September 2026 Summaries

Tines’ finance team built 3Billing, an internal application on its Tines 3B platform, to automate the flow of deal, billing, and revenue-recognition data between Salesforce and NetSuite, replacing a previously manual process that required staff to create customers, sales orders, line items, and invoices individually. After evaluating third-party and native NetSuite tools estimated at roughly $60,000 annually, with one option including an additional $60,000 implementation fee, the team rebuilt an earlier workflow and reported spending less than $1,000 in AI token costs to build and test the app, with minimal ongoing costs. The application processes every company deal, supports IFRS 15 revenue-recognition requirements, includes validation, reconciliation, audit logging, and controls for business-specific edge cases, and can be updated internally as processes change. Tines reports that 3Billing has reduced time spent on this billing work by about 75%, enabled the finance team to scale without additional headcount, reduced error risk, and improved audit readiness, while avoiding vendor pricing tied to deal volume or revenue.

AI governance monitoring focuses on proving that approvals, restrictions, and accountability controls actually operated during specific AI actions, distinguishing it from model monitoring, which evaluates accuracy, errors, and drift. The approach relies on runtime evidence from identity, OAuth, cloud audit, network, approval, and revocation logs to document the acting agent, tools and data accessed, control decision, timestamp, and resulting outcome. It addresses control drift, shadow AI adoption, outdated point-in-time assessments, and fragmented evidence by continuously tracking event-level records, program health metrics, and reassessment triggers such as new data sources, model versions, vendors, or deployment regions. Effective monitoring requires shared records across security, IT, business, compliance, and legal teams, while assigning accountable owners for individual AI applications. The discussion also notes that frameworks such as NIST AI RMF and ISO/IEC 42001 distinguish governance from model measurement, and that EU AI Act requirements for logging, traceability, and post-market monitoring increase the need for durable operational evidence.

Tines co-founder and COO Thomas Kinsella discussed the company’s origins, the growing governance challenges posed by “wild code,” and how organizations can combine deterministic automation, AI agents, and human judgment in workflows. Founded in 2018 after Kinsella and co-founder Eoin Hinchy found existing security automation tools too complex, Tines aims to help teams build, operate, and monitor critical workflows across security and IT. Kinsella described wild code as ungoverned scripts, agents, and automations created throughout enterprises without centralized oversight, and presented Tines 3B, launched in July 2026, as a platform intended to enable AI-driven building while maintaining organizational visibility and control. He argued that organizations should make governed tools the fastest and easiest option for users, and recommended deterministic automation for predictable tasks, AI agents for less-defined problems, and humans for consequential decisions requiring context. The discussion also emphasized monitoring workflows at scale, automatically identifying failures and proposing fixes, and converting duplicated, fragmented automations into reusable shared services rather than accumulating technical debt.

AI SOC platforms are portrayed as an increasingly commoditized cybersecurity category, as major SIEM vendors add native AI triage and numerous vendors compete with similar offerings, while faster alert processing can leave analysts facing a growing backlog of decisions. The text argues that security operations are shifting toward customizable agentic automation platforms that enable organizations to build workflows tailored to their own systems, processes, and governance requirements. It presents Tines 3B as a prompt-driven environment for creating automated security workflows, with capabilities such as alert investigation, contextual evidence aggregation, explainable triage, attack-path visualization, threat hunting, detection management aligned to MITRE ATT&CK, threat-intelligence ingestion, and automated containment actions. It also emphasizes controls for AI agents through sandboxed workflows, auditability, Retrieval-Augmented Generation using historical security knowledge, and Model Context Protocol integrations intended to let agents query internal resources without directly accessing sensitive credentials.

Tines has introduced updates aimed at expanding app development, case management, AI assistance, governance, and administrative controls. Apps can now access cases, records, and resources through Workbench functions, be exported as ZIP archives, have assigned owners, and require single sign-on for viewers, while self-hosted environments gain support for interactive Apps without external network access. Workbench is now embedded in cases with read-only “Ask” and action-oriented “Agent” modes, alongside manually triggered Tines 3B workflows, improved case list layouts, and API-based case template management. AI Agent actions now support reusable Agent Skills across Stories and Workbench, with Fable 5 and Opus 5 Claude models available, while visual change reviews, credential-level Workbench access controls, AI credit alerts, and identity-provider authentication for Collections strengthen collaboration and governance. Additional changes include date filtering for audit logs, customizable console banners for self-hosted deployments, and announcements for Builder Connect events and ongoing release information.

Tines reports that its AI-native building environment, Tines 3B, was introduced internally on June 2 ahead of its July public launch, enabling employees across nontechnical and technical teams to create apps, agents, and automations for tasks such as billing, recruitment, legal requests, sales operations, and security access management. The company says employees built more than 1,000 live 3B workflows in 90 days, compared with the more than seven years it took to reach that total with its existing Tines Stories workflow platform, with 67% of employees building in 3B and all employees using its outputs. Tines attributes the rapid adoption to AI-assisted creation and governance features including authorized connectors, continuous monitoring, SCIM-based provisioning, role-based access controls, and evolving workspace structures, while an internal hackathon accelerated experimentation and production use. Internal feedback also influenced product development, including an access-map feature first built by the company’s IT security leader. Tines says Stories will remain in use for established long-running workflows, while 3B increasingly handles new use cases and integrations; three months after rollout, 3B reportedly represented 20% of all workflows ever built at the company, alongside emerging considerations around AI costs, duplicate applications, and governance.

Tines has obtained TX-RAMP Level 2 certification for its Tines Stories and Tines 3B platforms, enabling Texas public-sector organizations to use them for workflows involving confidential data and moderate- or high-impact information resources. TX-RAMP, administered by the Texas Department of Information Resources, evaluates cloud services against standardized security requirements for state agencies. Tines Stories supports automation and orchestration across existing security and IT systems, while Tines 3B enables organizations to build, run, and monitor AI-generated work with governance and visibility controls. The certification reflects Tines’ broader public-sector strategy, which includes its security-focused product development and ongoing pursuit of FedRAMP authorization to support federal agencies.