About
Most Copilot and AI initiatives don’t fail because the technology is bad.
They fail…
They fail…
Activity
3K followers
-
Robert Crane shared this“Use the best available AI model” is not a sensible Copilot strategy. Not when the best-performing model can consume more than ten times the credits of another model that may still complete the job adequately. I’ve been running the different LLMs available through Copilot services against the same standard prompt. That gives me something more useful than vendor claims or synthetic benchmarks: comparable outputs from an actual business task. In my previous testing, Opus 5 in Cowork was the winner. Microsoft has now added Fable 5.1 to Cowork, so I repeated the test and used Grok to evaluate the result. Fable 5.1 is the new winner. It produced the best result from my standard prompt, but it was also the most expensive: → Fable 5.1: 3,453 credits → Opus 5: 2,240 credits → GPT 5.5: 1,200 credits → GPT 5.6 Terra: 260 credits That creates a far more useful conversation than simply asking which LLM is best. What is the work worth? For an important client report, proposal or complex analysis, paying more for the strongest result may make perfect sense. For a routine summary, first draft or internal document, using the most expensive model could simply be waste. The goal should not be to use the smartest model for everything. The goal should be to use the most appropriate model for the job. I’ve published the updated results, evaluation and supporting files here: https://lnkd.in/g2pex2xC A few questions I’d genuinely like your take on: 1. Do you measure the cost of individual AI-generated business outputs? 2. Where would you draw the line between “good enough” and “best available”? 3. Would automatic model routing make you more comfortable with this decision, or less? AI quality has a price. The real skill is knowing when that price is worth paying. #Copilot #Microsoft365 #MSP #LLM #AI
-
Robert Crane shared thisYour MSP may not need another idea. It may need an honest diagnosis. When business feels slow, the instinct is to add something. A new service. A different price. Another automation platform. An AI offer because another provider appears to be doing well with one. Activity returns, so it feels like progress. But copying another MSP’s move without understanding its context is dangerous. A large MSP with mature delivery can launch something that a smaller provider cannot yet support. A solo consultant can make commitments that become difficult to deliver consistently across a team. A cybersecurity specialist can position services differently from a generalist carrying years of legacy expectations. The move is not necessarily bad. It just belongs to a different level of the game. Before adding another service, I would use the information already sitting inside the business. → What problems keep returning in leadership meetings? → What themes appear repeatedly in customer escalations? → Where are projects slipping? → Which agreements are creating activity without acceptable profitability? → Which internal standards are discussed regularly but never completed? Microsoft 365 Copilot can help surface those patterns across Teams and Outlook. Excel can then help test whether the story matches the operational numbers. The important part is not the technology. It is having enough honesty to accept the answer. Perhaps the business does not need a managed Copilot service yet. Perhaps it needs cleaner SharePoint permissions. Perhaps it does not need another engineer. Perhaps it needs to understand agreement profitability. Perhaps it does not need a new marketing campaign. Perhaps it needs to fix onboarding so new customers stop creating the same problems. That work rarely looks impressive on LinkedIn. It does, however, create real progress. A decision gets made. A blocker is removed. A service standard improves. The team can see movement again. That is the argument behind my latest post: https://lnkd.in/gnpXxBPE I’d be interested in practical answers to these: 1. What move is your MSP considering because everyone else appears to be making it? 2. What evidence would tell you whether that move fits your current business? 3. Where would fixing the basics create more value than launching something new? The smartest-looking move is not always the right one. The right move is the one your business is ready to absorb. #MSP #Microsoft365 #Copilot #ManagedServices
-
Robert Crane shared thisLowering the bar does not fix a broken process. It often just makes poor work easier to defend. MSPs sometimes soften standards because documentation takes too long, tickets are inconsistent, handovers are incomplete or people complain that the process creates friction. But removing the standard does not remove the problem. It removes the shared definition of acceptable work. Then the organisation starts operating with several competing versions of “good enough”: → One engineer documents every client change. Another writes three vague lines in a ticket. → One account manager records agreed actions. Another relies on memory and an inbox search weeks later. → One person follows peer review for a sensitive change. Another skips it because the client needed the work quickly. The issue is not necessarily bad people. The issue is a business that has allowed different definitions of acceptable work to exist at the same time. That is also why Copilot cannot repair an unclear operating model. Copilot can help summarise a Teams meeting, prepare a client follow-up in Outlook or structure project notes in Word. It cannot decide what a proper handover should contain if the business has never agreed on one. It cannot clean up SharePoint permissions simply because someone wants to “ask Copilot”. It cannot create accountability where ownership was never assigned. AI makes good standards more valuable because it can help people apply those standards more consistently. It also makes missing standards more visible. My latest post explains why clear expectations protect the team, the client and the business: https://lnkd.in/gx8yrYvD Standards should not exist to make work unnecessarily difficult. They should tell people: “This is how we do things here.” A few questions for MSP owners and technology leaders: 1. Which process in your business currently has too many definitions of “good enough”? 2. When people resist a standard, do you lower it or make it easier to follow? 3. What has Copilot exposed about the consistency of your existing processes? Sometimes the kindest thing you can give a team is a clear bar. #MSP #Microsoft365 #Copilot #ServiceDelivery #LeadershipStandards Are Not Gatekeeping. They Are Protection.Standards Are Not Gatekeeping. They Are Protection.
-
Robert Crane shared thisThe MSP business model isn't dying. But parts of it are becoming obsolete. AI, automation, security, and changing customer expectations are reshaping what clients value and what they're willing to pay for. I've spent months analysing industry research, market data, Microsoft strategy, valuation trends and emerging technology shifts. The result? MSP 2030. A practical report designed to help MSPs: ✅ Increase profitability ✅ Strengthen security offerings ✅ Build recurring growth engines ✅ Adapt to AI disruption ✅ Create long-term competitive advantage The next decade will create huge winners. And some businesses won't make the transition. Which side of the equation will you be on? 📘 Read the report: MSP 2030 - https://lnkd.in/g_5kK2xV #MSP #AI #CyberSecurity #BusinessStrategy #ManagedServices
-
Robert Crane shared thisYour “Book a Demo” button is not a trust strategy. It is easy to believe that an MSP needs more leads, more ads, more landing pages and more sales activity. But visibility is not always the real problem. The problem may be that people have not yet been given a reason to care. A prospect who does not understand the risk is unlikely to value the solution. A prospect who does not recognise your name is unlikely to trust your recommendation. A prospect who has never received anything useful from you has little reason to give you their time. Yet many businesses immediately ask for commercial action. Book. Buy. Call. Subscribe. Request a proposal. That is a large request to make before providing any value. A better approach is to earn attention before trying to convert it. For an MSP, that could mean: → Turning a repeated support question into a short article → Explaining Conditional Access without unnecessary jargon → Publishing a checklist clients can use immediately → Sharing an honest lesson from a difficult implementation → Helping business owners understand where Copilot will and will not help → Showing the thinking behind your recommendation, not merely the product Some MSPs worry that sharing useful knowledge will reduce the need for their services. I believe the opposite is usually true. When a client reads a clear explanation and discovers five issues they had not considered, they do not conclude that expertise is unnecessary. They begin to understand why it matters. I’ve written more about that here: https://lnkd.in/gXWt-xZG I’d be interested in specific answers to these: 1. What does your MSP provide before asking a prospect to book a meeting? 2. Are you withholding useful knowledge because you fear giving away too much? 3. What is one client question that deserves a public, plain-English answer? Trust is not created by shouting louder at the end of the buying journey. It is built quietly, well before the proposal. #MSP #Microsoft365 #ContentMarketing #TrustedAdvisor
-
Robert Crane shared thisCopilot does not deserve a licence just because it can summarise an email. It has to earn its seat. The more I speak with SMBs about Copilot, the more I think the price objection is often covering a more important concern: “I cannot yet see how this changes the way we work.” That is not really a licensing problem. It is a business-case problem. Showing someone that Copilot can draft in Word, summarise in Outlook or recap a Teams meeting might generate interest. But a collection of interesting demonstrations is not a reason to license an entire business. The conversation becomes far more useful when it starts with one annoying, expensive and repeated problem: → Proposals that consume half a day → Weekly reports assembled manually every Friday → Service desk documentation that is always postponed → Client follow-ups buried across email and meetings → Internal knowledge that exists in SharePoint but remains difficult to find Give Copilot a job inside one of those workflows. Use it with the people closest to the problem. Use real business information and real work. Then measure what actually changed. Did the proposal get completed sooner? Did the owner recover time from the inbox? Did meeting actions become easier to capture and follow? Did service desk notes become more consistent? For SMBs, capacity is a better conversation than generic productivity. Can the same team complete more useful work without adding more people? That is why I favour focused pilots over broad licence rollouts. Start small. Pick the problem. Measure the movement. If the outcome is obvious, expanding Copilot becomes a much easier decision. If the outcome is not obvious, buying more licences will not fix it. Read the full article: https://lnkd.in/g9TtFDjt A few questions I’d genuinely like your take on: 1. What specific job would you give Copilot first in an SMB? 2. What evidence would convince you that a pilot had earned an expansion? 3. Are we spending too much time demonstrating features and not enough time identifying business friction? Copilot should not be handed a seat. It should prove that it belongs there. #Copilot #Microsoft365 #MSP #SMB #AI
-
Robert Crane shared thisThe people who look lucky are usually the people who were ready. We see someone become an authority on Copilot, start speaking at events, attract new clients or build a strong reputation. From the outside, it can look like they were simply in the right place at the right time. What we don’t see is everything that came before it. → The early morning study sessions → The testing in trial tenants → The failed experiments → The blog posts hardly anyone read → The presentations delivered to small audiences → The relationships built before they were needed When Microsoft 365 Copilot became a major opportunity, the people who had already invested hundreds of hours understanding the technology appeared lucky. They weren’t predicting the future perfectly. They had simply done the work before the opportunity arrived. The same pattern applies to business development. A LinkedIn post starts a conversation. The conversation leads to a meeting. The meeting becomes a project. The project creates a customer reference. The reference opens another door. No individual step looks life-changing. But repeated consistently, those small steps create momentum. It is much like using Copilot in Outlook. You may not suddenly save ten hours in a day. Instead, you save a few minutes writing an email, summarising a meeting or finding information. Those gains compound. So do learning, writing, experimenting, building relationships and sharing what you know. In my experience, luck is rarely something that simply arrives. It is something you quietly create by becoming ready before the opportunity appears. I’ve explored this idea in more detail here: https://lnkd.in/gK6b7fNg A few questions I’d genuinely like your take on: 1. What small activity has created an unexpected opportunity for you? 2. Where have you seen years of preparation mistaken for overnight success? 3. What are you working on now that may not deliver an immediate payoff? The harder you work, the more opportunities you recognise. The more opportunities you recognise, the luckier you appear. #MSP #Microsoft365 #Copilot #BusinessGrowth #ContinuousLearning
-
Robert Crane shared this“We already have AI.” That is no longer the end of the Copilot conversation. It is where the real conversation starts. Usually, it means people are already using a standalone AI tool to rewrite emails, summarise content, clean up notes or generate ideas. It works. It may cost less. So why pay more for Microsoft 365 Copilot? Because comparing licence prices alone misses the most expensive part of the process: the work people must do before the AI can provide a useful answer. If someone has to open Outlook, search Teams, find a document in SharePoint, check a spreadsheet, remove sensitive information, paste everything into another tool and then move the answer back into Microsoft 365, that friction has a cost. It just does not appear on the AI licence invoice. The better questions are: → How much preparation is required before the AI becomes useful? → How much time is spent finding and assembling context? → What business information is being copied into external tools? → Can the organisation govern how that information is accessed and used? → Does the resulting output remain connected to the actual flow of work? This is where Microsoft 365 Copilot needs to be assessed differently. Its value is not merely that it can generate text. Many tools can do that. Its value comes from working closer to the emails, meetings, files, conversations, decisions and commitments already held within Microsoft 365. That does not mean every employee needs a full Copilot licence from day one. Start with roles where context matters and where people spend too much time stitching information together: → Client service → Sales preparation → Project coordination → Internal reporting → Executive briefings Then measure the improvement in the work, not how impressive the first AI response looks. A cheaper standalone AI tool may still be the right tool for some jobs. I use different tools for different purposes. But the buying decision should compare the complete workflow, including the searching, copying, checking, retyping and second-guessing. I explore that distinction in more detail here: https://lnkd.in/g44CXSuc A few questions I’d genuinely like your take on: 1. When customers say they already have AI, what are they actually using it to do? 2. Are businesses measuring AI licence cost, or the total cost of the workflow around it? 3. Which job or process would you test first before considering a wider Copilot rollout? The real decision is not which AI tool produces the cleverest answer. It is which one improves the work. #MSP #Microsoft365 #Copilot #AIThe Real Copilot Conversation Starts After Someone Says “We Already Have AI”The Real Copilot Conversation Starts After Someone Says “We Already Have AI”
-
Robert Crane shared thisThe biggest Copilot risk is not that the technology fails. It is that the technology works, the licences remain assigned, and everyone quietly goes back to working the old way. I’ve seen this pattern with plenty of business systems. There is a launch email, a demonstration and perhaps a recording. People experiment for a week. Then the initial interest fades while the invoices continue arriving. That is the adoption problem organisations need to solve with Microsoft 365 Copilot. Access is not adoption. Giving somebody a licence does not mean they know: → When Copilot will genuinely help them → What they should ask it to do → How to judge the result → What information should not be included → How it fits into the work they already perform Most people will not complain when they cannot see the value. They will simply stop using it. The better approach is to start with the work, not the product. A salesperson needs to see how Copilot can summarise a long customer email thread before a call. A manager needs to see how it can extract decisions and actions from a Teams meeting. An employee facing a difficult message needs to see how it can help produce a clearer first draft. One useful moment creates interest. Several useful moments start creating a habit. Managers also matter far more than launch emails. If leaders use Copilot openly, share where it helped and connect it to real business outcomes, the team receives a clear signal that this is part of how work gets done. IT can enable and secure Copilot, but it cannot create adoption on its own. I’ve expanded on that here: https://lnkd.in/gNHZUrkJ A Copilot rollout needs a rhythm: short training, practical examples, internal champions, follow-up sessions, clear boundaries and regular review. A few questions I’d genuinely like your take on: 1. What practical Copilot scenario has produced the quickest win inside your organisation? 2. Are managers actively modelling Copilot use, or has adoption been left largely to IT? 3. How are you identifying licences that have been assigned but are not producing meaningful value? The technology enables the opportunity. Behaviour change creates the return. #Copilot #Microsoft365 #MSP #AIAdoption
Robert Crane replied to a comment
5h
Ben Hooker you get what you pay for in short.
Robert Crane replied to a comment
1w
Exactly. Without a baseline, it's difficult to prove improvement. AI can generate a lot of activity, but unless you're measuring outcomes such as time saved, revenue gained, costs reduced, or risks mitigated, it's hard to demonstrate genuine ROI.
Robert Crane replied to a comment
1w
I think that's often true. The newest model usually gets the attention, but not necessarily the job. Stability, cost, speed and consistency often matter more than benchmark scores. Opus 5 might still be the right choice in some scenarios, but "newer" doesn't automatically mean "better".
Experience & Education
Licenses & Certifications
-
Administering Office 365 for Small Businesses
Microsoft
-
Microsoft Certified Professional
Microsoft
-
Microsoft Certified Sharepoint Professional
Microsoft
-
Microsoft Office 365 Administration
Microsoft
-
Microsoft SharePoint 2010, Configuring
Microsoft
-
Microsoft Small Business Specialist 2003
Microsoft
-
Microsoft Small Business Specialist 2008
Microsoft
-
Microsoft Small Business Specialist 2011
Microsoft
Publications
-
Cybersecurity Verification Playbook
Gumroad
See publicationEmpower your business with confidence—verify your Microsoft 365 security like a pro!
Unlock the essential guide for small and medium businesses to ensure your Microsoft 365 Business Premium environment is truly secure. This step-by-step playbook is designed for business owners and non-technical users, providing easy, actionable checks to verify that your Managed Service Provider (MSP) has correctly configured critical security features. -
How to Effectively Use Microsoft 365 Copilot
Gumroad
See publicationHow to Effectively Use Microsoft 365 Copilot: A Step-by-Step Guide for Small Businesses
Unlock the full potential of Microsoft 365 Copilot with this practical, easy-to-follow guide designed specifically for small business owners and operators. This document provides:
Clear, actionable instructions for using Copilot in Word, Excel, Outlook, and Teams, with real-world examples and tips for marketing, finance, operations, and customer service1.
Best practices for prompting…How to Effectively Use Microsoft 365 Copilot: A Step-by-Step Guide for Small Businesses
Unlock the full potential of Microsoft 365 Copilot with this practical, easy-to-follow guide designed specifically for small business owners and operators. This document provides:
Clear, actionable instructions for using Copilot in Word, Excel, Outlook, and Teams, with real-world examples and tips for marketing, finance, operations, and customer service1.
Best practices for prompting Copilot to get the most accurate and helpful results, including how to structure requests, provide context, and refine outputs for your business needs1.
Step-by-step walkthroughs for drafting documents, analysing data, handling emails, and summarising meetings—making everyday tasks faster and easier1.
Templates and prompt examples for common business scenarios, so you can start using Copilot immediately, even if you’re new to AI tools1.
Australian English and small business focus, ensuring the advice is relevant and accessible for local businesses1.
Whether you want to save time, boost productivity, or empower your team to work smarter, this guide will show you how to make Copilot your virtual assistant—helping you create content, analyse data, communicate professionally, and collaborate more effectively.
Empower your business with AI—purchase this guide and start transforming your daily operations with Microsoft 365 Copilot today! -
MSP AI Playbooks
Gumroad
See publicationWhy Choose This Bundle?
Future-Proof Your MSP or SMB: Stay ahead of the curve by mastering AI adoption, Copilot integration, and security best practices.
Drive New Revenue Streams: Package and sell AI-powered services with confidence, backed by expert guidance and proven frameworks.
Boost Client Trust & Satisfaction: Deliver Copilot solutions that are not only innovative but also secure, compliant, and tailored to real business needs.
Honors & Awards
-
SMBiT Professionals Honorary Member
-
This month we awarded longtime Valued Contributor Robert Crane with an Honorary Membership to SMBiT Professionals in recognition of his outstanding contribution to the SMBiT Professionals community.
Recommendations received
90 people have recommended Robert
Join now to viewOther similar profiles
Explore more posts
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore MoreOthers named Robert Crane in Australia
17 others named Robert Crane in Australia are on LinkedIn
See others named Robert Crane