Why 2026 and 2027 are the best time to transform.

01Context

The way business runs is changing.

Every technology wave has a moment when it stops being seminar material and starts becoming daily work. For AI, that moment has already passed at the individual level. Generative AI reached about 53% population adoption in just three years, faster than PCs or the internet⁶.

Indonesia is among the fastest markets for AI use. The e-Conomy SEA 2025 report notes that revenue from AI-based apps in Indonesia grew 127% between the first half of 2024 and the first half of 2025, the highest in Southeast Asia⁵. The foundation is already there: 80.66% of the population, about 229 million people, is online⁷. In offices, 92% of knowledge workers already use generative AI, above the global average of 75%¹.

But using AI is not the same as transforming. Research by AWS and Strand Partners found that 28% of businesses in Indonesia have already adopted AI. Yet 76% still focus on basic use cases, and only 10% of AI-using businesses have reached the stage where AI is part of core decisions and the business model². At the leadership level, 48% worry their organization still lacks a vision and plan for applying AI¹.

Your team is probably already using AI. The real question: has your company's way of working changed with it?

That is the race. Over the next few years, what separates winners will not be who understands AI. Almost everyone does. What separates them is who first rewires processes, data, and team habits around these new tools.

02Opportunity

Repetitive work that quietly eats your team's week.

Try counting how many hours each week your team spends retyping data from one system into another, compiling reports from several spreadsheets, chasing confirmations over chat, or answering the same customer questions. In almost every business we meet, the number is larger than people assume.

McKinsey estimates that generative AI, together with other technologies, could technically automate activities that absorb 60 to 70% of employee time today³. That is technical potential, not a promise. But the direction is clear: the share of work machines can help with is far larger than a few years ago.

Evidence on the ground is starting to accumulate. Studies summarized in the Stanford AI Index 2026 show productivity gains of 14 to 26% in customer support and software development. In some hospital systems, doctors using automatic clinical documentation tools reported note-writing time falling by up to 83%⁶.

What matters: work like this sits mostly in daily operations, not only in big programs. And often the data already exists. It is just scattered across spreadsheets, shared drives, messaging groups, or the systems teams already use every day.

The goal is not to replace people. The goal is to give time back to work that needs humans: serving customers, making decisions, improving quality. Processes move much faster, the team stays lean, and energy shifts to what matters.

03Timing

Why now, not later.

It is fair to ask: why not wait until the technology is more mature? There are four reasons why 2026 and 2027 are the window that makes the most sense.

3.1Costs have already fallen sharply.

The drop did not stop in 2024. Epoch AI estimates that since 2023 the cost of reaching a given level of AI performance has fallen about 47% per quarter, or about 13 times per year⁹. One concrete example: when OpenAI released o3 in January 2025, Epoch estimates it cost about US$0.30 per question to hit a 75% score on GPQA Diamond, a PhD-level science exam. Less than 18 months later, GPT-5.6 Luna reached the same score for about US$0.0004 per question: a 725-fold drop⁹. Earlier, Stanford's AI Index had already documented a more than 280-fold fall in the cost of GPT-3.5-level inference, from US$20 to US$0.07 per million tokens between November 2022 and October 2024¹⁰. Capability that once sat only with global tech budgets is now far more reachable for organizations that already run day-to-day operating processes.

Capability rose with the price drop. On OSWorld, which tests agents on real computer tasks across operating systems, success rose from roughly 12% to 66.3% in 2025, though agents still fail about one in three attempts on structured benchmarks⁶. The tools are not perfect, but they are mature enough for administrative work with clear rules, with humans still checking the output.

3.2Early-mover advantage compounds.

BCG studied more than 1,250 companies and found that only 5% are already generating AI value at scale. Those firms report five times the revenue increase and three times the cost reduction of others. Meanwhile 60% get almost no meaningful value, even though they have invested¹¹.

The pattern looks like compound interest. Those who start earlier get results, reinvest those results, and pull further ahead. BCG calls it a widening gap¹¹.

Larger players are moving too. McKinsey notes that nearly half of companies with more than US$5 billion in revenue have reached the AI scaling phase, compared with 29% of those with less than US$100 million¹². In Indonesia, 41% of large enterprises already use AI, but only 22% of those AI adopters have a comprehensive AI strategy². Waiting means letting that distance grow.

3.3Habits and talent take time.

Software can be installed in a day. Team habits cannot. The WEF estimates that 39% of workers' core skills will change by 2030. 63% of employers cite skills gaps as the biggest barrier to transformation, followed by organizational culture and resistance to change (46%)¹³.

In Indonesia the pressure is sharper. The Ministry of Communication and Digital Affairs estimates a need for 9 million digital talents by 2030, while current capacity is only about 100 to 200 thousand a year¹⁴. No surprise that 57% of businesses in Indonesia name the digital skills gap as the main barrier to expanding AI use².

That means AI talent cannot simply be bought on the open market. It has to be grown inside the company, through real work. That process cannot be rushed in one quarter. Starting in 2026 means your team is already fluent when competition demands more speed.

3.4The cost of waiting keeps rising.

Meanwhile, the cost of running the old way keeps going up. Most provinces raised the 2026 provincial minimum wage by 5 to 7%¹⁵. For organizations with many operating roles, that flows into real cost.

The upside of moving is also clearer. Among Indonesian businesses that have adopted AI, 59% reported revenue gains averaging 16%, and 64% expect average cost savings of 29%². Every month of delay is another month of paying people to do work a machine could handle.

04Differentiator

Winners are decided by how you work, not the tools.

If AI is already this cheap, why do 60% of companies still get almost no meaningful value? The answer is consistent across studies: the problem is rarely the technology. It is how people work.

The Work Trend Index 2026 finds that organizational factors such as culture, manager support, and talent practices explain more than twice the AI impact of individual factors⁴. McKinsey finds that AI high performers are nearly three times as likely to have fundamentally redesigned workflows¹².

There is also a paradox to watch. 65% of AI users fear falling behind if they do not adapt quickly, yet 45% feel safer focusing on current targets than redesigning the work. Only 13% feel rewarded for trying new ways of working⁴. Without a push from leadership, teams will choose the safe option.

Transformation is not buying new software. Transformation is deciding which work no longer needs a human, then reorganizing the team around that.

This is good news. You do not need a giant transformation budget to redesign one time-consuming workflow. What you need is clarity: which process is most painful, what data you already have, and what outcome you want. Then a team that can turn that into a tool people actually use.

The impact also shows up in people. PwC finds that industries most able to use AI saw three times higher growth in revenue per employee, and workers with AI skills command an average wage premium of 56%¹⁶. The IMF estimates that about 40% of jobs in emerging markets like Indonesia are exposed to AI, and some of those can benefit through higher productivity¹⁷. Teams that learn to work with AI become more valuable teams.

05Roadmap

Start with one process.

Successful transformation rarely starts with a hundred-page master plan. It usually starts with the most painful process, gets that working until the result is felt, then expands.

1. 01
   Week 0-1

   Talk and diagnose

   Map the repetitive work, measure the time it takes today, and check the data you already have. The output: one or two priority processes with a clear target.

2. 02
   Week 2-4

   Pilot with your own data

   A first version built and used by the real team: an app, portal, dashboard, or automation. Not a demo. Not slides.

3. 03
   During the pilot

   Measure and improve

   Compare with the baseline: hours returned, fewer errors, reports that land faster. Improve from what the team tells you.

4. 04
   Next

   Expand to the next process

   Once one process is running, move to the next. New habits grow from success the team can see for itself.

Five questions for your next board meeting.

• What work does our team repeat every week?
• How many hours does it take, and who does it?
• What data do we already have that is not being used?
• If one process were fixed in a month, what would that mean for our customers?
• Who on the team is most ready to drive the change?

06Partner

You do not need a big consultancy to start.

Many directors delay because they picture transformation as a large consultancy, months of project work, and a thick report that ends up in a drawer. That is not what is needed first.

What is needed is a partner who understands your business and builds the tool. That is what we do at Akses.

We already run this way ourselves. Aksesmedika grew out of labs and clinics drowning in MCU work: data retyped over and over, conclusions written from scratch, client HR asking one by one. AksesData grew out of leaders waiting too long for reports. Both started with one repetitive job, then were built around the way of working that was already there.

AksesmedikaMCU platform for labs and clinics, with a dashboard for client HR. AksesDataLeadership reports from the data you already have.

Products are just examples of how we work. What we offer is the outcome, for your team's own needs. And we will be honest when something is outside what we can help with. See what we can do.

07Closing

This wave belongs to those who move.

If you lead an organization in Indonesia and want to move, this paper is for you. Start with one process, use the data you already have, and help your team build new habits. What matters is clear intent.

If more organizations can work faster with the same team, the impact shows up in business results and in competitiveness across industries that employ millions. That is a part of the contribution we want to make.

The 2026 and 2027 window will not stay open forever. The tools are affordable, those who moved earlier are already scaling, and your teams need time to adapt. Those who move now will help decide what their industry looks like a few years from now.