The Problem
Small businesses don't lose people more often. They lose more when they do.
In a ten-person firm, one resignation takes a tenth of the team with it. There's no bench and no second person who knows the client, the process or where the files are. A large company absorbs a departure. A small one feels it in every client it serves.
Replacing that person is slow. From resignation to a new hire working at full speed takes about five to eight months, by our estimate.1 In Singapore, 16% of professional and managerial vacancies stay open for six months or more.2 And in firms of 3 to 30 people, losing a worker costs one to two years of their salary to replace.3
Absence hurts too. Nearly half of Singapore businesses say they can't spare people even to send them for training.4
5–8 months
from a resignation to a replacement working at full speed. Our estimate for an experienced hire from the same sector.
Agent Hero estimate for an experienced, same-sector hire, built from notice periods, hiring times (Ashby) and ramp-up studies (Oxford Economics). Cross-sector hires take 10–13+ months.
| Phase | Typical length |
|---|---|
| Leaver works their notice | 1–3 months |
| Find and hire a replacement | 6–9 weeks |
| New hire works their notice | about 1 month |
| New hire ramps up to full speed | 12–15 weeks |
The AI Gap
Large companies are pulling ahead on AI. Small ones are falling behind.
In 2024, 14.5% of Singapore SMEs used AI, against 62.5% of larger firms. Both grew, but the gap widened from 40 points to 48 in a single year.5 Europe shows the same pattern: 17% of small firms use AI, against 55% of large ones.6
The gap runs deeper than the headline. SMEs that use AI apply it to three parts of the business; larger firms apply it to five.5 In the US, only 7% of small firms using AI have fully built it into how they work.7
Most small businesses have tried AI. Very few have made it part of the business.
14.5%
of Singapore SMEs used AI in 2024, against 62.5% of larger firms. The gap grew from 40 to 48 points in a year.
IMDA, Singapore Digital Economy Report 2025.
| Year | SMEs | Larger firms | Gap (points) |
|---|---|---|---|
| 2023 | 4.2% | 44% | 40 |
| 2024 | 14.5% | 62.5% | 48 |
The Hidden Risk
The AI your team uses belongs to them, not to your business.
Most AI comes into a small business through its staff. 84% of Singapore AI users bring their own tools to work,8 and worldwide, 70% of employees who use AI rely on free public tools.9 The prompts, shortcuts and workflows they build live in personal accounts the company never sees.
Paying for business plans doesn't fix it. On Claude's Team plan, for example, colleagues lose access to a removed member's chats, and skills that person never shared stay with them.10
So a small firm's AI often rests on one or two people. Six in ten Singapore SMEs using AI say losing their main AI person would disrupt the business.11 When that person leaves, much of what the firm learned leaves with them.
84%
of Singapore AI users bring their own AI tools to work. What they build stays in their accounts, not the company's.
Microsoft & LinkedIn, 2024 Work Trend Index, Singapore.
| Group | Share |
|---|---|
| Bring their own AI tools | 84% |
| Don't | 16% |
Why It's Still Unsolved
AI got cheap. Making it work for a small team didn't.
Price is no longer the wall. Business AI costs about US$20 a person a month, and the large seat minimums are gone.12 Yet among small European firms that considered AI and passed, 71% cited a lack of expertise, against 39% who cited cost.6
Today's tools assume someone in-house will set them up, connect them to the firm's systems, write the instructions and keep them working. Large companies have teams for that. A twelve-person accounting firm doesn't. And agents still need people to guide them: the best one completes only about 30% of realistic office tasks on its own.13
So small businesses get a chat window, not a workforce.
71%
of small firms that considered AI and held back blamed a lack of expertise. Only 39% blamed the cost.
Eurostat, EU firms with 10–49 staff, 2025. Firms could give more than one reason.
| Reason | Share of firms |
|---|---|
| Lack of expertise | 71% |
| Doesn't fit existing systems | 42% |
| Cost | 39% |
Our Solution
An AI workforce your business owns. It learns your work and stays when people leave.
Agent Hero gives small and medium businesses AI that belongs to the company, not to any one employee. We set up AI executives for each department and connect them to the tools you already use. Your staff hand them work and guide them, the way they would a new colleague.
Every task teaches them more about how your business runs: your clients, your processes, every correction your staff make. What they learn is kept as skills the whole company shares. When someone is on leave, the work keeps moving. When someone leaves, what they taught the AI stays.
It helps the people who join, too. In one large study, support agents with an AI assistant performed at two months the way their colleagues without one did at six.14 With Agent Hero, a new hire starts with the company's know-how on day one.
2 months
for new support agents with an AI assistant to perform like six-month veterans without one.
Illustrative shape of the finding in Brynjolfsson, Li & Raymond, Generative AI at Work (QJE, 2025): support agents with AI at two months matched agents without it at six.
| Group | Months to reach the six-month level |
|---|---|
| With an AI assistant | about 2 |
| Without | 6 |
Sources
1. Oxford Economics for Unum, The Cost of Brain Drain (2014). UK survey of 500+ firms on time to full productivity; with Ashby, Startup Hiring Report, for time to hire at firms under 25 staff. Agent Hero's 5–8 month figure is an estimate built from these and Singapore notice rules.2. Ministry of Manpower, Job Vacancies 2025. Singapore establishments with 25 or more employees.3. Jäger & Heining, How Substitutable Are Workers? Evidence from Worker Deaths, NBER Working Paper 30629. German social-security records.4. Singapore Business Federation, National Business Survey 2025: Manpower & Wages. 555 businesses, 85% of them SMEs.5. IMDA, Singapore Digital Economy Report 2025. AI adoption by firm size, 2023–2024, and a pulse survey of about 500 AI-using firms.6. Eurostat, Use of artificial intelligence in enterprises (isoc_eb_ai), 2025. EU firms with 10–49 staff (small) and 250+ (large).7. Federal Reserve Banks, 2026 Small Business Credit Survey: Report on Employer Firms. US small employer firms, surveyed in 2025.8. Microsoft & LinkedIn, 2024 Work Trend Index: Singapore (May 2024).9. KPMG & University of Melbourne, Trust, Attitudes and Use of AI: A Global Study 2025. About 48,000 people in 47 countries, Singapore included.10. Anthropic Help Center, What happens to a user's data when they are removed from a Team or Enterprise organization.11. AWS & Strand Partners, Singapore SME AI maturity research (May 2026). 1,500 SMEs in financial services, manufacturing and healthcare.12. Microsoft, Bringing the full power of Copilot to more people and businesses (15 January 2024), which removed the 300-seat minimum; and Claude pricing, Team plan for 2 to 150 seats.13. Xu et al., TheAgentCompany: Benchmarking LLM Agents on Consequential Real World Tasks (NeurIPS 2025).14. Brynjolfsson, Li & Raymond, Generative AI at Work, Quarterly Journal of Economics (2025). 5,179 customer-support agents at one firm.
People change. The know-how stays.
