Digital Revolution Awards

We were delighted to attend the Digital Revolution Awards, celebrating the organisations and individuals driving meaningful change through technology across the UK. It was a fantastic evening recognising real innovation and impact, and a great reminder of the quality of work happening across our industry. Congratulations to all of the nominees and winners on the night.

The sun came out just in time for our annual Summer Social at Bondi Green. Hosted with Seer 365 and sponsored by our friends at TrueCommerce, iDynamics Business Solutions, Yavrio - Open Banking, Vertex Inc., Xe.com, Formpipe, Revolent Group, and Binary Stream Software - Microsoft ERP Solutions, the aim was, as always, to keep things simple - no slides. no presentations or pitches - just great conversation and a few cocktails, all enhanced by the gorgeous backdrop of the canal, thanks to Bondi Green's superb location.


These kinds of nights are a reminder that real community matters. Thanks to everyone who joined us - if you'd like to join us next year, please just let us know at info@oaka.studio.

IAMCP Events:

European Power Platform Conference: June 29 – July 2

Microsoft Partners Don't Need an AI Practice. They Need Something Customers Will Buy. 

We've had roughly the same AI conversation with Microsoft partners for a couple of years now. It usually opens with some version of "we know we need to do something with AI". 


Fair enough. The next part is where it gets less useful. Teams start debating which Microsoft technology to invest in, whether they need more Copilot Studio skills, how much Azure AI capability to build and whether they should start calling themselves an AI partner. 


Meanwhile the customer has a much simpler concern. They've got a problem and they'd quite like someone to solve it. 


That matters because the AI market moves fast. Six months ago customers were asking different questions, and six months from now they probably will be again. Building a practice around whichever technology is getting the most attention this quarter is an expensive way to keep reorganising your business. 


The better starting point is what customers are actually prepared to spend money on now

Customers have mostly moved past the excitement phase 


Most customers no longer need convincing that AI might be useful. They've seen the demos, their leadership team has probably been to a conference and someone internally is already using ChatGPT, sanctioned or not. 


The conversations we're seeing now are far more practical. Customers want productivity, because hiring is expensive and there's pressure to do more with the people they've got. They want faster, better decisions, because new competitors are arriving without the same legacy systems and processes. They're looking at AI-enabled customer service, because the technology is finally usable enough to put in front of customers without producing an embarrassing screenshot for social media. 


And they're worried about risk. Quite reasonably. 


The market is shifting from "show me what AI can do" to "show me how to use it without creating another problem". That's a much better market for partners. 


The blockers are part of the opportunity 


A lot of partners still treat customer hesitation as something to get past before they can sell the AI project. We tend to see it differently. The hesitation is often the project. 


Customers keep running into four fairly predictable problems. They don't know which use cases are worth pursuing, their data is a mess, their governance isn't ready and they have no sensible operating model for managing AI once it starts spreading through the organisation. 


Take data. Most organisations have spent years piling up SharePoint sites, Teams channels, documents and permissions. There are five copies of the same presentation, three of them called "FINAL", and nobody's entirely sure which one is. Then someone switches on Copilot and is surprised the answers reflect the information underneath them. 


AI didn't create that problem. It just made it much harder to ignore. 


For Microsoft partners, sorting out the data, permissions, security and governance around AI is commercially valuable work in its own right. You don't have to wait for the glamorous project that comes afterwards. 

Don't put AI into something Power Automate already does perfectly well 


There's another urge worth resisting: using AI because it's there. 


If a process reads "when field X equals this and field Y equals that, do Z", you've got a deterministic workflow. Power Automate has been handling that sort of thing happily for years. Adding an agent doesn't make it more sophisticated. It gives you a more expensive and less predictable way of getting the same answer. 


AI earns its place where judgement is needed, where inputs are unstructured or where the answer can't be neatly expressed in rules. That sounds obvious written down. It becomes surprisingly less obvious once everyone's in a workshop with "agentic AI" on the slide. 


Four sensible places for most partners to start 


We mapped ten AI offers against what we're seeing customers buy. They range from fairly accessible services to much bigger bets, and for most Microsoft partners four sit at the sensible end of the investment curve. 


AI security and governance comes first. Customers need policies, controls, ownership and auditability before they can safely expand their use of AI, and a security partner already has much of the credibility and capability to lead that conversation. 


Data access and permissioning is the second. Microsoft 365 Copilot can only be as sensible as the environment you expose it to, so SharePoint structures, permissions and sensitivity need attention before AI starts reasoning across an organisation's data. 


Third is automation and integration quick wins. There are plenty of processes where Power Automate and Copilot Studio work well together, provided someone has first checked that AI actually belongs there. 


Fourth is Copilot readiness and value assessments. This is less about deploying technology and more about finding worthwhile use cases, defining what success looks like and putting guardrails in place before money gets spent. 


None of these are especially exotic, which is partly the point. They map to problems customers are already trying to solve, and for many partners they can be delivered with capability that's already sitting somewhere in the business. 


The more interesting work comes later 


Further along, the opportunities get bigger and so do the capability requirements. 


Line-of-business augmentation is one. Rather than asking someone to leave Dynamics or another business application to have a separate conversation with an AI tool, AI sits inside the work they're already doing. Summaries, next-best actions and knowledge retrieval become part of the application instead of another place to go. 


Customer service is another area we expect to grow considerably. Organisations are getting more comfortable letting AI deal directly with customers, particularly in authenticated environments where context and audience can be controlled. The potential is significant. So is the damage from getting it wrong. 


Then there's data readiness. "Every AI project is a data project" has been said often enough to risk becoming wallpaper, but annoyingly it's still true. Customers need trustworthy data, a clear view of the gaps and information they can actually use before more advanced AI work is worth doing. 


Beyond that you're into multi-step agents with human oversight, repeatable industry solutions and eventually proprietary data or pre-trained models that can make up for customers whose own data estates are years from ready. Potentially substantial opportunities, and not where every partner should start. 

You probably shouldn't become an "AI partner" 


This is where a lot of strategy conversations go wrong. A Power Platform partner spots the AI opportunity and decides it needs security, data engineering, Azure AI, adoption and governance capability. A security partner decides it needs a team of AI engineers. Everyone starts hiring. 


There's a more boring and usually cheaper answer: start with what you're already good at. 


For a Power Platform partner, that might mean workflow automation, Copilot Studio, implementation and adoption. For a security partner, AI governance, Purview, permissions and data protection are far more natural extensions. An Azure or data partner may have a credible route into the more engineered end of the market. 


Your existing expertise isn't something you need to abandon to take part in AI. It's probably your best way in. Customers aren't hunting for companies that have added "AI" to the top of their website. They want people who understand their problem and can apply the technology to it safely. 


Some capabilities are better borrowed than built 


Once you accept you don't have to do everything yourself, the practice conversation gets much easier. 


A Power Platform partner may need security expertise around permissions, Purview, DLP and sensitivity labels to deliver a broader AI engagement properly. That doesn't automatically mean hiring a security team. A security partner may need Power Platform capability without wanting to become a Power Platform practice. 


So partner. There's nothing especially noble about owning every link in the delivery chain. If another Microsoft partner has complementary capability, works with similar customers and can help you deliver a stronger proposition, work with them. Joint offers and referral arrangements can fill gaps now while you decide whether those skills deserve permanent investment later. 


The important decision is knowing which capabilities you'll own, which you'll develop over time and which you've no intention of building. That last list is useful. More partners should have one. 


Copilot Chat is a starting point, not much of a business model 


Microsoft's current AI stack gives partners several places to play. 


Copilot Chat is a fairly straightforward entry point that gives employees a sanctioned general-purpose AI tool instead of leaving them to find their own. There's some training work around it, but not much implementation opportunity. 


Microsoft 365 Copilot gets more commercially interesting because organisational data enters the picture, which means use cases, permissions, governance and adoption all need attention. 


Copilot Studio moves further towards agents embedded in business processes, often integrated with other systems. 


Azure AI Foundry sits at the heavily engineered end, with more customisation and correspondingly heavier requirements around engineering, data pipelines and governance. 


It's tempting to look at that progression and assume moving further along makes you a more mature AI partner. It doesn't. It makes you a partner doing more technically complex AI work, which isn't the same thing. A profitable governance service that solves a recurring customer problem is a better business than an impressive Foundry capability nobody has agreed to buy. 


Build offers customers can understand 


"We do AI" isn't an offer. Neither is "we can help you unlock the power of Copilot", however much website space is currently devoted to variations of it. 


Customers need to know what you'll do, what problem it solves and what's different when you've finished. 


For new customers, Copilot readiness can be a useful way in, with the engagement moving on to automation quick wins, line-of-business augmentation and eventually customer service AI. With existing customers, security and governance is often the more natural opener, because you already know their environment and have enough trust to talk about what Copilot can see, what it should see and which controls need to exist. 


That gives salespeople something concrete to sell. It also gives customers a sensible starting point without signing up to a huge AI transformation programme before anyone knows whether the first use case works. 


Some of what you build will have a short shelf life 


This is the bit technology businesses find uncomfortable. Some AI services you build today may need substantial rework within a year. 


That doesn't automatically make them bad investments. Things that are difficult now may be trivial in twelve months, and waiting for the market to settle sounds prudent until you realise it may not settle for some time. 


The real judgement is whether the investment needed is proportionate to the revenue and customer value available while the capability stays useful. A fairly cheap capability with strong demand for the next year can be a perfectly good business. A big hiring programme for a narrow technical service Microsoft might turn into a button next spring deserves rather more scrutiny. 


Customers face the same question. Sometimes the sensible AI project is one that delivers value quickly, even if everyone accepts it'll be rebuilt as the technology improves. Not everything needs to become a family heirloom. 

The hard part is deciding what you won't do 


Most partners can produce an impressive list of AI opportunities. That's the easy bit. The useful work is narrowing it down. 


Which industries and customer problems already fit your business? Which existing capabilities give you a credible starting point? Which offers can you sell without spending six months building a team first? Which capabilities are worth developing and which should come from partners? 


Then the question that tends to get skipped: what are you deliberately not going to build? 


A security partner doesn't need to become an Azure AI engineering house just because Foundry exists. A Power Platform specialist doesn't need to recreate a mature security practice before selling its first Copilot Studio engagement. There's enough opportunity in AI without every Microsoft partner trying to become every type of Microsoft partner at once. 


We've watched plenty of partners spend months debating AI strategy when a perfectly viable offer was sitting one step from what they already do. The answer is usually less dramatic than expected. 


Take something you already understand. Apply it to a problem customers are paying to solve right now. Package it clearly. Borrow the capabilities you don't need to own. 


Then sell it. You can worry about becoming an AI company later. 


Working out where AI fits in your business? 

Get in touch

If you'd like help identifying the AI offer that sits closest to what you already do, and what it would take to package and sell it, email us. We'll look at your current capability with you and give you a straight view of where to start and what to leave alone. 

info@oaka.studio

< Older Post Newer Post >

Creating value: what really moves the dial

By Lucy Bourne • October 21, 2026
That was the starting point for September's Oaka Partner Insights. Consolidation across the channel isn't slowing down, and the partners we speak to keep asking the same things. What does a succession plan look like? How do I position my business? Which metrics actually matter to the people writing the cheques? Our consultant Nigel Ridpath put those questions to three people who have seen deals from different sides of the table. Adam Kerr is Managing Partner and Founder of Primas Law, specialising in corporate transactions. Jack Henry is a barrister and owner of Henry Chambers, who has been through acquisitions as a business owner. Neil Lomax spent 17 years at SoftwareOne, latterly responsible for around $22 billion in Microsoft revenue, before co-founding Onyx. Here's what we took away.

Your Microsoft Designation Isn’t a Badge. Stop Treating It Like One.

By Lucy Bourne • October 21, 2026
A partner works hard for a Solutions Partner designation. Customers get associated, certifications get chased, the Partner Capability Score finally creeps over the line and everyone breathes out. Then the designation gets left alone. A few months later someone opens Partner Center and finds the score has slipped, a certification has lapsed, customers have gone missing or the renewal window is approaching with the business sitting on 41 points. Panic follows, as it tends to. The issue is rarely that the partner doesn’t understand Microsoft’s requirements. It’s that designation management has been treated as a project with a finish line rather than an ongoing part of running a Microsoft practice. This year, that difference matters more than it used to. 

The new FY27 Incentives Don’t Break Your Business Model. But They Might Have Exposed It.

By Lucy Bourne • October 21, 2026
Every July, LinkedIn fills up with the same posts declaring that Microsoft has changed everything. Most of them are written before anyone's actually opened the rebates guide. The honest version is fairly simple: Microsoft has doubled down on net-new business. In our experience, a lot of partners still haven't. The part that catches people out the most has been the pace of change - This didn't creep in over several renewal cycles. The direction changed sharply between FY26 and FY27 and it's still moving and every quarter the gap widens between the model Microsoft is funding and the one most partners are still running. Treat it as a slow drift you'll adjust to eventually and you'll be building a plan for a market that has already moved on. As always – and as Oaka often refer to in our “Partner Genome” being ambiguity tolerant is not just something for the technical delivery leads within the practices we serve, but also the partnership focussed leaders.
Show More