# AI Productivity Impact: What Controlled Experiments and Field Studies Actually Show

> Experiments put the AI productivity impact at 12–56% depending on the task. Learn who gains most, why it rarely reaches profit and how to measure it.

Source: https://digitalbridge.com.tr/en/blog/ai-productivity-impact-research.html
Publisher: Digital Bridge Teknoloji (https://digitalbridge.com.tr)
Topic: Artificial Intelligence
Published: 2026-09-25 · Updated: 2026-10-22

AI productivity impact is the gain measured when employees use AI tools to do the same task faster, in greater volume or better. In controlled experiments, customer support agents resolved 14% more issues per hour and developers completed 26% more tasks. Gains depend on the task and worker experience, and reach profit only when workflows are redesigned.

## Key takeaways

- In controlled experiments, the AI productivity impact ranged from 12% to 56% depending on the task; there is no single universal figure.
- In NBER's study of 5,179 support agents, the gain reached 34% for novices but was minimal for experienced agents.
- In the 2023 HBS–BCG experiment, AI users were 19 percentage points less likely to be correct on a task beyond AI's frontier.
- According to St. Louis Fed researchers, US workers using generative AI saved an average of 2.2 hours a week in November 2024.
- Individual gains reach profit only when freed time is redirected to new work and compared against a measured baseline.

## The question on every leadership agenda

Some of your staff are almost certainly using AI already, to draft emails, summarise documents or fix code with [generative AI](https://digitalbridge.com.tr/en/glossary.html#generative-ai) tools. The question that reaches the board is always the same: is this producing measurable productivity, or is it just a habit that feels useful? Vendor slides full of confident percentages rarely answer it, because they do not say what was measured, for whom, or under what conditions.

This article starts with **primary field and experimental research** rather than opinion surveys: studies with a control group, a defined sample and a defined task. We then look at the evidence on why those individual gains so often disappear at company level. Calculating return on investment is a separate exercise, which we cover step by step in our guide to [measuring AI project ROI](https://digitalbridge.com.tr/en/blog/measuring-ai-project-roi.html).

## What do controlled experiments show about AI productivity impact?

As of 2026, the strongest evidence comes from experiments in which two groups do the same work and differ only in access to AI. The table below sets five of the most prominent studies side by side. Each row applies to one task and one sample; reading any single percentage as "AI raises productivity by X" would be a mistake.

| Study | Who, which task | Measured result |
|---|---|---|
| [NBER — Generative AI at Work](https://www.nber.org/papers/w31161) | 5,179 customer support agents | Issues resolved per hour up 14% on average |
| [GitHub Copilot experiment, arXiv (2023)](https://arxiv.org/abs/2302.06590) | 95 professional developers writing an HTTP server | Task finished 55.8% faster (71 vs 161 minutes) |
| [MIT — Cui et al., three field experiments](https://economics.mit.edu/sites/default/files/inline-files/draft_copilot_experiments.pdf) | 4,867 developers at three companies | Completed tasks up 26.08% |
| [HBS–BCG experiment (2023)](https://www.hbs.edu/ris/Publication%20Files/24-013_d9b45b68-9e74-42d6-a1c6-c72fb70c7282.pdf) | BCG consultants, 18 realistic tasks | 12.2% more tasks, 25.1% faster, over 40% higher quality |
| [MIT — Noy and Zhang (2023)](https://economics.mit.edu/sites/default/files/inline-files/Noy_Zhang_1_0.pdf) | 444 professionals, workplace writing | Time down 37%, grades up 0.45 standard deviations |

Two patterns stand out in the table of controlled experiments. Single-task, lab-style experiments (Copilot, writing) show large speed gains, while months-long field studies in real workplaces (NBER, Cui et al.) show more modest but still meaningful effects. And none of the studies measures "the whole job"; each measures one specific, repeated task.

## Who benefits most from AI, and where does the effect reverse?

The most consistent finding of the AI productivity experiments is that gains are unevenly spread. In the [NBER study](https://www.nber.org/papers/w31161) the improvement reached 34% for novice and low-skilled agents, with minimal impact on experienced ones. In the [HBS–BCG experiment](https://www.hbs.edu/ris/Publication%20Files/24-013_d9b45b68-9e74-42d6-a1c6-c72fb70c7282.pdf) consultants below the average improved by 43% and those above it by 17%. AI acts as a way of passing the know-how of experienced staff on to newcomers, which is why the split of tasks between people and tools needs deliberate design, as we discuss in [human-AI collaboration at work](https://digitalbridge.com.tr/en/blog/human-ai-collaboration-at-work.html).

The other side of the HBS–BCG experiment gets less attention. On a task deliberately chosen to lie outside AI's capability frontier, consultants using AI were 19 percentage points less likely to reach the correct answer than those working without it. So the tool does not speed up every job; on some, it lowers quality with confident but wrong answers. We look at how to reduce that risk in our article on [reducing AI hallucinations](https://digitalbridge.com.tr/en/blog/reducing-ai-hallucinations.html).

The practical lesson from these findings is to look for early gains in knowledge-heavy tasks with long onboarding curves. In customer service, that means an assistant that suggests replies and summarises cases for the human agent, described in our piece on [agent assist for customer service](https://digitalbridge.com.tr/en/blog/agent-assist-ai-customer-service.html).

## Why don't individual AI gains reach the bottom line?

The speed-ups seen in controlled experiments do not show up on the balance sheet to the same degree. Time savings at worker level are real but limited: in a [November 2024 survey analysed by St. Louis Fed researchers](https://www.stlouisfed.org/on-the-economy/2025/feb/impact-generative-ai-work-productivity), US workers using generative AI saved an average of 5.4% of their hours, or 2.2 hours in a 40-hour week. Across all workers, including non-users, the figure falls to 1.4% of total hours.

> In McKinsey's "The state of AI in 2025" survey, only 39% of respondents attributed any level of enterprise EBIT impact to AI, and most of them said less than 5% of EBIT was attributable to AI use. ([McKinsey — The state of AI in 2025](https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai))

An [NBER study by Humlum and Vestergaard](https://www.nber.org/papers/w33777), linking adoption surveys to Danish administrative records, rules out effects larger than 2% on earnings and recorded hours two years after ChatGPT's launch; what changes is how tasks are reorganised. The executive view is similar: [IBM's 2025 CEO study](https://newsroom.ibm.com/2025-05-06-ibm-study-ceos-double-down-on-ai-while-navigating-enterprise-hurdles) found only 25% of AI initiatives delivered the expected ROI. And in [Microsoft and LinkedIn's 2024 Work Trend Index](https://www.microsoft.com/en-us/worklab/work-trend-index/ai-at-work-is-here-now-comes-the-hard-part), 59% of leaders worried about quantifying AI's productivity gains.

The reason for the gap between experiment and balance sheet is simple: two hours saved by one employee vanish unless the workflow routes them to other work. In the same McKinsey survey, the roughly 6% of "high performers" attributing 5% or more of EBIT to AI were nearly three times as likely as others to have fundamentally redesigned their workflows. BCG's [2024 research](https://www.bcg.com/press/24october2024-ai-adoption-in-2024-74-of-companies-struggle-to-achieve-and-scale-value) likewise reports that AI leaders put 70% of their resources into people and processes.

## How do you measure AI productivity impact? Six steps from the research

The design of the AI productivity experiments offers a method you can reuse internally. The steps below are not a spreadsheet; they are a framework for setting up the measurement:

1. **Define the task narrowly.** Not "speed up the sales team" but "write the first draft of a quotation email": a task with a clear start and end. Every controlled experiment worked this way.
2. **Measure the baseline.** Before handing out the tool, record time per task and error or rework rates for a few weeks. Without a baseline, any gain you claim later is open to dispute.
3. **Set up a comparison group.** Ideally one team uses the tool while another doing the same work does not; failing that, compare the same team before and after.
4. **Measure quality as well as speed.** The HBS–BCG experiment showed speed can cost accuracy on tasks outside AI's frontier. Sample outputs and have a person score them.
5. **Split results by experience.** Gains may be large for newcomers and small for experts; an average on its own misleads.
6. **Plan where the saved time goes.** If freed hours are not directed to backlog, more customers or quality checks, they will never appear in the accounts.

Use this checklist to choose which task to start with:

| Task characteristic | Gains likely | Needs care |
|---|---|---|
| Type of output | Draft, summary, classification, translation | Calculation or judgement with one right answer |
| Verification | A person can check it quickly | Spotting errors requires deep expertise |
| Knowledge source | Company documents are accessible | Information is scattered or out of date |
| User | Newcomer, someone searching for information | Expert who already does it very quickly |
| Volume | Repeated dozens of times a day | Done a few times a month |

If the task follows fixed rules, you may not need AI at all; we discuss that distinction in [AI versus rule-based automation](https://digitalbridge.com.tr/en/blog/ai-vs-rule-based-automation.html). To score and rank candidate tasks department by department, use our [AI use case prioritisation matrix](https://digitalbridge.com.tr/en/blog/ai-use-case-prioritisation-matrix.html).

## What do public company case studies add?

Beyond the controlled experiments, companies publish their own results; these are not independently audited and often rest on user self-reports. According to [a customer story Microsoft published in December 2024](https://www.microsoft.com/en/customers/story/19346-vodafone-microsoft-365-copilot), participants in a 300-person Copilot trial at Vodafone reported saving four hours per person per week, and the company is extending use to 68,000 employees.

A narrower, more measurable example comes from manufacturing. In a Microsoft customer story from November 2024, [Eaton says](https://customers.microsoft.com/en-us/story/1837633288515694884-eaton-microsoft-365-copilot-discrete-manufacturing-en-united-states) that across 1,000 standard operating procedures drafted with Copilot's help, time per procedure dropped from one hour to ten minutes (an 83% reduction). Both cases echo the experimental lesson: the gain is clearest on a repeated writing task with a known format.

## How we approach this at Digital Bridge

We start with measurement, not productivity claims. In the discovery phase of our [digital transformation consultancy](https://digitalbridge.com.tr/en/consultancy/digital-transformation-consultancy.html), we map your teams' repeated tasks, the time they take and where the knowledge behind them sits. The thinking behind that needs analysis is set out in our article on [software requirements analysis](https://digitalbridge.com.tr/en/blog/software-requirements-analysis.html).

We then design an [AI pilot project](https://digitalbridge.com.tr/en/glossary.html#ai-pilot-project) around a single task that fits the task-selection checklist. For knowledge-heavy work we build an [enterprise LLM assistant](https://digitalbridge.com.tr/en/technology-hardware/enterprise-llm-assistant.html) that answers from your own documents and cites its sources; the technical basis is explained in [RAG for enterprise LLMs](https://digitalbridge.com.tr/en/blog/rag-enterprise-llm.html). The baseline and comparison method are agreed in writing before the tool goes live.

So the gain is not lost, we place the tool inside the screens people already use, such as ERP, CRM or email, through [AI integration](https://digitalbridge.com.tr/en/technology-hardware/ai-integration.html) and, where needed, [system integrations](https://digitalbridge.com.tr/en/technology-hardware/system-integrations.html). Where part of the task is rule-based, we handle that step with [RPA process automation](https://digitalbridge.com.tr/en/software-solutions/rpa-process-automation.html) and keep AI for the steps that need judgement. We also recommend a clear [company AI acceptable use policy](https://digitalbridge.com.tr/en/blog/company-ai-acceptable-use-policy.html) so staff know which tools they may use with which data.

Where writing, summarising and document search are concentrated in email and file screens, our in-house [Smart360](https://digitalbridge.com.tr/en/smart360.html) family builds that placement in: [SmartMail](https://digitalbridge.com.tr/en/smart360/smartmail.html) handles message summaries, classification and translation in the inbox, and [SmartFiles](https://digitalbridge.com.tr/en/smart360/smartfiles.html) runs analysis on upload and lets people ask questions of a document in the file panel. In both products AI is a separate permission granted person by person, and usage is tracked against the organisation's monthly allowance in the admin panel, which makes it easier to limit a pilot group and see actual use.

At the end of the pilot we report plainly which task and which group of employees saw gains, and where they did not. If you decide to scale, we tie that step into your [digital transformation roadmap](https://digitalbridge.com.tr/en/blog/digital-transformation-roadmap.html). For the general order of play, read [where to start with AI in business](https://digitalbridge.com.tr/en/blog/ai-in-business-where-to-start.html).

## Next step

This week, list three writing, summarising or information-search tasks that your teams repeat at least ten times a day, and note roughly how long each takes today. That list is the starting point for a measurable pilot. Send it to us through our [contact page](https://digitalbridge.com.tr/en/contact.html) and we will use the discovery call to agree which task to test first. For more on the topic, browse our [artificial intelligence articles](https://digitalbridge.com.tr/en/blog/topic/yapay-zeka.html).

## Frequently Asked Questions

### How much does AI increase productivity?

There is no single figure. In controlled experiments the result varied by task: customer support agents resolved 14% more issues per hour, developers across three companies completed about 26% more tasks, and in a 2023 single-task coding experiment speed rose by more than 55%. These numbers belong to specific tasks and samples; only a pilot with a proper baseline shows the effect in your own company.

### Which employees benefit most from AI?

In the experiments, the largest gains went to newcomers and below-average performers. In the customer support study, novice agents improved by 34% while experienced agents saw minimal change. That makes AI a strong tool for shortening onboarding and spreading the know-how of experienced staff across a team, rather than a way to make top performers dramatically faster.

### Does AI improve productivity on every task?

No. In the 2023 Harvard and BCG experiment with consultants, on a task chosen to lie outside AI's capability frontier, those using AI were 19 percentage points less likely to reach the correct answer than those without it. For tasks with one right answer, where errors are hard to spot, human review and a system that cites its sources are essential.

### If employees save time, why doesn't it show up in profit?

Saved time is lost unless it is redirected to other work. Surveys show that companies reporting meaningful profit impact from AI are a minority, and those capturing value tend to have redesigned their workflows. Rolling out a tool is not enough; you need to plan in advance where freed hours will go and how the result will be measured.

### How do you measure AI's productivity impact in your own company?

You do not need a large dataset. For one narrowly defined task, a few weeks of baseline timings, an error or rework rate and, ideally, a comparison group that does not use the tool are enough. Scoring a sample of outputs for quality, not just speed, and splitting the results by experience level stops a single average from misleading you.

### How much time does AI save employees each week?

A few hours on average, according to US data. In a November 2024 survey analysed by St. Louis Fed researchers, workers using generative AI saved an average of 5.4% of their working hours, or about 2.2 hours in a 40-hour week. Across all workers, including non-users, the saving falls to 1.4% of total hours: real, but limited.

### Which task should an AI pilot start with?

The best candidate is a task repeated dozens of times a day, whose output is a draft, summary, classification or translation, and which a person can check quickly. Accessible company documents and a task done by newcomers raise the odds of a gain. Tasks with one right answer, errors that need expertise to spot, or a frequency of a few times a month are poor first pilots.