# Brian Fontanella

Brian Fontanella shares practical guidance for small businesses on using AI and managing cybersecurity risks, including advice on Microsoft 365, AI tools, and managed IT.

- Profile: https://b2binfluencer.marketing/creators/brian-fontanella/
- LinkedIn: https://www.linkedin.com/in/bfontanella
- Median of last 8 posts: 3 reactions, 0 comments (checked 2026-11-02)
- Topics: artificial-intelligence, leadership, cybersecurity, ai-agents, ai-adoption, business-automation
- Based on 11 dated posts, latest 2026-09-24

## Summary

Brian Fontanella is most useful for owner-level AI and cybersecurity guidance for small businesses.

- Best for: Managed IT or security brands wanting thought leadership on AI and cyber risk for small businesses.
- Not for: Enterprise software vendors needing architecture-level, long-form technical content for CIOs and platform engineers.
- Check first: Ask for recent LinkedIn audience demographics and performance for comparable AI and cybersecurity topics.

### Takeaways

- Before enabling Copilot, review who can already see shared files.
- An AI agent deleted a production database in nine seconds.
- Cyber insurance renewal requires evidence that security controls actually exist.

## Main ideas

- **AI rollouts need human oversight before speed**: A recurring priority is that AI agents should not be given unchecked access. Brian Fontanella argues for logging, independent controls, and a named person who can pause or stop a model mid-task. The reasoning is that AI can act faster and further than a person expects, so the ability to intervene matters more than the feature list.
- **Before turning on Copilot, review who can already see what**: The Microsoft 365 posts pair feature enthusiasm with an access-control warning. The argument is that AI search does not create new permissions, it makes existing access easier to stumble into, so the risk sits in old sharing settings and company-wide folders.
- **Cyber insurance renewal is an evidence exercise**: Rather than treating insurance as paperwork, Brian Fontanella argues that carriers now want proof of specific controls. The risk is answering yes on the application when the records do not exist, because the gap surfaces at claim time.
- **Budget AI like a utility, not a per-user subscription**: Brian Fontanella argues that usage-based AI billing changes software budgeting from a predictable per-seat cost to consumption that can surprise a company. The practical priority is to set a monthly cap, control who can turn on new agents, and review the usage report.
- **Start automation with the biggest weekly time loss**: The recurring method is to avoid a broad AI strategy and instead identify the one task where the business loses the most hours each week, then automate that. The argument is that a narrow first deployment produces a concrete win and a manageable scope.
- **Know what your team is putting into AI prompts**: Brian Fontanella treats prompt data handling as an owner-level decision. The argument is that most owners cannot say which AI tools their team used that day, so the practical move is to choose tools the business is comfortable with and write simple rules about what stays out of a prompt.
- **The AI workforce story is retraining and growth, not only cuts**: Brian Fontanella uses survey data to counter the simplest headline that AI replaces workers. The argument is that layoffs and retraining are happening at the same time, and the businesses pulling ahead use AI to help existing people do more meaningful work.
- **Skepticism about Copilot turning into measured hope**: An evolution in Brian Fontanella's Microsoft commentary is worth noting because it shows a willingness to revise a view. The earlier pattern felt like ambitious features arriving before they were ready, while the 2026 updates are described as different because agent mode works iteratively and remembers context.
- **Small-business cyber risk is concrete, and managed IT fills a capability gap**: The security posts repeatedly make risk tangible with statistics and incidents, then argue that many small teams lack the security operation to respond alone. A related point is that some risks arrive through vendor software, so patching is an ongoing coordination problem rather than a one-time fix.

## Work to read

- [Microsoft's OneDrive and SharePoint Copilot upgrade, with an access-control warning](https://www.linkedin.com/posts/bfontanella_onedrive-just-got-a-major-upgrade-new-copilot-activity-7520498954053074946-DAJZ)
- [An AI agent deleted a production database in nine seconds](https://www.linkedin.com/posts/bfontanella_i-violated-every-principle-i-was-given-activity-7455982278788370432-oabs)
- [Cyber insurance renewals as an evidence exercise](https://www.linkedin.com/posts/bfontanella_cyber-insurance-requirements-for-small-business-activity-7515471452930953217-8oSc)
- [AI budgeting: usage-based billing changes the forecast](https://www.linkedin.com/posts/bfontanella_microsofts-cloud-growth-accelerates-and-activity-7523061149454716928-luXZ)
- [KPMG CEO survey: AI job cuts and retraining at the same time](https://www.linkedin.com/posts/bfontanella_kpmgs-ceo-on-why-ai-safety-isnt-another-activity-7518675729354760192-HLjC)
- [Where to automate first: 58% of small businesses use AI agents](https://www.linkedin.com/posts/bfontanella_ai-agents-for-small-businesses-how-to-add-activity-7452355578733068289-S2W1)
- [Nadella's emergency brake for AI](https://www.linkedin.com/posts/bfontanella_microsofts-nadella-says-ai-needs-an-emergency-activity-7516629087667257344-0jh5)
- [What is new in Microsoft 365 Copilot, with an honest caveat](https://www.linkedin.com/posts/bfontanella_whats-new-in-microsoft-365-copilot-march-activity-7447999261918597120-Ok98)
- [PHP2Shell and the websites you do not control](https://www.linkedin.com/posts/bfontanella_websites-across-world-hit-by-millions-of-activity-7519127077628055552-trt0)

## Brand relationships

- **Microsoft** (mention)

## Partnerships

No sponsorship or partnership program is documented in the material reviewed for Brian Fontanella. The reviewed items are Brian Fontanella's own LinkedIn posts on AI adoption, Microsoft 365 changes, cyber insurance requirements, and small-business security, plus one hiring notice. None is labeled as a sponsored placement, names a paying partner, or sets out deliverables, rates, approval rules, or an inquiry process. That absence reflects the reviewed material, not a conclusion that Brian Fontanella never accepts sponsorships. For a thought leadership partnership on AI and cybersecurity for small businesses, the practical implication is that there is no published package to buy into. A buyer would need direct contact and a from-scratch agreement covering scope, disclosure language, review and approval steps, and how sponsored content would be distinguished from the commentary Brian Fontanella already publishes. The closest fit in the existing work is co-authored or clearly sponsored explainer content on subjects Brian Fontanella already covers, such as AI guardrails, Microsoft 365 Copilot rollout, or cyber insurance evidence. Any such arrangement would be original, and no published terms exist to plan against.

## Alternatives

- Nick Machol: Choose Nick Machol when the collaboration's core is AI as an operating and workforce decision rather than a security one. The reviewed posts argue that AI is moving from assisting staff to doing the work itself — closing books, drafting contracts, handling claims — and that the durable move is pairing automation with training so people build the judgment AI cannot replace. A managed IT brand whose campaign angle is "automate the work and train the people at the same time" gets a sharper thesis here than from a tool-rollout explainer, and both reviewed posts are recent, from March and July 2026, which suits a current-cadence series.
- Ali Ahad: Choose Ali Ahad when the campaign is about automating a specific process rather than a category of risk. The clearest reviewed post argues against "just replace the spreadsheet" and instead keeps the system a team already understands while automating the copy-and-paste around it; a second works through one industry's compliance duplication, where the same certificates sit in three systems and drift out of sync, leaving a compliant landlord looking non-compliant. That is directly useful for a managed IT brand selling workflow automation or a vertical offer, and it also demonstrates a build-oriented approach with a named automation tool, which matters if the collaboration is meant to show a working build rather than only an opinion.
- Phil Beavan: Choose Phil Beavan when the campaign is cybersecurity-first for professional-services firms. The reviewed posts draw the line a managed security buyer needs: insurance is a parachute, not a shield, and the reactive cover has to sit on top of 24/7 monitoring, penetration testing, phishing simulation, device and identity management, and certification. A second post supplies the urgency case with ransomware numbers — nearly half of businesses pay, only 7% of those that pay recover their data in full, and 76% are targeted again. For a brief whose deliverable is a security-hardening campaign rather than an AI-adoption one, this is the closest subject match of the three, because the reader is asked to buy layered defence.

---
AI-assisted summary of public posts. Audience demographics are not available. Engagement appears only where we checked recent posts. Method: https://b2binfluencer.marketing/methodology/