Where Is AI Going? By the alliantDigital team

In our conversations across dozens of industries – with PE firms, CEOs, and accounting firm managing partners alike – a consistent set of questions keeps surfacing. They may sound different at first, but they share a common underlying theme: the inevitable impact AI will have on the future of professional services. The questions may start with AI, but they never end there. Here are some of the most prevalent ones we hear.

Our answer is always the same, regardless of industry: start small. Go for a single or a double – the most important thing is just getting on base. Most companies are not as far behind as they assume, and very few are actually “ahead.” Only 6% of organizations qualify as AI high performers, meaning they generate 5%+ EBIT impact from AI (1). The overwhelming majority are still in the early stages. If you start small and build a solid base, you’re likely ahead of the curve already.

The approach we advocate is a four-step framework: plan, pilot, self-fund, then scale with agents. Start with ideation – where do you wish you had a digital assistant? Then pick a pilot that is lower complexity, well-defined, and has a clear ROI. Once that pilot is working, it funds the next one. By the end of the year, you’re often running a zero-budget AI transformation.

One method we utilize is providing clients with solutions that already have a proven and established framework – before customizing those solutions to a client’s specific tech stack, workflows, and strategy. Starting with demonstrable, repeatable wins is a great way to begin with AI.

Think of it like black ice. When you drive over it, the first instinct is to brake, but that’s exactly the wrong approach. You steer into the skid, straight into the direction the car is sliding. AI transformation is the same. Fear is natural. But the answer isn’t to stop or slam on the brakes – it’s to steer deliberately, with a plan, proper governance, and security, toward the outcome you want.

We’re still in the early stages of a genuinely transformative technology. Some skidding is inevitable. The numbers reflect that. According to MIT’s NANDA Initiative, 95% of enterprise generative AI pilots deliver zero measurable P&L impact. Our own surveys show that while 75% of executives consider AI a priority, only 25% say they’re realizing value.

In our view, it’s not a technology problem – it’s a people and process problem. We’ve seen this play out time and again with our clients. The change management piece has to come first. Getting the right stakeholders aligned on the steps and the plan before building anything is what separates successful deployments from failed ones.

The biggest mistake companies often make is starting too big – driven by the belief that they’re much further behind than they actually are. We often say that without a destination, any direction will do. We won’t let a client do ten things at the same time. We find a pilot, we ensure it’s successful, then build from there. Iterative, fast, and self-funding is the strategy.

What we advocate is a culture of experimentation – where some failure is not just tolerated but expected. Organizations that treat AI as a one-time project will find themselves perpetually behind. The ones that get it right build for a moving target. This is especially true with AI, since the very nature of AI is that it learns with you as you go.

Now about three years into AI, we’ve seen which use cases actually get adopted and which ones fade. If you’d asked us three years ago to name the ten highest-value applications for an accounting firm, our 8879 E-Filing tool wouldn’t have made the list. But it’s become one of our most requested. It touches a workflow that every firm needs, and it works end to end – from tax return completion to client sign-off, to IRS verification, to filing.

The same goes for IRS and state taxing authority notices. Firms are building agents that read the notice, pull the relevant client file, draft a response, and route it for review. What previously took hours of experienced professional time now takes minutes.

Billing and collections is another area where we’ve seen enormous demand – and where AI finally delivers. We’ve spent years hearing partners say: if you can solve billing automation, we will carry you out on our shoulders. AI is the first technology that actually delivers on that need, and we love deploying this solution to our clients.

One we’re particularly proud of was a client retention solution. It quickly retained 30% of clients who were already on their way out the door and saved nearly $900,000 in revenue that would have been lost otherwise. It started with one simple capability: reading call scripts and identifying at-risk customers. The “save” team went from being underutilized to having a full, objective pipeline of customers worth engaging, and continues to deliver with the help of the AI retention solution.

What makes all of this scalable – and what most firms are beginning to realize – is that these AI components are reusable. Once you’ve built the capability to upload, read, and act on a document, that same component can be deployed across a billing agent, a contract review agent, or an RFP agent. You’re not starting from scratch each time, you’re stacking. We call these Lego bricks. The more you build, the faster the next deployment happens.

And for a significant number of these use cases, we already have a partially-built solution before we even begin. The architecture, the logic, and the agent are already there. What remains is configuring it to your tech stack, your workflows, your client base. That’s weeks – sometimes days – to a meaningful ROI, or even a complete payback. Not a nine-month ERP implementation. A well thought-out AI strategy provides a much shorter ROI runway.

It’s a reasonable concern. You’ve invested heavily in your current technology stack – the last thing you want is to blow it up in pursuit of something new.

Luckily, you don’t have to. According to recent data, 78% of organizations now use AI in at least one business function – up from just 55% in 2023. The vast majority of them didn’t get there by replacing what they had. They got there by layering AI on top of it. That’s precisely the philosophy behind our approach: platforms that sit on top of existing tech stacks, not in place of them.

The same principle applies to your people. The workforce question and the technology question are really the same thing – as is the answer. What we’re talking about is a reinvention of your operating model: a new complement of offshore, nearshore, onshore, human, and non-human resources, working in concert. Agents don’t replace your team – they become part of it. They take on the work your people don’t want to do, free them up for the creative, strategic work they’re passionate about, and get better over time as your business evolves.

But we also want to be direct about something. The question we hear underneath all of this is: can I afford to wait?

We believe there will be winners and losers in this era, and the dividing line isn’t going to be size, or reputation, or how well things are going today – it’s going to be who moved and who didn’t. The firms that treat this as the reinvention – pervasive, strategic, fully committed – will not only survive, they’ll be unrecognizable in the best possible way. You don’t need to blow up what you’ve built. You just need to build on what you have.

We’re moving toward agentic AI, and quickly. The next phase isn’t about individual tools or one-off automations. It’s about AI taking on entire motions within an organization – quote to cash, intake to delivery – with co-working stops where the human comes back in and says: slow down, rework this, go back. We love workflow that doesn’t just go forward, but can take steps back. That’s what agentic AI can actually enable.

The platform question is also going to sort itself out fast. Right now, the market looks a lot like 1999 – over 100,000 AI startups, most of them doing one narrow thing, most of them looking for an exit. That’s not a platform, it’s a vendor. And when they cash out and no longer support their products – they shake your hand and say sorry. The firms that win this era will be the ones that consolidate their AI into a single layer – one place where your data, your agents, your automations, and your people all come together. That’s the bet we’ve made at alliantDigital.

Most of our conversations in the marketplace start with AI – the technology, the use cases, the ROI, the governance – but they never end there. They inevitably become about people, change, and transformation. AI is a revolutionary technology. But it’s very human at its core.
We’ve spent our careers doing one thing: taking organizations – sometimes small teams, sometimes entire enterprises – from where they are today to where they need to be. To do that, you have to be able to see the future clearly enough to bring others along with you. There has never been a moment where that work felt more urgent, or more possible, than right now. The question is whether your organization is heading there with intention, or waiting to see what happens. That’s what alliantDigital is built to do – and there has never been a better moment to do it.

References:

1. https://www.fullview.io/blog/ai-statistics

2. https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai