Latest from LinkedIn

Integrations shouldn’t take weeks.

A design and engineering consultancy was retyping every won job from its quoting tool and CRM into its delivery system: client details, fees, billable stages, project numbers, job folders. Nothing could be booked or invoiced until someone did it, and later changes never made it across.

We connected the systems they already had, with every flow following the same pattern:

1. Watch for an accepted quote or a won deal
2. Resolve whether the job already exists, and stop and report if it's unclear
3. Reconcile: create what's missing, update only what changed
4. Record every outcome

Now won jobs appear in the delivery system on their own, with numbers, stages, fees and folders in place, and staff time has moved from re-entering to a quick check.

Delivered in 5 work days. A simple integration, saving hours of work.

Read the full case study →

Most businesses don't have a data problem. They have a "too many systems that don't talk to each other" problem.

CRM in one place, finance in another, HR somewhere else. So people re-key, export, import and reconcile, and every team ends up with its own version of the numbers.

We connect them:

→ Ingest data from any source (CRM, ERP, HRIS, e-commerce, finance, operations) into one place

→ Sync master data, so when a customer changes in Salesforce, NetSuite and Shopify follow automatically

→ App-to-app integration, so your applications work together rather than just feeding a report

Each dataset gets one primary source, changes flow to everything else, and we reuse the systems you already have rather than replacing them.

Struggling with disconnected systems? Let's talk about where your data lives today. Talk to us →

BI or AI? Wrong question.

BI shows where you stand. AI finds the why. They work best together, each doing what it's good at.

Most organisations have the opposite problem: dozens of dashboards, drill-downs nobody uses, and the answer to any real question still three clicks deep.

Our approach:

→ A small number of well-designed views that tell you where you stand at a glance

→ AI behind them doing the harder work: finding the patterns and correlations, explaining the movement, and suggesting what to do next

Analysis working on your behalf, rather than another dashboard to go and read.

See it working in the Insyte Experience →

The biggest time saver from AI agents isn't the question you ask. It's the work that happens without anyone asking.

Take the monthly sales and marketing analysis report. It spans commercial apps and our own in-house application: the pipeline from HubSpot, website traffic from Google Analytics, LinkedIn performance, and access requests from the Insyte Experience, which we built ourselves. Insyte Agents pulls it all together, analyses what's driving it, creates a PDF in our standard template, emails it to the leadership team and lets them know on Slack. Nobody has to ask.

That's the important part. Agents that only work inside one vendor's product can only automate that product. Your processes run across off-the-shelf SaaS and the applications you've built in-house, so your automations need to as well.

Personal prompt libraries don't get you there alone. They live with one person, run when that person runs them, and leave when they do.

With Insyte Agents, an automation is defined once, owned centrally and put to work:

→ Build and test: describe it in your own words and AI shapes it into a workflow that understands the agents, so the instruction is just right. Test individual agents until each response is right, then test the workflow as a whole

→ Approved before it runs: the agents it can use are locked in, approvers sign off, and anything not pre-approved is skipped and logged

→ Dry run first: see exactly who would be emailed, what would be posted, and which records would change, before anything is sent

→ Scheduled: it runs on its own, with the same permissions as a live request, so it can never reach something its owner couldn't

→ Every run audited and costed: so one that quietly gets expensive shows up long before the invoice does

Built on Microsoft Agent Framework and Foundry, across your commercial apps, your in-house applications and the channels your team already uses.

We went from prototype to a platform the whole business uses in weeks, and it’s now an accelerator we can use with our clients. Read how it came together, and the six decisions behind it, in From One Agent to an AI Platform the Whole Business Can Depend On →

Data, reporting and AI were all key topics at last week's Together Vision Technology User Group.

Melanie Poonai from the InsyteGroup team co-presented the Reporting & BI session with Ben Marriott from Together Vision Technology, and three things kept coming up in the room and in the conversations afterwards:

1. Excel is often still the last mile.

Many people regularly export data from Business Central into Excel so they can add logic the standard reports don't include, such as custom tax and duty calculations. That logic could live in the report itself and run automatically on live data, instead of being reapplied by hand every month.

1. Your data isn't only in your ERP.

Logistics and transport data, marketing and HR systems, customer conversations tracked in Outlook. The fuller picture comes from having that data alongside Business Central, rather than looking at each in isolation.

1. AI is only as good as the data behind it.

Agents and plain-English questions are exciting, but they only work well when the data underneath is connected, clean and modelled properly. BI and AI aren't separate projects. Good BI is what makes practical AI possible.

The new Business Central mirroring to Fabric got a lot of attention on the day, and for good reason. It removes the need to export in the first place, gives you a natural place to bring other data alongside Business Central, and lays the foundation that both BI and AI depend on.

Thanks again to Together Vision Technology and Bevica for a great day.

"Don't do anything different. You've clearly figured out what great customer experience and execution are."

That's from a new Clutch review by the COO of a field service and construction software company.

What they said impressed them most: "Sense of urgency. Responsiveness. Executes with extreme accuracy. Measures twice, cuts once. Collaborative. Innovative with AI best practices."

And the line that means the most to the team: "InsyteGroup has never missed a deadline and takes pride in responding to us as soon as possible."

Thank you for the trust, and for taking the time to write it up.

Read the full review on Clutch →

Integrations shouldn’t take months.

One flow or enterprise scale, the same principles decide whether an integration can be trusted. We’ve built both:

→ One flow: new deals and won jobs flowing straight from CRM into delivery. All projects, all systems, no retyping

→ Enterprise scale: 15,000 sales and inventory movements a day, with finance, ERP, POS and online in step and fully reconciled

What makes it quick is that we don’t start from scratch. Our integration accelerators (connectors, sync patterns, transformation and monitoring) are ready from day one, so the effort goes on what’s unique to your business.

Two recent integrations each went live in about five days.

Our lead engineer, Nazarii Kalichynskyi, has written up the nine principles we build by: One flow or enterprise scale: the integration principles that hold either way →

Thank you to everyone who came to the Together Vision Technology User Group 2026 at Microsoft, Paddington yesterday.

It was great to meet so many Business Central users, both on our stand and in the session, and to hear what you’re working on. The conversations about getting more from your data, and where AI fits in, were the highlight of the day for us.

A big thank you to Together Vision Technology and Bevica for having us, and to everyone who stopped by for a chat (and a chocolate).

If you didn’t manage to scan the QR code but would like to try the reporting and AI tools we demonstrated for yourself, here’s the link →

And if we didn’t get to talk on the day, or you’d like to pick up where we left off, we’d love to hear from you.

Ten years ago this month, we started InsyteGroup.

We began as a management consultancy. To run our first client project, we built a Power BI data room. The client liked it so much that they asked for it in their own business.

Overnight, we were a technology provider.

We went all in on technology in 2020. For the last twelve months, the focus has been expanding our capabilities with AI, building alongside our clients to solve problems in new ways and make possible what wasn’t before.

The pattern has been the same all the way through. Every capability we have, we added because a client needed it. Either to fix something that was costing them, or to reach an opportunity they could see but couldn’t get to.

Today that's five things:

• Data and application integration
• Data architecture and engineering
• Dashboards and reporting in Power BI
• AI-powered insights and analysis
• Enterprise AI agents, on Microsoft Agent Framework and Foundry

Thank you to every client who trusted us with a first project, every colleague who built something they're proud of, and everyone who has introduced us to someone else.

Here's to the next ten.

If you'd like to see the best of what we're doing today, take a look at the Insyte Experience →

We’ll see you at the Together Vision Technology User Group 2026 at Microsoft, Paddington on 15 October. Thanks to Together Vision Technology for having us.

It’s a day of Business Central, AI and automation. On our stand we’ll be talking about Fabric powered BI & AI on Business Central: your data landed and modelled once in Microsoft Fabric, then used both by Power BI reports and by AI agents that can explain what’s behind the numbers.

We’ll have the experience we’ve built for Bevica running live, with Power BI dashboards and AI agents to play with.

If you’re going, come and find us.

"Which stores missed target last month, and why?"

In most businesses, that question means a ticket to the BI team and a wait for a report.

In the Insyte Experience, you just ask.

The answer doesn't come from one place. Our agents query your database and semantic models directly, pull from curated 360s, search your unstructured data such as notes and documents, and an analysis agent brings it together into an answer with the why, not just the what.

And you can see exactly where each part came from.

Try it on worked examples across sales, retail, services and finance →

Most organisations don't need another dashboard.

They have dozens already. Drill-downs nobody uses. And the answer to any real question is still three clicks deep.

We think BI and AI should work together, not compete. A few well-designed views that show where you stand at a glance, with AI doing the harder work behind them: finding the patterns, and getting to why the numbers are what they are.

So rather than send a deck about it, we've opened up what we've built. The Insyte Experience:

→ InsytesPortal: embedded Power BI, bespoke dashboards and natural-language agents in one place, running on your own Fabric capacity, so reporting reaches far more people without a licence for every viewer.

→ InsytesAgents: agent workflows built on Microsoft Foundry and the Agent Framework, running in your own Azure tenant. Not another SaaS subscription. Reaches your non-Microsoft systems too, and comes with its own MCP server, so it works from Copilot, Teams or Claude.

It's the same portal we put in front of clients, with worked examples across sales, retail, services and finance. Ask one of the agents a question and watch where the answer comes from. It goes beyond natural-language queries direct to your database and semantic models, to curated 360s, RAG search over your unstructured data, and analysis agents that bring it all together into comprehensive reports and recommendations.

Explore it here →

Hit the ceiling with Copilot Studio? Talk to us and we'll walk you through a live workflow doing real work.