Summer ’26 is live. And July’s security deadlines are already here.
The new release is in your org, and the first hard enforcement dates land this month. A practical checklist for the org you already run: what to lock down now, the data work that makes AI actually work, and two quiet Summer ’26 features worth switching on today. Plus a new CirriusTalk episode.
02 AI-Ready Data
03 Field Access
04 Reporting Wins
July’s Security Deadlines Are Already Landing
Three Summer ’26 security enforcements reach production this month, and every live org is affected. A quick sandbox check now beats a user lockout later.
- Phishing-resistant MFA for admins, plus step-up auth for report exports (July 1)
- Transaction-security limits on large record exports (July 13)
- Org-wide multi-factor authentication (July 20)
- Unconfigured step-up policies block users from exporting reports
- Admins without enrolled MFA methods lose privileged actions
- Dates follow Salesforce’s roadmap. Verify against your schedule.
Free to Ingest Isn’t the Same as AI-Ready
Ingesting your core-cloud data into Data 360 no longer burns credits. Genuinely good news, but pouring in more data doesn’t make it AI-ready.
- Duplicates and half-filled records can lower your overall data quality
- An agent grounded on messy data returns wrong answers that look right
- The real work is unchanged: dedup, identity resolution, and modeling
One Screen for Total Field-Access Clarity
Summer ’26 adds a Field Access Summary (Object Manager → Field Access): one read-only view showing exactly where a field’s access comes from, across profiles, permission sets, and permission set groups.
No more chasing “why can this user edit this?” across scattered settings. It’s also a natural on-ramp to a least-privilege access review.
Reporting Just Got a Quiet Upgrade
Three low-effort reporting wins are live now, usable by ops and business stakeholders, not just admins:
- Brand color palettes so dashboards match your brand
- Up to two row-level formulas per report (no more duplicate reports for one calculation)
- Embeddable Lightning components in dashboards (GA)
AI agent strategy: seven deliverables to lock in before development
This CirriusTalk episode walks through the seven deliverables to lock in before you build an agent, so scope, data, and guardrails are settled up front instead of discovered mid-build. A natural companion to this issue’s readiness theme: get the foundation right, then build.
Secure it, ready it, use it.
The thread through all of this: Summer ’26 rewards orgs that tend what they already have. Lock down access before the July dates, get your data honest so AI can trust it, and switch on the features that hand time back to your team. Not more tools, better use of the ones you own.
Beat the July Enforcement Dates
Not sure where your org stands on MFA coverage, step-up policies, or permission-set readiness? We are offering managed services clients a complimentary Security Health Check to validate your configuration before the enforcements bite. Talk to your Cirrius Consultant to align it with your 2026 roadmap.
*Enforcement dates follow Salesforce’s published security roadmap and may vary by instance; verify against your org’s release schedule.
