While large global enterprises are commanding attention and headlines with big AI investments, the most immediate returns are being seen from a different corner of the market.
Small to midsize businesses (SMBs) are deploying AI with speed, focus and intent, often seeing measurable impact within weeks rather than months.
European startups developing AI agents that can be deployed across workflows have already raised €1bn this year, according to Sifted data.
Smaller companies are seeing faster and more valuable returns by deploying pre-built AI, according to Samantha Wessels, president of EMEA at intelligent content management platform Box.
“These companies achieve faster time-to-value by skipping the infrastructure phase and deploying turnkey AI to solve specific resource constraints immediately,” she says.
When using AI at work, SMBs often move from pilot to production faster than their larger counterparts, which often suffer from so-called “data gravity”, says Wessels.
“Large enterprises are bottlenecked by decades of fragmented content across disconnected systems, putting them in an AI prototype purgatory with endless pilots that never reach production," she says.
Enterprises run broader transformation programmes, whereas SMBs can scope tightly and scale by starting with one workflow, proving value and then expanding from there.
Smaller businesses sort through data faster by deploying existing digital-native stacks such as Slack and Salesforce within their workflows, while larger companies often spend too much time trying to build their own platforms.
Traditionally, complex organisational structures also slow down AI adoption within larger companies, whereas smaller businesses with a flatter structure often have the agility to avoid these hurdles, says Wessels.
“Enterprises run broader transformation programmes, whereas SMBs can scope tightly and scale by starting with one workflow, proving value and then expanding from there,” she says.
The companies that are able to get ahead with AI are those redesigning workflows and for SMBs, who are able to remain agile, this is often easier.
For larger enterprises, this may mean moving past incremental pilot projects and engineering a complete overhaul of business processes.
In order to transform work processes, businesses should consider starting by giving AI repeated tasks that often take up a lot of time. SMBs have seen quick wins by automating these kinds of tasks across finance processing, HR onboarding and sales.
Live entertainment producer RWS Global is one company showing how an SMB can integrate AI effectively, according to Wessels. It works with thousands of performers on a contractual basis and has managed to automate contract approval workflows to save time on manual processing.
The company didn’t build custom infrastructure but instead leveraged Box’s existing tools to automate approvals and processing time. Box’s ability to centralise digital asset libraries meant teams at RWS Global could find, collaborate on and distribute content without switching between disconnected systems
Getting employees acquainted with using AI is key to transforming workflows. Companies, whether big or small, should work to align the views of the C-suite with the realities of those lower down in the company.
Box adheres to a four-phase approach and encourages companies to use the same framework. The four stages are ideation (identify AI opportunities by examining pain points and inefficiencies), pilots (build and test a set of agents over three to six months), rollout (transform successful pilots into solutions) and scaled adoption (maximise adoption through change management and workflow redesign).
Guardrails
But moving fast doesn’t mean moving recklessly. SMBs should balance speed with responsible deployment by putting guardrails in place from day one.
AI tools must also provide citations linking to source files so humans can easily verify outputs.
Core day-one guardrails include zero-training policies (AI providers that don’t use customer prompts or data to update their base models), permissions-based AI (AI systems that enforce strict data access controls and limitations), and citation and transparency, Wessels says.
“AI must strictly adhere to existing user access controls, preventing internal leaks by only analysing files the user can already view,” she says. “AI tools must also provide citations linking to source files so humans can easily verify outputs.”
There should also always be a “human-in-the-loop,” Wessels adds. The technology should source the information and summarise it, but humans should always verify and make any final decisions.
The competitive advantage is shifting from ‘who has the most computing power’ to ‘who can adapt workflows fastest’.
In the next five to ten years, Wessels believes the AI gap between SMBs and enterprises will narrow significantly. Workplaces are adapting from simply using agentic AI as an assistant to using it as a colleague.
“Agentic AI that actively executes work rather than just answering questions are being natively embedded into everyday SaaS platforms,” she says.
“The competitive advantage is shifting from who has the most computing power to who can adapt workflows fastest. This means SMBs, unburdened by rigid change management, will effortlessly inherit enterprise-class intelligence and operate at massive scale.”
Click here to learn more about how your business can adopt AI with Box.
