into a public AI chatbot. Is it still privileged? A federal court has already weighed in. In United States v. Heppner, the Southern District of New York held that chats with public AI are not attorney-client communications. Put privileged material into a public tool and you can waive the privilege. This is an ethics question now, not only an IT one. ABA Formal Opinion 512 ties generative AI directly to the duties of competence, confidentiality, and communication, and North Carolina's 2024 ethics guidance warns specifically against feeding client-specific information into public AI that trains on user inputs. The answer is not to ban AI. It is to run it privately: tenant isolation, a contractual ban on training, and full audit logs, with attorney supervision still in the loop. Our latest piece breaks down private legal AI vs public AI tools, what the rules actually require, and how to define approved tools and banned data classes before you scale. Read it: hubs.ly/Q04xnk0h0 #LegalAI #LawFirms #AttorneyClientPrivilege #LegalTech #ABA512 #LegalOps

When a data-heavy R&D tool is hard to use, the science slows down. Elsevier came to DOOR3 with ChemSel, a research and analytics tool built to support product development across the chemical industry, a field that invests tens of billions in R&D every year. The information was all there. The experience of getting to it was not. We ran a focused discovery and design engagement: mapping how chemical R&D teams actually search, interpret, and act on information, then reworking the interface through wireframing and rapid iteration, all while staying true to Elsevier's brand. The goal was simple to say and hard to do, make a dense scientific tool faster and more intuitive for the people who depend on it. This is the work we do best: taking complex, data-rich enterprise software and making it feel effortless for the experts who live in it. Full story: hubs.ly/Q04x715x0 See our verified client reviews on Clutch: hubs.ly/Q04x714F0 #EnterpriseUX #ProductModernization #CustomSoftware #DigitalProduct #UXDesign

AI can read a contract in seconds. It cannot tell you which risk is worth walking away from. That gap is where most legal AI decisions go wrong. Our new piece lays out what an AI contract review assistant genuinely does well, and where it quietly fails. What it does well: - Extracts clauses, parties, obligations, and renewal dates with source citations - Flags deviations from your playbook and templates - Compares redlines and summarizes changes across large document portfolios Where judgment still sits with counsel: - Novel deal structures and edge-case risk - Multi-jurisdiction conflicts - Anything client-facing, and privilege when a vendor trains on your matter text The firms getting real value are not buying the most-hyped tool. They pilot one reversible document class, require pin cites and named reviewers, measure omission rates alongside speed, and keep matter text in controlled systems. If you are evaluating contract review AI for a law firm or in-house team, this is the adoption playbook that separates real value from a flashy demo. Read it: hubs.ly/Q04x09Ps0 #LegalAI #ContractReview #LegalTech #AIadoption #InHouseCounsel

74% of legal professionals using AI use it to summarize documents (Thomson Reuters). It is the most common AI use case in law, and the one that is least forgiving when it goes wrong. Use cases, risks, and the review gates that keep it defensible: hubs.ly/Q04wRS0G0 #LegalTech #LegalAI #AIforLawFirms #LawFirmLife #LegalOperations #GeneralCounsel #AITools