Dominic Jainy has spent years weaving AI, machine learning, and blockchain into operational systems, and he’s candid about what actually moves the needle in Dynamics 365 Business Central. In this conversation, he explains how 2026 release wave 1 puts intelligent agents right where work happens, why finance automation and self-billed invoices change the rhythm of close, and how refreshed Power BI dashboards and new APIs lift trust in reporting. We cover day-to-day mechanics—guardrails, queues, and reconciliations—alongside sustainability data capture, localization, and Microsoft 365 workflows across Outlook, Excel, and Teams. Throughout, Dominic shares pragmatic steps, governance patterns, and change tactics that turn the promise of AI into measurable progress.
Big picture: Where will intelligent automation in ERP deliver the fastest wins this year, and what early missteps should teams avoid? Share a specific example with timelines, metrics, and the step-by-step approach that made it work.
The quickest wins land where Business Central 2026 release wave 1 puts agents directly on the pages your team already uses—think order intake and invoice creation—so the “new” feels like the “now.” The common early misstep is jumping to custom logic before exhausting the native agents and Copilot prompts, which are built to work end-to-end in one task pane and cut setup friction. A practical example is converting emailed POs into purchase invoices: we started with inbox rules, used the agent to parse key fields, and mapped them to vendor, item, and tax data in Business Central, then finished with a human review. The timeline hinged on progressive enablement: week one focused on agent configuration and email parsing, week two on page-level validation and posting policies, and week three on tightening exception rules and documenting approvals under the same 365 security roles.
AI agents now process tasks like creating purchase invoices from emails. How do you design guardrails, review steps, and exception handling so accuracy stays high? Walk through a real workflow, including error rates before/after and human-in-the-loop checkpoints.
Start by constraining scope to the vendors that already send consistent layouts and let the agent work only within one page and one document type at first. The human-in-the-loop happens twice: once in the task pane to verify header fields and once at posting to confirm taxes and dimensions; that mirrors how 365 roles already gate access. Exceptions trigger when mandatory fields are blank, when totals mismatch line sums, or when tax logic conflicts with advanced tax calculations; the agent routes these to a review queue. In practice, the workflow moves from email capture to agent draft, to header check, to line validation and tax cross-check, and finally to posting with audit notes stored via the new APIs for downstream review.
A task pane now centralizes agent activities. How should operations teams triage, prioritize, and audit agent queues day-to-day? Describe alert thresholds, KPIs you track, and how you tune workloads during peak periods.
Treat the task pane as your control tower: one view, three lenses—age, risk, and value. We triage by oldest-first for compliance-sensitive items, then highest-value documents, and finally low-risk routine drafts; that keeps auditors happy and cash moving. KPIs include items awaiting approval, exceptions by root cause (vendor data, taxes, or price mismatch), and cycle time from agent draft to post; these can roll to Power BI so leaders see trends in 365 with no extra clicks. During peaks, we tune by narrowing agent scope to the highest-accuracy vendor or customer profiles and pausing noncritical automations so the queue stays green instead of flooding reviewers.
Order-to-cash and procure-to-pay see speed gains from background data entry. What cycle-time reductions and DSO/DPO impacts have you seen, and what process changes unlocked them? Break it down by handoffs, approvals, and reconciliation.
The speed comes from deleting handoffs you don’t need and streamlining the ones you do: data entry moves to the background, leaving one approval at the right point rather than many. In order-to-cash, we eliminated the handoff between sales and finance by letting the agent enrich orders on the same Business Central page and alert approvers in Teams. In procure-to-pay, the approvals happen after the agent confirms vendor terms and advanced tax calculations, and reconciliation is auto-suggested from the same ledger view. The practical outcome is fewer touches, a cleaner audit trail via the new APIs, and a healthier rhythm for DSO and DPO because approvals aren’t waiting on manual keying.
Finance agents read documents and auto-fill payables and expenses. What data validation rules, tolerances, and cross-checks keep postings clean? Share a step-by-step setup and the defect rate trend after go-live.
Keep validation layered: header totals must equal the sum of lines, vendor and item master data must exist and be active, and tax logic must align with the advanced tax calculations already configured. Cross-checks include vendor bank data against approved lists and dimension rules against project or item policies, especially when environmental data is linked to items. Setup is straightforward: enable the agent, map document fields to Business Central columns, define tolerances for price and quantity, then lock posting to roles already defined in 365. After go-live, we watch defects by category in one Power BI view and tune mappings weekly, which steadily reduces rejects without adding custom code.
Self-billed invoices and advanced tax calculations simplify complex billing. Which industries benefit most, what edge cases cause trouble, and how do you test for multi-jurisdiction tax scenarios? Include sample configurations and reconciliation routines.
Self-billed invoices shine in manufacturing and distribution where the buyer sets the quantity and price at receipt, and advanced tax is vital for healthcare and cross-border supply chains. Edge cases appear when returns, rebates, or multi-rate rules collide, so we model those cases on a sandbox first with the same 2026 release wave 1 features. A sample configuration starts with tax groups for items and customers, jurisdiction rules, and invoice templates that include self-billing logic; then we bind those to workflows the agent can follow in one pass. Reconciliation means comparing posted entries to tax ledgers and ensuring the self-billed document numbers roll up cleanly for audit under the new workflow and permission APIs.
Period close is expected to move faster. How do you re-sequence close tasks, segregate duties, and automate reconciliations without losing control? Provide a close calendar example, target durations, and audit evidence you produce.
I move reconciliations earlier by letting agents draft entries as the month unfolds, so the last-day crunch narrows to approvals and analytics. Duties stay segregated by using 365 roles: one prepares, one reviews, and one posts, and the task pane keeps each step in plain view. An example calendar pins bank and subledger recs before management adjustments, and reserves the final window for disclosures and Power BI refresh validation. For evidence, we export workflow histories through the APIs, attach supporting docs from Outlook to the ledger entries, and save the dashboard refresh logs, giving auditors one coherent trail.
Financial reports and Power BI dashboards now refresh automatically. How do you align data models, security roles, and refresh cadences to keep trust high? Share a dashboard example, adoption metrics, and a remediation plan for data drift.
Start by mapping Business Central entities to a star schema that your team recognizes, then lock row-level security to the same 365 groups you use in the ERP. Refresh cadence mirrors operational tempo—near real time for cash and inventory and scheduled windows for period-end statements—so leaders open Power BI and see what they expect. A practical example is a cash and profitability dashboard that blends payables, receivables, and project costs, with drill-throughs to the underlying entries on the same day. When data drift shows up, we pause risky transformations, compare today’s refresh to a known-good snapshot, and use the new APIs to trace permission or workflow changes that might have shifted the numbers.
New APIs expose workflows and permissions for audit. How would you structure continuous controls monitoring, alerting, and evidence collection? Walk through an audit trail example and the queries or API calls behind it.
I frame continuous controls monitoring around three pillars: workflow state changes, permission assignments, and posting anomalies. Using the APIs, we pull every approval state change and map it to the user’s 365 role; anything outside policy lands in one alert queue. For evidence, we store the before-and-after workflow states with timestamps, and we attach the original document from Outlook so auditors can trace it end-to-end. A typical query fetches approval updates for high-value documents and correlates them with permission changes in the same window, closing the loop between intent and authority.
Teams increasingly embed environmental data into projects or items. How do you capture emissions factors, attribute them to costs, and report on ESG KPIs alongside financials? Provide a data model sketch and a monthly reporting cadence.
The model starts at the item or project level where we store emissions metrics as attributes and link them to cost entries so every posting carries both financial and environmental context. When agents create or update documents, they inherit those attributes, so reporting aligns naturally with ledgers and dimensions. In Power BI, we surface ESG KPIs next to profitability and inventory so leaders see both impact and cost in one dashboard. For cadence, we align monthly ESG reporting with the same refresh rhythm as financials, letting 365 security determine who can view or export the combined view.
Localization and compliance need rapid tweaks. What’s your playbook for configuring regional tax and reporting without custom code sprawl? Share a concrete change request, versioning approach, and rollback strategy.
My playbook favors configuration over code: use advanced tax calculations, localization packs, and feature flags in 2026 release wave 1 to contain change. A common request is adding a jurisdiction with special rates; we introduce a new tax group and rule, test in a sandbox, and document the workflow impact before promoting. Versioning happens in extensions with clear release notes and a tagged baseline, so rollback is just moving back to the previous package. We keep one change log that ties configuration changes to posted transactions, so auditors can trace why and when outcomes shifted.
Usability upgrades include personalized forms, resizable help panes, and smarter search. Which changes drive measurable productivity, and how do you quantify it? Offer before/after task timings, user feedback methods, and a quick wins checklist.
Personalized forms reduce hunting and clicking, and the resizable help pane keeps guidance in view so users don’t context-switch. We quantify by time-to-complete for frequent actions and by queue age in the one task pane; when search gets smarter, those metrics improve without extra training. Feedback starts with short in-app surveys and ends with Teams roundups that capture friction points within 365. Quick wins include saving views, pinning filters, trimming fields to what’s essential on a page, and bookmarking the help topics most used during close.
Microsoft 365 integration lets teams attach quotes from Outlook, analyze in Excel, and start Teams chats from records. What cross-app workflows save the most hours, and how do you secure data movement? Provide a step-by-step scenario with roles and permissions.
The biggest win is staying in the flow: a sales rep attaches a quote from Outlook, finance reviews it in Business Central, and a controller opens Excel to analyze margin without exporting data to a random share. Security flows through 365 groups, so the same roles that permit posting also gate who can see the attachment and who can launch a Teams chat from the record. Step-by-step, the rep sends the quote, the agent ingests it, the approver reviews in the task pane, and a Teams thread captures the decision with a link back to the record. Nothing leaves the authorized boundary, and the APIs keep an audit of who saw and did what in that chain.
When selecting an implementation partner, how do you evaluate industry expertise, custom development capability, and support SLAs? Share a scorecard you’ve used, the thresholds that matter, and an example of a tie-breaker insight.
I score partners on three axes: industry fit, extension craftsmanship, and operational support, all grounded in Business Central and Dynamics 365 experience. Industry fit means they can speak to manufacturing, distribution, or healthcare processes using the language of the 2026 release wave 1 features, not generic promises. Development capability shows in how they’ve built extensions that avoid technical debt, and support is evidenced by response patterns your team can trust. The tie-breaker is often their approach to training on Copilot and agents—if they bring playbooks and direct lines into Microsoft, like a Microsoft Solutions Partner would, they tend to help you realize value faster.
Customized extensions can unlock value but add risk. How do you govern extensions, manage lifecycle updates, and prevent technical debt? Describe branching strategies, testing gates, and a rollback story with lessons learned.
Governance starts with a catalog of extensions, why they exist, and which native 2026 wave 1 features might replace them later. Branching separates configuration from custom code, and testing gates include unit checks for calculations, page-level validations, and end-to-end posting cycles with agents active. Rollback is enabled by versioned packages and one clear baseline, so when a posting anomaly appeared in a new build, we reverted, inspected the workflow via APIs, and re-released with a fix. The lesson is simple: ship small, document changes, and keep a path back to stable.
Training on new Copilot features can make or break adoption. What curriculum, change management tactics, and reinforcement loops work best? Offer a 30-60-90 day plan with usage targets and coaching touchpoints.
I anchor training in real work: start with the common tasks agents can assist, then expand to edge cases, so users see value on day one. Change tactics include show-and-tell in Teams, quick reference cards in the resizable help pane, and champions who can translate features into the team’s language. Reinforcement loops blend weekly office hours and Power BI usage tracking, so leaders can nudge adoption where it lags inside 365. Over one quarter, we phase from guided demos to supervised use to confident autonomy, always tying back to the same pages and task pane users touch every day.
For organizations seeking fast ROI, where should they start—AI agents in AP, reporting automation, or Microsoft 365 workflows? Share a staged roadmap with cost ranges, payback periods, and the metrics you track to prove value.
Begin where friction is felt most: AP agents if invoice volume is heavy, reporting automation if leadership lacks timely visibility, or 365 workflows if context-switching burns hours. The roadmap rolls out in stages so you stabilize one area before adding the next, keeping everything inside Business Central’s native 2026 release wave 1 capabilities where possible. Proof comes from measuring queue age in the task pane, exception rates, and dashboard adoption in Power BI, all visible in one place. Because you’re staying close to standard, you avoid surprises and move toward ROI without speculative customization.
What is your forecast for AI-driven ERP over the next three years?
Over the next three years, AI in ERP becomes less of an add-on and more of the operating fabric, with agents embedded on the exact pages where people decide and act. Finance will feel it first with self-billed invoices, advanced tax, and automated close sequences that keep auditors satisfied through APIs and clear trails. Reporting will be expected to refresh itself, and ESG will sit beside profitability in one dashboard without a separate data chase. The most successful teams will pair disciplined governance with everyday usability—personalized forms, smarter search, and Microsoft 365-native collaboration—so the system feels both intelligent and familiar from day one.