Start a project
Menu
ServicesAll servicesCybersecurityCloud EngineeringApplication ServicesEnterprise PlatformsData & AnalyticsAI & Machine LearningInternet of ThingsMobilize Talent
PlatformsAll platformsIdentity and accessPrivileged access and secretsERP, HCM and financeCRM, ITSM and automationHealthcareInsurance and bankingData and integrationCloud and AIIndustrial, utilities and mainframeSkills matrix
Case studiesAll case studiesIdentity and accessPrivileged accessSecurity operationsCloud and infrastructureApplicationsERP and HCMCRM, ITSM and automationIndustry platformsData and analyticsAI and agents
IndustriesAll industriesEnergy, oil & gasHealthcare payersBanking & financial servicesInsurancePublic sectorManufacturingLocations
InsightsField notesLearning tracks
CompanyAbout CohashPortfolioSkills matrixCareersTrust and securityContact
Start a project
low-code & automation

Your makers keep building Power Platform apps, and IT can still govern them.

Citizen development turns into shadow IT when nobody sets the rules. We build Power Apps and Power Automate solutions and set up Copilot Studio agents. We also put environment, DLP and ALM rules in place.

Talk to a Microsoft Power Platform engineerWhat changed recently
platform dossierMicrosoft Power PlatformLow-code & automation
Vendor
Microsoft
Components
Power Apps, Power Automate, Copilot Studio, Dataverse
Current update
2026 wave 2, North America 16 to 19 Oct 2026
Roadmap model
Continuous AI at Work roadmap from Sep 2026
release watch
  • Release waves replaced by the AI at Work roadmap25 Aug 2026
  • 2026 wave 2 general availability
connects with
in short·15-second read

Cohash builds and governs Microsoft Power Platform solutions.

We create canvas and model-driven Power Apps, Power Automate flows and Copilot Studio agents on Dataverse. We also design environment strategy, data loss prevention policies and pipelines for moving solutions to production. As Microsoft shifts from release waves to a continuous roadmap, we help admins track changes and test mandatory updates first.

you probably need this if
  • Hundreds of flows run in the default environment, and half the people who built them don't work here anymore.
  • Solutions move between environments as exported zip files on someone's laptop.
  • Someone wants a Copilot Studio agent, and security wants to know exactly which connectors it can reach before saying yes.
Ask an engineer facts checked 2026-10-20 · 3 sources
how it runs
  1. 01Tenant inventory
  2. 02Environment and DLP model
  3. 03Build as solutions
  4. 04Pipelines to production
  5. 05Handover to the CoE
you keep
  • Environment strategy with managed environments and an owner per environment
  • DLP policy set with connector classifications and the reason for each
  • Inventory of apps, flows and connectors with orphaned items flagged
+ 2 more documents
who does it
Power Apps developer (canvas and model-driven)Power Automate developerCopilot Studio developer Microsoft Certified: Power Platform Developer Associate Microsoft Certified: Power BI Data Analyst Associate
platforms
proof
7 case studies on this
power-platform · release timeline
  1. 25 Aug 2026 Release waves replaced by the AI at Work roadmap New Power Platform and Dataverse capabilities publish continuously from September 2026. Release Planner retires by 15 November 2026. source ↗
  2. 17 Sep 2026 September 2026 feature update App-builder skill for model-driven apps and the model-driven header and navigation refresh are generally available. source ↗
  3. today · Oct 22, 2026
  4. no date yet2026 wave 2 general availabilityMandatory platform update deployed by region. North America is scheduled for 16 to 19 October 2026.source ↗
the story · microsoft power platform estate

From the first call to a setup your own team can run.

Here's what usually prompts a call about your Microsoft Power Platform estate, what we do about it and in which order, and where you end up.

  1. chapter 01 · where it hurts

    Why teams reach out

    • Hundreds of flows run in the default environment, and half the people who built them don't work here anymore.
    • Solutions move between environments as exported zip files on someone's laptop.
    • Someone wants a Copilot Studio agent, and security wants to know exactly which connectors it can reach before saying yes.
  2. chapter 02 · what we change

    How we solve it

    • Environment strategy and governance
    • Power Apps development
    • Power Automate flows
    • Copilot Studio agents
    • ALM and deployment pipelines
  3. chapter 03 · how it runs

    Our approach

    • Tenant inventory
    • Environment and DLP model
    • Build as solutions
    • Pipelines to production
    • Handover to the CoE
  4. chapter 04 · where you end up

    What changes

    • Critical flows live in owned, governed environments.
    • Each employee sees only their own data through the agent.
    • Environment strategy with managed environments and an owner per environment
    • DLP policy set with connector classifications and the reason for each
what we do

Our Microsoft Power Platform work

5 kinds of Microsoft Power Platform work, each run by engineers who have done it on live tenants.

  1. Environment layout, managed environments, data loss prevention policies and connector rules are agreed with your security team. Then they're enforced through the admin center.

  2. Model-driven apps on Dataverse for structured processes and canvas apps where the screen design matters, built as solutions from day one.

  3. Cloud flows for approvals and integrations, and desktop flows where a legacy system has no API. Error handling and run history are designed in.

  4. Agents built on your knowledge sources and Dataverse data, with authentication through Entra ID and actions limited to what each user may already do.

  5. Solutions moved from development to test to production through Power Platform pipelines or Azure DevOps, with environment variables and connection references set per stage.

capability 01 / 05in production

Environment strategy and governance

Environment layout, managed environments, data loss prevention policies and connector rules are agreed with your security team. Then they're enforced through the admin center.

talk to an engineer about this →
capability 02 / 05in production

Power Apps development

Model-driven apps on Dataverse for structured processes and canvas apps where the screen design matters, built as solutions from day one.

talk to an engineer about this →
capability 03 / 05in production

Power Automate flows

Cloud flows for approvals and integrations, and desktop flows where a legacy system has no API. Error handling and run history are designed in.

talk to an engineer about this →
capability 04 / 05in production

Copilot Studio agents

Agents built on your knowledge sources and Dataverse data, with authentication through Entra ID and actions limited to what each user may already do.

talk to an engineer about this →
capability 05 / 05in production

ALM and deployment pipelines

Solutions moved from development to test to production through Power Platform pipelines or Azure DevOps, with environment variables and connection references set per stage.

talk to an engineer about this →
typical situations

Microsoft Power Platform situations we walk into

We pieced these together from Microsoft Power Platform situations that come up a lot. They're patterns. You won't find a specific client in them.

case file 01

4,000 flows in the default environment and one part-time admin

composite · not a client story
  1. Beforethe situation

    The default environment holds about 4,000 flows and 600 apps, many owned by people who have left. A few run payroll exceptions and vendor onboarding, and DLP lets any connector talk to any other.

  2. What we dothe work

    We inventory everything with the CoE Starter Kit or the admin APIs, find the items the business depends on, and move them into managed environments as solutions with owners. DLP tightens in stages, starting with connectors that move data outside the tenant, and makers get a request path for new environments.

  3. Afterwhere you end up

    Critical flows live in owned, governed environments. The default environment is limited to personal productivity, and each DLP rule has a stated reason.

case file 02

An HR agent in Copilot Studio that answered with too much data

composite · not a client story
  1. Beforethe situation

    HR built a Copilot Studio agent on top of a SharePoint site and a Dataverse table of employee records. In testing it answered questions about other people's leave balances, because it was running on the maker's connection.

  2. What we dothe work

    We switch the agent to authenticate users through Entra ID with their own connections, narrow knowledge sources to the policy site, and replace direct table access with an action that returns only the signed-in user's record. DLP is reviewed for the connectors it uses.

  3. Afterwhere you end up

    Each employee sees only their own data through the agent. The maker's credentials are out of the loop, and the agent reaches production through the same pipeline as other solutions.

case studies

Microsoft Power Platform case studies

All case studies
AI rollout AI and agents

Classic RPA bots to agentic automation

Moving classic RPA bots to agentic automation in UiPath or Power Automate. APIs first, agents only where judgment is needed, and controls on both.

read the case study
AI rollout AI and agents

Microsoft 365 Copilot or Gemini Enterprise rollout with real adoption

Rolling out Microsoft 365 Copilot or Gemini Enterprise with oversharing fixed first, use cases tied to real tasks and adoption measured past login counts.

read the case study
migration ERP and HCM

Dynamics AX 2012 to Dynamics 365 Finance & Operations

We take AX 2012 R2 or R3 systems to Dynamics 365 finance and operations apps, and the code, data, attachments and integrations come across with them.

read the case study
migration CRM, ITSM and automation

Dynamics CRM on-premises to Dynamics 365 online

We move Dynamics CRM or Customer Engagement 9.x off your servers and onto Dynamics 365 online, rebuilding customizations, reports, integrations and sign-in.

read the case study
migration CRM, ITSM and automation

Siebel or homegrown CRM to Salesforce or Dynamics 365

We replace Siebel or homegrown CRMs with Salesforce or Dynamics 365 and move business logic, data and integrations deliberately.

read the case study
migration Data and analytics

Power BI Premium P SKUs to Microsoft Fabric F SKUs

Moving Power BI Premium per-capacity P SKUs to Microsoft Fabric F SKUs before the agreement ends, sized from real load and reassigned in validated batches.

read the case study
approach

How a Microsoft Power Platform engagement runs

We don't start a step until the last one has produced something you can check.

  1. step 01

    Tenant inventory

    We list environments, apps, flows, agents and connectors from the admin center and CoE data, then mark what is orphaned, risky or worth keeping.

  2. step 02

    Environment and DLP model

    Development, test and production environments are agreed with your security team, with managed environment settings and connector rules for each tier.

  3. step 03

    Build as solutions

    Apps, flows and Copilot Studio agents are built inside solutions from the start, with environment variables and connection references instead of hard-coded values.

  4. step 04

    Pipelines to production

    Power Platform pipelines or Azure DevOps promote managed solutions through each stage, with an approval before production.

  5. step 05

    Handover to the CoE

    Your admins keep the governance rules, a maker onboarding guide and a routine for reviewing new connectors and stale flows.

ai on microsoft power platform

Where AI fits on Microsoft Power Platform, and where it needs guardrails.

Power Platform's AI is mostly Copilot Studio and Dataverse. In Copilot Studio, makers build agents that call tools from MCP servers. Dataverse can be an MCP server itself. Makers can also build model-driven apps with the app-builder skill now. For engineering teams the hard parts are authentication and DLP, and then who pays for the Copilot credits.

AI agentasks for data or an action
Scoped identityonly the access it needs
MCP serverthe one door into the platform
Logged, approveda person signs off on writes
Microsoft Power Platformdoes the work

Microsoft publishes the Dataverse MCP server, and Copilot Studio can call external MCP servers. We register MCP connectors per environment under DLP rules and use end-user authentication through Entra ID rather than maker connections. Allowed-client lists stay short. Agents need approval before they delete records or change schemas.

Copilot Studio agents with MCP tools

Copilot Studio agents connect to MCP servers and use their tools and resources, which update on their own when the server changes. You need generative orchestration on. We choose the MCP servers each environment can reach, run agents as the signed-in user through Entra ID, and promote them through pipelines like any other solution.

Dataverse as an MCP server

Dataverse exposes tables and records to MCP clients, with tools for search, SQL read queries, record writes and schema changes. Whether the server is on, and which clients may connect, is set per environment. For us, schema tools only run in development environments. Microsoft renames tools from time to time, and each time we recheck the client allow-lists.

App-builder skill for model-driven apps

The September 2026 feature update made the app-builder skill for model-driven apps generally available. Makers can use it in managed environments. It generates tables, forms and security roles, and we review all three before an app leaves development.

Credit and DLP review

Copilot Studio usage costs Copilot credits, and so do Dataverse MCP calls from agents built somewhere else. So before launch we estimate volumes, and we check the DLP policy on every connector an agent can reach.

Microsoft Learn: Extend your agent with MCP in Copilot Studio ↗Microsoft Learn: Connect to Dataverse with MCP ↗Power Platform blog: September 2026 feature update ↗
architecture

How a Power Platform solution moves from maker to production

Makers build apps, flows and Copilot Studio agents in a development environment. Pipelines promote them as managed solutions into a production managed environment, where DLP rules control which systems they can reach.

How a Power Platform solution moves from maker to production Makers build apps, flows and Copilot Studio agents in a development environment. Pipelines promote them as managed solutions into a production managed environment, where DLP rules control which systems they can reach. 01 · MAKERS AND IDENTITY 02 · PROMOTION 03 · PRODUCTION 04 · CONNECTED SYSTEMS solution solution deploy security group read · write shared tables connectors desktop flows Makers Power Apps · PowerAutomate Copilot Studio agents · topics · actions Ex Entra ID sign-in · securitygroups Power Platformpipelines managed solutions ·approvals PP Managedenvironment DLP · sharing limits Dataverse tables · row-levelsecurity D365 Dynamics 365 shared Dataversetables SharePoint and Outlook standard connectors Legacy desktop app no API · desktop flows
Read the diagram as text
  1. Makers → Power Platform pipelines: solution
  2. Copilot Studio → Power Platform pipelines: solution
  3. Power Platform pipelines → Managed environment: deploy
  4. Entra ID → Managed environment: security group
  5. Managed environment → Dataverse: read · write
  6. Dataverse → Dynamics 365: shared tables
  7. Managed environment → SharePoint and Outlook: connectors
  8. Managed environment → Legacy desktop app: desktop flows
use cases

Typical Microsoft Power Platform projects

The work teams usually bring us, and the platforms Microsoft Power Platform has to talk to while we do it.

  1. POW-01 Hundreds of personal flows run in the default environment with no owner common request
  2. POW-02 A business team wants an app this quarter and IT can't build it common request
  3. POW-03 You want a Copilot Studio agent that answers from internal policies common request
  4. POW-04 Solutions are exported and imported by hand between environments common request
  5. POW-05 A legacy desktop system needs automation but has no API common request
Microsoft Power Platform · connected platforms
Microsoft Dynamics 365Microsoft FabricMicrosoft Entra IDSalesforceOpenAI API
you keep

What a Microsoft Power Platform engagement leaves behind

These are working documents, written against your configuration, so your team can run them without us.

  • document01

    Environment strategy with managed environments and an owner per environment

    yours to keep
  • code02

    DLP policy set with connector classifications and the reason for each

    yours to keep
  • inventory03

    Inventory of apps, flows and connectors with orphaned items flagged

    yours to keep
  • tests04

    Pipeline setup for dev, test and prod with environment variables per stage

    yours to keep
  • report05

    Copilot Studio review checklist covering authentication, knowledge and actions

    yours to keep
talent and delivery · microsoft power platform

Certified engineers, wherever you need them.

Some clients want a project team, some want a couple of engineers on theirs. Either works. Onsite, nearshore or offshore, with handovers across time zones that keep things moving.

Onsite

Engineers at your sites in the US

Nearshore

Overlapping hours with US teams

Offshore

Overnight progress on your backlog

All time zones

Coverage handed over across time zones

24-hour coverage · US Central time
OnsiteUS business hours
Nearshoreoverlapping US hours
Offshoreovernight progress
Combinedhandover, not hand-off
certifications our engineers hold
roles we staff
  • Power Apps developer (canvas and model-driven)
  • Power Automate developer
  • Copilot Studio developer
  • Power Platform administrator (environments, DLP)
  • Power Platform solution architect
  • Power BI developer
due diligence · microsoft power platform

Questions to ask any partner, including us.

Put these to anyone bidding on your Microsoft Power Platform work. The ones who've done it answer differently from the ones who've read about it.

  1. Q1

    How will you decide what stays in the default environment and what moves out?

    Moving everything breaks personal productivity, and moving nothing leaves critical processes unowned. The rule needs criteria such as business impact, data sensitivity and audience.

  2. Q2

    Which connectors will you block, and how will makers request exceptions?

    DLP that blocks without a request path pushes makers to workarounds. Classifications should come with a documented exception process.

  3. Q3

    Will our Copilot Studio agents act as the signed-in user or as the maker?

    An agent on the maker's connection can expose data the user should never see. End-user authentication and narrow actions keep the agent inside the user's own permissions.

  4. Q4

    How will solutions reach production, and who approves each release?

    Changes made directly in production are hard to audit or roll back. Pipelines with environment variables and connection references make each release repeatable.

  5. Q5

    How will our admins hear about platform changes once release waves end?

    Microsoft is moving to the continuous AI at Work roadmap, and Release Planner retires by 15 November 2026. Admins need a filtered view and a regular review, plus early access testing.

01Should makers build in the default environment?

For personal productivity, yes, as long as the DLP policies are tight. Anything a team depends on belongs in a managed environment with an owner, a solution and a path to production. We usually set that boundary early.

02Do we need Dataverse, or can we use SharePoint lists?

SharePoint works for small, simple apps. Dataverse is the better choice when you need relational data, row-level security, auditing or Dynamics 365 integration. Licensing differs between the two, so we compare cost as well.

03How do Copilot Studio agents access our data safely?

We configure authentication so the agent acts as the signed-in user, limit knowledge sources and actions to what that user can already reach, and review DLP policies for the connectors the agent uses.

04Can you clean up an existing Power Platform tenant?

Yes. We inventory environments, apps, flows and connectors, find orphaned or risky items, and agree an environment and DLP model before moving what should stay.

05What drives Power Platform licensing cost?

Mostly how many users need premium connectors or Dataverse, plus Copilot Studio usage and any capacity add-ons. Design choices such as Dataverse versus SharePoint change the bill, so we compare options against Microsoft's current licensing guide before building.

services

Related services

Projects on Microsoft Power Platform usually end up touching one of these, so they're worth a look.

you are hereMicrosoft Power Platform
sources 3 sources linked last checked 2026-10-20 content fingerprint 3471f31d
  1. 01Microsoft Dynamics 365 blog: moving beyond release waves (25 Aug 2026)microsoft.com ↗
  2. 02Microsoft Learn: general availability deployment schedulelearn.microsoft.com ↗
  3. 03Power Platform blog: September 2026 feature updatemicrosoft.com ↗
Reviewed by the Cohash team Microsoft Power Platform is a trademark of its owner, used here to identify the platform. how we handle facts and data →
next step

Tell us what you're running. We'll tell you what we'd do first.

A 30-minute working session with a senior engineer. No slides.

Book a working session See the case studies
useful to bring
  • The platform and the version you're on
  • Where you need to get to, and by when
  • What's custom: rules, plugins, integrations
  • Any audit finding or deadline driving it