If you work in B2B software right now, it can feel like every second product suddenly “does AI”.
For owners and operations leaders in product-based businesses, most of that noise doesn’t help you ship one more order, fix a stock issue, or close the month any faster. It’s a lot of chatbots and slideware – not much that makes Tuesday afternoon in the warehouse any easier.
That’s the gap The Access Group’s Small Business Division is trying to close.
In a recent conversation, Ben Carter, Product Director for the division, sat down with Chartered Accountant, Podcaster and Digital Transformers Founder, Ryan Pearcy, to talk through some of the very practical ways his teams are already using AI across the Small Business stack.
The examples they touched on are just four points on a much bigger AI roadmap inside The Access Group. But they’re a useful window into the philosophy:
“Less hype, more hard yards in the background to get customers to value faster.”
How quickly can we prove ourselves?
Unleashed Small Business Division alongside Prospect CRM and Mintsoft – together covering inventory, CRM and warehousing for product-based SMEs. The division’s north star, as Ben puts it, is simple:
“Our vision is really to become the most recommended software providers for small businesses. Being the most recommended is the best indicator of success.”
To get there, they’ve set themselves a very specific product mission:
“We talk a lot about achieving value, and value in about seven minutes… The people who buy our products are the ones that use them. So we’ve got to establish value quickly or they’ll just abandon us.”
AI, in this world, isn’t a gimmick. It’s one of the main ways they’re shrinking the gap between “I’ve signed up” and “this is genuinely helping me run the business.”
Here’s how.
1. AI that makes the trial personal
Empty demo accounts don’t sell inventory software. Seeing your own SKUs, quirks and edge cases in the system does.
Ben shared a stat that underpins a big chunk of their AI work:
“Did you know that if you upload your own data into one of our trials, you’re twice as likely to then go on and purchase?”
“Did you know that if you upload your own data into one of our trials, you’re twice as likely to then go on and purchase?”
Ryan’s reaction was exactly what you’d expect from an advisor who’s spent years rolling out systems:
“It makes complete sense – you want to see your own products, your own nuances coming through, so you can see it’s real and actually applies to you.”
The catch? For most SMEs, getting that data in has historically been painful – fiddly CSV templates, reformatting exports, time they don’t really have.
So, the Unleashed team built an AI agent to do the grunt work for Unleashed trials:
“We built a little AI agent, and you just upload any old document, and we parse it, and we say, ‘Oh, there’s your product, there’s your SKU lines,’ and we take it. So it’s really, really easy.”
How this compares to the usual AI story
In a lot of SMB software, “AI in onboarding” is still a glorified checklist or a chatbot pointing you at help articles.
What Ben describes is more tangible - using AI to remove a job nobody enjoys (massaging data into the right format), in a way that:
- Brings Unleashed’s core value – clarity and control over your stock – to life faster.
- Boosts the chance a trial turns into a long-term customer, because the system actually feels like their business.
It’s not magic, and it still needs decent source data, but it’s a concrete step beyond “here’s an empty sandbox, good luck.”
2. Creating live interfaces on site
The next thread in the Ben–Ryan conversation tackles an old problem: the gap between what a customer wants or feeds back and what gets built.
Historically, product teams gather requirements in one place, disappear into a backlog, and weeks or months later ship something that only sort of matches what was asked for.
Ben’s team is using AI-assisted tools to compress that loop:
“Often we hear requirements, we get ideas from customers, but then by the time we’ve got back to the office or we’re trying to implement them, they might have changed or we’re not sure.”
Now, they’re able to sit with customers and build interfaces in real time:
“We can now do prototyping with customers on site, in real-time… I was actually at a customer event just yesterday, and I was demo-ing my ability to create new interfaces in real-time with customers. The pace of the way we work and the momentum that we get – it’s gone up massively.”
Ryan’s response:
“That’s crazy – just being able to build on the fly when out with customers… that’s something I couldn’t even imagine.”
Under the hood, this is a mix of low-code platform, AI-assisted config and good old-fashioned product taste. But for Unleashed customers, the effect is simple:
- Your feedback on how you want to see stock, orders or BOMs can be shaped into a working screen while you’re still talking.
- There’s less lost-in-translation between ops teams, advisors and product managers.
How this stacks up in today’s market
Lots of vendors talk about being “customer-led”; fewer can sit down and literally reshape parts of the product during the meeting.
Again, this isn’t science fiction – there are guardrails, and not every request becomes a new interface on the spot. But it’s a good example of AI being used to speed up genuine collaboration, not just generate mock-ups for a slide deck.
3. Integrations that “heal”
The next topic will probably resonate (or trigger flashbacks), as it did for Ryan. He describes the frustration of so-called “integrations” that are basically glorified exports:
“We get so frustrated about seeing poor quality integrations out there, where someone’ll say it integrates, whereas actually it’s just an export file. That’s not a proper integration.”
For an Unleashed customer, inventory rarely lives in isolation. You’re typically connecting:
- Unleashed ↔ accounting (Xero, QuickBooks)
- Unleashed ↔ ecommerce (Shopify, Amazon, WooCommerce)
- Unleashed ↔ WMS/3PL (Mintsoft or others)
- Unleashed ↔ CRM (Prospect CRM)
Every change to a field, schema or workflow creates opportunities for things to break in subtle ways. This is where Ben’s passion for the subject is most obvious:
“It actually does fix itself and heal itself, which is incredible. So there’s a working example of giving an AI a focused task specifically around maintaining integrations and watching it go.”
Ryan imagines the prompt and sketches the shape of that task:
“You can say to an AI agent, ‘I want to integrate this tool. This is the data I want to share – design and define what that looks like, and then also, if anything changes that’s preventing that happening, restructure it to make sure that you’re still adhering to the objective that we want from this’ – within the parameters that we’ve set.”
So what’s actually happening?
- An AI agent monitors integrations for failures or drift.
- Within carefully defined rules, it can adjust mappings or flows when something upstream changes.
- Only when it hits a real unknown does it escalate to humans.
Compared to most “AI-powered integrations”
Elsewhere in the market, “AI integration” often means “we’ve got some pre-built connectors and a nice wizard.”
What Ben outlines is closer to an integration steward: not replacing engineers, but handling day-to-day nudges and fixes so they don’t become customer-visible issues.
It’s still evolving, and it doesn’t mean nothing ever breaks. But if you’ve lived through inventory–accounting reconciliation nightmares, even a modest reduction in silent failures is a big deal.
4. AI in context: Access Evo
The last theme from the conversation is about bringing AI into the everyday Unleashed experience, not leaving it as a separate “labs” experiment.
Across the Small Business Division, that’s happening through Access Evo – an AI-enabled experience layer that introduces things like:
- A Copilot-style assistant that lets you ask questions in natural language
- A Feed that surfaces important events, exceptions and suggested actions
- Best-in-class security, with a private AI-instance that never shares your information outside the business.
In Unleashed, that means moving towards an inventory experience where you can:
- Ask: “Which SKUs are most at risk of stock-outs next month?”
- See: “Which customers are ordering less than they did last quarter?”
- Get nudged: “These POs need attention today if you want to hit lead times.”
All in the context of clean, structured inventory data – not a generic chatbot that happens to sit next to your stock list.
Prospect CRM and Mintsoft are tapping into the same Evo capabilities in their own domains – CRM and warehousing – so over time the AI experience feels consistent across stock, customers and fulfilment.
A quick reality check
Lots of vendors now have some form of “AI assistant” in the UI. The honest distinction here is:
- Unleashed’s advantage is the quality and granularity of inventory data it already holds – BOMs, batches, landed costs, multiple warehouses.
- Evo’s job is to sit on top of that and make it easier to see what matters, faster.
It’s not about replacing operations managers or advisors – it’s about cutting the time between “something changed” and “we’ve seen it and acted.”
So how does this compare to AI in other SME software?
Stepping back, Ben’s examples land in an interesting place compared to the broader AI landscape in SMB B2B software:
More than marketing gloss
These aren’t just “we added AI to our homepage copy.” They’re tackling very specific pain points that Unleashed customers and Partners talk about all the time: getting started, building the right screens, keeping integrations healthy, spotting issues early.
Less than “AI runs your business for you”
None of this is claiming the system will automatically design your supply chain, negotiate with suppliers or replace your FD. Humans remain firmly in charge; AI is the power assist.
Firmly grounded in operations
Every example ties back to Unleashed’s core promise: giving product businesses clarity, control and confidence over inventory so they can grow without chaos.
And these four threads are just the tip of the iceberg. Across The Access Group there are many more AI initiatives in play – from analytics to helping enhance the experience of other products in the Small Business Division.
What this means for SMEs and their Advisors
Bringing it back to earth – if you’re running, advising or implementing systems for product-based SMEs, what should you take from all this?
For product businesses using (or considering) Unleashed
Faster proof, less faff
Getting your own data into an Unleashed trial with AI help means you can quickly see if it will give you the visibility and control you need – without losing a day to spreadsheet gymnastics.
Software that reflects how you actually work
With the ability for Unleashed’s product team to prototype interfaces in real-time with customers, it’ll improve screens and workflows that feel natural to your operation, not something designed in a vacuum.
A more resilient stack
AI-assisted integration maintenance won’t eliminate every issue, but it nudges your Unleashed–Xero–Shopify–warehouse stack towards “boringly reliable” instead of “it mostly works unless someone changes a field and forgets to tell us.”
Less firefighting, more proactive decisions
With Evo-style capabilities surfacing risk and opportunity inside Unleashed, you can spend fewer hours hunting for problems and more time deciding what to do about them.
For Advisors
Cleaner, quicker trial experiences
When you recommend Unleashed, AI-assisted trials make it easier to show clients real value early – crucial for buy-in from non-technical teams.
New advisory angles, grounded in real data
AI-powered insights on stock, margins and demand don’t replace your judgement – they give you more raw material for conversations about product mix, pricing and working capital.
Less reputational risk around integrations
Knowing there’s active monitoring and self-healing logic around key integrations gives you more confidence recommending a multi-app stack with Unleashed at the centre.
A vendor that’s actually shipping, not just shouting
Perhaps the biggest takeaway is cultural: these examples show a product organisation quietly using AI to shave friction off real workflows, not just announcing “we do AI now.”
Closing thoughts: beyond the buzz, back to the work
It’s easy to get lost in grand AI narratives – AGI, disruption, robots taking jobs.
What comes through in Ben’s conversation with Ryan is much simpler and, for most SMEs, much more useful:
- Get people to value faster.
- Close the gap between what customers say and what gets built.
- Keep the plumbing working.
- Surface the right information at the right time.
If AI can help Unleashed do those things better – from trials to day-to-day inventory decisions to keeping everything connected – that’s the kind of “AI strategy” most product businesses can get behind.
