Research and analysis

The impact of technology diffusions on growth and productivity: findings from a human-only rapid evidence review

Published 23 April 2025

Note:

This research was supported by the R&D Science and Analysis Programme at the Department for Culture, Media & Sport. It was developed and produced according to the Behavioural Insights Team’s research team’s hypotheses and methods. Any primary research or findings do not represent government views or policy. Please note that this research was commissioned under the previous government of 11 May 2010 to 5 July 2024, and before the founding of the UK Metascience Unit.

1. Executive summary

This rapid evidence review examines the impact of technological diffusions on growth and productivity. It places special focus on identifying the mechanisms that drive these changes, and on deriving UK-specific conclusions based on the findings. The review is not systematic – it prioritised the identification of significant research findings over an exhaustive review of the literature. Papers were identified using specific search terms on Google and Google Scholar, and then refined down through application of detailed inclusion/exclusion criteria.

This review was commissioned by the Department for Science, Innovation and Technology (DSIT) and the Department for Culture, Media & Sport (DCMS) through the R&D Science and Analysis Programme. The choice of technologies under review was informed by DSIT’s Science and Technology Framework, which emphasises the importance to future UK success of technologies such as artificial intelligence, robotics, and broadband internet.

Overall, the review finds a generally positive, albeit varied, impact of these technologies on firm growth and productivity. It underscores the importance of contextual factors and firm characteristics in leveraging their full potential, and identifies 6 key mechanisms which determine the extent to which firms benefit from them:

• Complementary intangible firm assets are crucial to unlock the benefits of new technologies.
• Firms with more in-house expertise and human capital reap more benefits from new technologies.
• Larger firms with access to more resources are better able to absorb the high costs of technological investments.
• Young firms that are not already reliant on existing technologies are more likely to capitalise on the arrival of new ones.
• Infrastructure and regulatory environments influence the ease and incentives to adopt new technologies.
• Whether firms unlock the benefits from new technologies depends on both time and timing.

Based on these findings, we produce the following conclusions:

• Financial incentives and support, especially for smaller firms, could help businesses to adopt and commercialise new technology. One barrier to investing in new technologies is cost. Given the finding that larger firms with more resources are better positioned to absorb these kinds of costs, SMEs may require special attention to help overcome initial cost barriers and facilitate their adoption of productivity-enhancing technologies. Optimised design of existing support measures, such as R&D tax relief, through the use of timely prompts and targeted campaigns, may help achieve this.
• Testing and evaluating ways to improve take-up can help upgrade existing business management and tech support schemes. In recent years, the UK government provided two significant programmes to encourage SMEs to boost their productivity and growth: ‘Help to Grow - Digital’ and ‘Help to Grow - Management’. The former has now closed following poor take-up; the latter remains open but also continues to experience low take-up. Greater use of peer-to-peer networking could encourage take-up of this programme, which could include commissioned business support services incorporating peer components, and tailoring support schemes depending on firm size, sector and age.
• De-shrouding business-to-business (B2B) markets can improve the quality of complementary investments. Complementary investments (e.g. training, IT infrastructure upgrades) are essential to unlock the full value of new technologies. Being able to quickly and reliably access high-quality providers of these services could therefore help speed up the process of technology diffusion. The creation and dissemination of transparent reputation systems (akin to Tripadvisor) can enable quick, reliable assessment of B2B firms providing IT and management training services, as well as technology suppliers and firms who assist with the redesign of organisational processes.

Table 1: Glossary of terms

Term	Definition
Innovation	The creation and application of new knowledge to improve the world.[footnote 1] The OECD identifies four types of innovation: product, process, marketing and organisational.[footnote 2]
Technology	The application of scientific knowledge for practical purposes, especially in industry.
Technology diffusion	The process by which the use of an innovation spreads and grows.
Growth	An increase in business value over time, interpreted broadly as including an increase in business revenues, employee headcount and profits.
Productivity	A measure of outputs relative to inputs used. An increase in productivity results from increased output for the same or less input.
Labour productivity	A measure of economic performance that calculates the amount of goods and services produced per hour worked.
Mechanism (that affects technology diffusion)	The key element of an explanation, which depicts the driving force that generates a certain effect or outcome in a particular condition or setting.[footnote 3]
Information and Communication Technologies (ICTs)	All digital technologies that facilitate the capture, processing, storage, and exchange of information in various forms.
Instrumental variable (IV)	A statistical method used to estimate causal relationships by utilising variables that influence the predictor variable of interest but have no direct effect on the outcome measure.
Structural equation modelling (SEM)	A statistical technique that allows for the testing and estimation of complex relationships among observed and latent variables.
General-purpose technology (GPT)	Technologies that have a wide range of applications across different industries and sector, such as electricity, the internet, and artificial intelligence
Predictive analytics	Using historical and current data to predict future outcomes

2. Introduction

The UK has a long history of technological innovation – from the advances of the industrial revolution to modern leadership in Artificial Intelligence (AI) research. These innovations have in turn led to significant improvements in productivity and economic growth.

To help ensure the UK remains a leader in science and technology in the 21st century, under the previous Conservative government the Department for Science, Innovation and Technology (DSIT) published a Science and Technology Framework in March 2023, with the ultimate goal of ensuring that the UK becomes ‘the most innovative economy in the world.’ This framework identified a number of technologies considered important to future UK success, such as artificial intelligence, robotics, and broadband internet.

Purpose of this review and our research question

This report aims to answer the research question: What is the impact of recent technological diffusions on growth and productivity? However, rather than focusing on quantifying effects, we place a greater emphasis on identifying the underlying mechanisms that drive any impact. In other words, the focus is on the ‘how’ rather than the ‘what’ – isolating the mechanisms through which the diffusion of new technologies has improved growth and productivity in a UK context, rather than merely documenting the after-the-fact impact of these new technologies. Given the adoption of a technology is a necessary precursor to potential growth and productivity impacts, we also comment on mechanisms related to acquisition – although this is not the focus of this review.

The subsequent section, ‘Methodology’, describes the procedures and strategies employed in our search for evidence, highlighting both its advantages and limitations. The ‘Findings and discussion’ section presents a structured analysis, organised by distinct mechanisms. Lastly the Conclusions section covers the practical takeaways from this work.

3. Methodology

We approached this review as a rapid evidence review, which involves a rapid search of high-quality empirical research about a topic. Although all of the evidence is sourced and compiled in a structured way, rapid evidence reviews are different to systematic reviews. They do not seek to comprehensively summarise all the literature on a topic – they are a pragmatic attempt to quickly assess the most relevant studies.

3.1 Inclusion and exclusion criteria

To ensure the literature we found was robust and relevant, we applied the following inclusion and exclusion criteria. Note that unpublished government resources were also not considered for review.

Inclusion:

• Sources grounded in empirical evidence, (e.g. papers on historic, specific technology diffusions with verifiable underlying datasets) rather than theory (e.g. papers conducting econometric modelling exercises).

• Prioritise research published since the year 2000 (not before that, although individual exceptions may be made on a case-by-case basis).

• Include ‘grey’ literature, e.g. government reports which are not technically academic publications.

• Focus on impact at the business level (productivity = measured as outputs relative to resource used / inputs; growth = increase in business value over time, interpreted broadly as including an increase in business revenues, employee headcount and profits), within the UK (but where evidence is scarce, we will look to other regions that share similar socioeconomic characteristics to the UK e.g., US, France, Germany).

• Focus more on recent (late 20th century / 21st century) technological changes.

Exclusion:

• Non-English-language material.

• Studies examining multi-faceted technological / societal shifts rather than discrete technology changes.

• Macro-history texts which only cover the topic as part of a broader narrative.

3.2 List of prompts used

The papers were identified using specific search terms on Google and Google Scholar, and then through a snowball effect by searching for relevant references in the identified papers. The list of prompts used to search for relevant literature are listed in the table.

Technology	Prompt
General	The impact of technology diffusions on growth and productivity.
General (UK-focused)	What is the impact of technology diffusions on growth and productivity in the UK?
Artificial Intelligence (AI)	What is the impact of AI diffusion on growth and productivity in the UK?
Engineering biology	What is the impact of engineering biology diffusion on growth and productivity in the UK?
Semiconductors	What is the impact of semiconductors diffusion on growth and productivity in the UK? How semiconductors diffusion impact productivity in the UK?
Quantum technologies	What is the impact of quantum technologies diffusion on growth and productivity in the UK?
Wi-fi & 3G/4G/5G networks	What is the impact of Wi-fi & 3G/4G/5G networks diffusion on growth and productivity in the UK? What is the impact of Wi-fi diffusion on growth and productivity in the UK? What is the impact of 3G network diffusion on growth and productivity in the UK? What is the impact of 4G network diffusion on growth and productivity in the UK?
Internet & internet of things	What is the impact of internet diffusion on growth and productivity in the UK?
Smartphones	What is the impact of smartphones diffusion on growth and productivity in the UK?
Mobile phones	What is the impact of mobile phones diffusion on growth and productivity in the UK?
Cloud computing	What is the impact of cloud computing diffusion on growth and productivity in the UK?
3D printing, robotics & laser tech	What is the impact of 3D printing, robotics & laser technology diffusion on growth and productivity in the UK? What is the impact of 3D printing diffusion on growth and productivity in the UK? What is the impact of robotics diffusion on growth and productivity in the UK? What is the impact of laser technology diffusion on growth and productivity in the UK? How laser technology diffusion impacts UK growth and productivity?
GPS & aerial imagery (e.g. Agriculture 3.0 / precision farming)	What is the impact of GPS & aerial imagery diffusion on growth and productivity in the UK? How GPS & aerial imagery diffusion impacts UK growth and productivity?
Renewable energy (solar, wind, hydroelectric, biomass, geothermal, tidal, wave, hydrogen)	What is the impact of renewable energy diffusion on growth and productivity in the UK? How renewable energy diffusion impacts UK growth and productivity? How renewable energy diffusion impacts UK firms productivity?
Lithium-ion batteries (e.g. increase in capacity)	What is the impact of lithium-ion batteries diffusion on growth and productivity in the UK? How lithium-ion batteries diffusion impacts UK growth and productivity? How lithium-ion batteries diffusion impacts UK firms productivity?

3.3 Screening process

Application of the search terms produced an initial longlist of 77 studies. Deeper scrutiny of these studies and systematic application of the inclusion/exclusion criteria then reduced this longlist to a shortlist of 20 studies, which made up the final sample that was used in the analysis stage.

Breakdown of final sample	Number of studies
Final shortlisted studies	20
Published in academic journals	16 (80%)
Published elsewhere	4 (20%)
Published 2018-24	16 (80%)
Published 2010-17	3 (15%)
Published 2000-09	1 (5%)
UK focused[footnote 4]	4 (20%)
Not UK focused	16 (80%)
Examining impact of technology on productivity	13 (65%)
Examining impact of technology on growth	5 (25%)
Examining impact of technology on both	1 (10%)

4. Findings and discussion

The main findings from our evidence review are presented in this section. The section is structured according to the specific mechanisms through which technology diffusion impacts growth and productivity. The mechanisms are roughly ordered by the strength of evidence supporting their existence.

The amount of evidence available varied notably for different technologies. For example, there was relatively little on quantum computers, and the highly specialised nature of semiconductors, which are also often integrated with other technologies, makes it difficult to isolate and study their impact. Our review ended up covering broadband/internet access (4 articles), cloud computing (3 articles), information and communication technologies (ICT) (4 articles), robotics (2 articles), artificial intelligence (2 articles), green energy technologies (2 articles) and data analytics (2 articles). Most papers examined one of these technologies in isolation.

The type of evidence also varied. Some of the studies in the sample are correlational, in that they examine associations between firms’ use of technology and their productivity/growth outcomes but are unable to confirm the direction of causality (i.e. it could be that better-performing firms tend to also be more willing to adopt new technologies, rather than the technologies driving their better performance). Other studies used more robust quasi-experimental methods to help clarify the causal relationship, for example by using instrumental variable (IV) and regression-discontinuity design methods. Beyond the methodologies employed in the studies, there is also variation in the industries, countries, and time periods examined, which makes it more challenging to identify their applicability to the current UK.

4.1 Mechanism 1: Training, management expertise and upgrading organisational processes are crucial to unlock the benefits of new technologies.

The evidence for this mechanism is supported by four papers examining causal impact through IV (UK, Canada and 2 US), one survey that critically reviews more than 50 published research articles on computers and productivity (mostly US) and one large structural equation modelling study using over 50 years of US firm data.

A common view in the literature is that productivity gains from new technologies do not materialise instantaneously after initial adoption, but instead often follow a “J-curve” trajectory, where initial investments may not be immediately reflected in performance metrics. This phenomenon is largely attributed to the need for complementary intangible investments – such as developing new business processes or organisational adjustments, enhancing managerial experience, and training workers – that are critical but difficult to isolate or quantify.

For example, a major 2021 observational study examined the value of intangible investment required by general-purpose technologies like computer hardware, software and AI.^{[footnote 5]} After assessing a wide range of data (e.g. market values, book values, R&D capital, organisational capital) from all publicly held companies in the U.S. between 1961 and 2017, the authors concluded that in order to benefit from new general-purpose technologies (GPTs), firms must invest in a range of complementary intangible investments like creating new business processes, developing managerial experience, and training workers.

Similarly, a 2003 review of over 50 articles examining the impact of IT investment on productivity found that, overall, at both the firm and country level, greater investment in IT was associated with greater productivity growth.^{[footnote 6]} It also found that the wide range of performance of IT investments among different organisations could be explained by whether those firms also made complementary investments in organisational capital such as decentralised decision-making systems, job training, and business process redesign. The trainability of these kinds of skills, such as management ability, is also supported by a 2019 study showing that participation in a management training programme improved the long-run productivity of managers in Italy by better equipping them with the skills needed to make use of advanced machinery.^{[footnote 7]}

This past trend is also supported by more recent individual studies. For example, a 2022 study of 2,000 large UK firms, using an instrumental variable approach, examined the impact of digital adoption on firm productivity.^{[footnote 8]} It found that companies which made large investments in digital tools often also required a broader suite of significant organisational adjustments and investments to unlock their full benefits.

Another 2021 study, also using an instrumental variable approach, examined the impact of predictive analytics on firm performance using data from 30,000 US manufacturers.^{[footnote 9]} They found that organisations that used predictive analysis were significantly more productive. They found that firms were more likely to realise productivity gains from predictive analytics when investments in these tools were combined with at least one of IT capital investment, educated workers, or flow-efficient production processes. The authors also noted the importance of managerial capacity for responding to data-driven insights.

Moreover, a 2021 instrumental variable study using data representative of businesses across the Canadian economy investigated the relationship between robot adoption and firms’ employment and organisational practices.^{[footnote 10]} They found that robots did not affect employment within the firm uniformly – they led to net increases in the headcount of non-managerial employees but decreases in the headcount of managerial employees. The two explanations for this were that (i) robots reduced the number of human errors in the production chain, which in turn reduced the amount of human supervision required, and (ii) the adoption of robots increased the number of highly-skilled employees who in turn required less managerial supervision. They also found that robot adoption which produced changes in employment required complementary changes in organisational practices. Some of these changes could be quite significant. For example, in some cases investments in robotics reduced the cost of monitoring employee performance, such that it then became more feasible to implement performance-based pay programmes.

Lastly, a 2019 instrumental variable study examined the relationship between data analytics capabilities and firm-level innovation (measured using patent records) using detailed longitudinal data on 331 large publicly traded US firms.^{[footnote 11]} The authors found a significant complementarity between process improvement practices and data analytics capabilities; specifically, among these firms, data analytics capabilities were more valuable for firms who were more oriented around process improvement and who created new tools by combining existing technologies (vs creating entirely new technologies).

4.2 Mechanism 2: Firms with more in-house expertise and human capital reap more benefits from new technologies.

This mechanism is supported by five papers examining the impact of new technologies using causal methods (3 in the UK, 1 in Norway and 1 in Canada), and three more using descriptive analyses (in the US) and structural equation modelling (Spain, Portugal).

A 2023 regression discontinuity study examined the effect of broadband diffusion firms’ growth and productivity using data covering economic activity in all economic sectors (except agriculture and finance) from 1997 onward in the UK.^{[footnote 12]} It found that for both urban and rural firms, broadband caused an increase in firm size, but not labour productivity, and that these increases were concentrated in knowledge-intensive industries. These results suggest that firms characterised by substantial R&D expenditures and a highly skilled workforce stand to gain the most from broadband adoption. Another important firm-level difference identified in a 2018 paper by the same authors was that firms with qualified IT staff and programmers benefit more from these digital technologies.^{[footnote 13]}

A 2020 paper using data on Spanish manufacturing SMEs from 2008 and 2015 found that the use of robotic devices was associated with better performance and higher firm productivity.^{[footnote 14]} Using a structural equation modelling approach, they systematically tested 12 hypotheses concerning variation in labour productivity; one significant finding was that employee learning and training were a prerequisite for taking advantage of the efficiency brought about by industrial robotics. Similarly, the 2014 study examining cloud adoption in the manufacturing and services sectors in Portugal also found that firms possessing workforces with the requisite skills and technical competence were more likely to integrate cloud computing technologies.^{[footnote 15]} A 2021 study of Canadian business presented a more mixed picture: it found that firm robot adoption led to decreases in middle-skilled employment, but increases in low and high-skilled employment.^{[footnote 16]}

A 2015 quasi-random natural experiment investigated the effect of broadband diffusion on labour productivity and workers’ wages using data from all nonfinancial joint-stock firms over 2000–2008 in Norway.^{[footnote 17]} The study exploited a public program with limited funding that did a staggered roll-out of broadband access points, thereby providing semi-random variation in the availability and adoption of broadband internet in firms, enabling a relatively clean examination of its effect. It found that broadband internet improved the labour market outcomes and productivity of skilled workers but worsened them for unskilled workers – the key mechanism being that broadband seemed to enhance the ability of skilled workers to execute nonroutine abstract tasks, but substitute for unskilled workers in performing routine tasks. Similarly, the adoption of AI and robotics for automation may further increase demand for skilled workers (not necessarily managers) but reduce opportunities for low-skill jobs.^{[footnote 18]}

A 2022 study using an instrumental variable approach examined the role of digital adoption and its link to firm productivity.^{[footnote 19]} Using panel data for 2,000 large UK firms over 2015 to 2018, it found that larger firms were more digital-intensive and more productive. Additionally, for companies less reliant on digital technologies, transforming their operations to become intensely digital-driven and productivity-focused required significant up-front organisational adjustments and investments. This aligns with an observation in the literature that firms may experience a temporary reduction in productivity after implementing new digital systems before realising eventual longer-term performance improvements (the ‘J-curve’ effect). They also found important differences among firms – those with in-house ICT specialists tended to be more productive than firms with bought-in capabilities.

Not all studies on this topic identify a mechanism of firm-level staff expertise making a significant difference to the impact of high-speed broadband to firms’ commercial performance.^{[footnote 20]} For example, a 2015 instrumental variable study examining the impact of broadband adoption in Ireland across 2,200 manufacturing firms over 2002-09 found it had no significant impact on firms’ productivity. This was true on average, and also when examining the firms by a range of factors (e.g. size, narrow defined industry). However, that study did not examine whether this impact varied depending on the availability (or absence) of the kind of staff expertise examined in the other studies examined in this section.

4.3 Mechanism 3: Larger firms with access to more resources are more able to absorb the high costs of technological investments.

The evidence for this mechanism is supported by one paper examining causal impact through an Instrumental Variable approach (in the UK), a multi-country literature review on green technologies, one descriptive analysis (in the US) and one structural equation modelling study (in Portugal) which identified it as a key mechanism.

A finding common across several studies is that large firms are more likely to adopt advanced technologies compared to their smaller counterparts within the same industries and cohorts. This trend is attributed to the ability of larger firms to absorb the substantial fixed costs associated with these technologies.^{[footnote 21]} For example, a 2014 study investigated the determinants of cloud adoption in the manufacturing and services sectors using data from 369 firms in Portugal.^{[footnote 22]} It found that cloud computing’s ability to improve the quality of business operations, perform tasks more quickly, increase productivity, and provide new business opportunities allowed firms to reduce cost. Using a structural equation modelling approach, the authors found several important mechanisms which encouraged firms to adopt and benefit from cloud computing. One is size: larger firms tend to be more likely to adopt the technology. The authors judged this to be due to these firms having greater resources to cover the cost and investment risk of an emerging technology – smaller firms frequently lacked the requisite resources, making it challenging for them to engage with these technologies. They also found that firms with an established technology infrastructure and workforce with the necessary skills and technical competence were better suited for cloud integration (see also mechanism 2). Finally, ‘top management support’ (which we infer as seniors/executives) in the form of committing financial and organisational resources and engaging in the process of embedding it into the firm was an important driver of adoption (but this finding held only for the service sector).

Additionally, a 2021 literature review of 18 empirical studies, all using econometric techniques designed to identify causal relationships, examined whether the deployment of clean technologies led to lower production costs of various energy consuming sectors (industries, power utilities, transportation, buildings, and agriculture).^{[footnote 23]}The review included peer reviewed journal articles as well as documents produced by international organisations, research institutions, and government departments. The analysis found that most studies show a positive relationship between clean technology investments and firm productivity. This link was affected by the size of the company – larger companies, particularly those operating in energy-intensive sectors, were more likely to invest in green/clean technologies. Smaller companies, especially those with lower energy consumption, may not find these investments profitable due to their limited capacity to absorb the associated costs. In certain instances, innovations in green/clean technology were negatively associated with company productivity, primarily because of the extra direct expenses related to R&D for innovation and altering production design and processes.

An exception to this trend is when the high costs of new technologies – such as creating broadband infrastructure – are not directly borne by firms.^{[footnote 24]} For example, a 2018 instrumental variable study examining the introduction of high-speed broadband to the UK found that it helped to grow the scale (income and employment) of firms, with the effect being larger for small firms – broadband enabled them to grow by creating websites and engaging in e-commerce, without improving their productivity (i.e. output per worker remained the same even as firm size and income may have increased). This insight underscores the transformative potential of digital technologies in expanding opportunities for smaller enterprises by broadening their market access and operational scale.

4.4 Mechanism 4: Young firms that aren’t already reliant on existing technologies are more likely to capitalise on the arrival of new ones.

The evidence for this mechanism is supported by two recent high-quality studies (1 UK, 1 US), which both use an instrumental variable approach to support their causal claims – however, both pertain specifically to cloud computing. A third descriptive study (in the US) focusing on 5 novel technologies also supports this mechanism.

The first, a 2023 study, used a complete census of UK firms’ data to investigate the impact of cloud adoption on firm performance.^{[footnote 25]} The data is very rich, allowing the authors to track the births and deaths of firms and their geographic dispersion. It found that, compared to incumbent firms, younger firms benefit more from cloud technologies adoption in terms of increased employment and sales. The mechanism is that these services allow young firms to substitute away from the capital-intensive model of owning their own IT equipment, thereby reducing their fixed costs and increasing their operational flexibility. Furthermore, the propensity for younger firms to adopt advanced technologies is higher than that of older firms within the same industries. This trend can be attributed to younger firms facing fewer organisational barriers to the adoption of new technologies.^{[footnote 26]}

A 2019 study, also using an instrumental variable approach, looked at the effect of cloud computing adoption on firm survival and performance, using representative panel data of US manufacturers from 2006 to 2014.^{[footnote 27]} They found the association between IT services spending and firm survival, growth and performance was much stronger for younger companies – older companies mostly saw little or no benefit to extra IT spending, except in certain specifications and sector contexts. That said, they also found that business-specific IT investments were risky for young companies – they increased the likelihood of these firms exiting the market, but also improved the productivity of the firms that survived. A key benefit of the cloud therefore seems to have been in how it allowed firms to experiment with readily available, flexible IT solutions early in their lifecycle.

Lastly, a study using 2019 data from a national and representative survey of 300,000 US firms documented the relationship between the adoption of 5 novel technologies (AI, robotics, dedicated equipment, specialised software, and cloud), firm characteristics and workforce outcomes.^{[footnote 28]} They found that adoption of these technologies is still relatively limited – half of firms do not use any of them, and the proportion using AI (3.2%) or robotics (2%) is very low in absolute terms. They also found that younger firms are more likely to adopt advanced technologies than older firms in the same industries, possibly because they face fewer internal organisational barriers. A limitation of this study was that it was purely descriptive – it described correlations between use of these technologies and firm outcomes but was unable to make strong causal claims.

4.5 Mechanism 5: Infrastructure and regulatory environments influence the likelihood of adopting technologies.

The evidence for this mechanism is supported by one paper examining causal impact through IV (Germany) and one multi-country structural equation modelling study (Germany, Austria, and Switzerland).

A 2023 study used a dataset of firms in Germany with at least 5 employees who were representative of the manufacturing, mining, utilities, and business-oriented service sectors to examine whether their use of AI translated into higher sales.^{[footnote 29]} The authors found that although only a small proportion of firms were currently using AI, those that were using it experienced a significant increase in sales – the studies use of IV analysis means that the authors attribute this relationship as causal. Although they did not clearly identify the mechanism through which these AI tools translate into greater sales, they concluded that managers should be better informed about the potential of AI to increase productivity, and also noted other barriers to its widespread adoption such as a lack of robust IT infrastructure (e.g. secure cloud computing) and data privacy regulations.

A 2019 study used a representative sample of firms in Germany, Austria and Switzerland to investigate the relationship between firms’ investments in green energy technologies (such as wind or hydroelectric power plants or solar systems, or energy-saving technology for production, ICT, transport or building technology) and productivity.^{[footnote 30]} It found that, while the productivity effects of investment in green energy technologies were significantly positive for firms with high energy costs – these accounted for only 19% of the firms in the sample. There were no significant effects for firms with medium energy costs – and the effects were even significantly negative for firms with low energy costs (i.e. they got a negative return on their investment in this technology) for the other 81% of firms. This finding implies that, without external incentives, firms with low-medium energy costs may be reluctant to increase their investments in green energy technologies.

4.6 Mechanism 6: After adopting new technology, a period of learning-by-doing may be needed to unlock their benefits

The evidence for this finding is supported by a 2022 study which investigated the impact of the employment of ICT specialists (in-house and external) and the use of digital technologies (cloud and big data) on productivity and labour share.^{[footnote 31]}

Using an instrumental variable approach and data from 1,065 French manufacturing firms with at least 20 employees, it found the benefits of employing external ICT specialists and the use of cloud technology took around five years to trickle through to enhanced firm productivity – which the authors attribute to a ‘learning-by-doing’ effect. In contrast, the adoption of big data analytics and the employment of in-house ICT specialists produced a ‘second mover advantage’. Early adopters of these technologies bear the burden of appropriation costs, including the initial investment and the challenges of integration. Consequently, late adopters could capitalise on the lessons learned and experiences gained by early adopters, enabling a more efficient and cost-effective integration of these technologies. 

5. Conclusions

A high-level takeaway from this review is that many policies which are good for growth in general, are also good for improving tech diffusion – such as building high quality general infrastructure, investing in skills, and maintaining a regulatory environment that minimises barriers to adoption of new technology. Beyond this, we have three specific conclusions:

5.1 Conclusion 1: Financial incentives and support could help businesses adopt and commercialise new technology, especially smaller firms

One barrier to investing in new technologies is cost. Given the finding that larger firms with more resources are better positioned to absorb these kinds of costs, SMEs may require special attention to overcome initial cost barriers and facilitate their adoption of productivity-enhancing technologies (Mechanism 3).

This applies to costs associated with direct R&D spend and costs related to wider complementary investment, which may often be required to take advantage of the latest technological developments (for instance, IT and capital investment). There are already a wide suite of financial incentives and support to encourage higher investment, for example government measures to improve access to finance, and investment-related tax reliefs.

However, more can be done to optimise their design. For example, existing measures such as R&D tax relief may be limited by friction in accessing these support systems, or by behavioural challenges such as present-bias (the tendency to be overly influenced by immediate, tangible costs compared to more abstract, longer-term benefits). Simple awareness is also a challenge – small businesses don’t have time to investigate all the potential support and incentives available to them.^{[footnote 32]}

Existing incentives, like R&D tax relief, could provide more upfront support that makes the longer-term benefits more salient during decision-making. This could involve schemes like advance assurance that give smaller firms certainty around future relief claims before they invest. The use of timely prompts and targeted campaigns may also help to raise awareness and understanding of these upfront R&D incentives among under-investing sectors and businesses – for example, by prompting companies to reinvest their tax relief proceeds at the moment they receive confirmation their claim was accepted (Mechanism 5).

More broadly, small business should be encouraged to invest, as well as reviewing the effectiveness of incentives in shifting decisions to invest, and identify potential improvements in the design and gaps in support. Lastly, we note that financial incentives are likely not sufficient on their own to address the issue, given that a firm’s ability to successfully adopt new technology also partly depends on its management – this is where pairing with the insights discussed under Conclusion 2 will be important.

5.2 Conclusion 2: Testing and evaluating ways to improve take-up can help upgrade existing business management and tech support schemes

In recent years, the UK government provided two significant programmes to encourage SMEs to boost their productivity and growth: ‘Help to Grow - Digital’ and ‘Help to Grow - Management’. These have had mixed success, as evidenced by the recent closure of ‘Help to Grow - Digital’,^{[footnote 33]} following take-up being only 1% of what was initially planned. As of September 2023,^{[footnote 34]} the ‘Management’ programme has also seen very low take-up, with only 5,290 participants completing the training course thus far.

Given that this review identified management ability as a key way to unlock (Mechanism 1), testing and evaluating ways to improve and sustain take-up and engagement with management schemes could help to make the most of lessons learned for future programmes. For example, the evaluation report^{[footnote 35]} of the ‘Digital’ scheme made a number of recommendations for improving other similar programmes, including using localised channels of promotion to ensure the target audience is reached and improving connections between all business support through other avenues. Given the strong link between management capability and successful tech adoption, businesses that are currently investing in new technology, for instance, recent R&D tax credit claimants, may benefit the most.

Greater use of peer-to-peer networking can also be a particularly effective way^{[footnote 36]} to encourage productivity-enhancing investments among firms. This review identified knowledge sharing as a potential facilitator for unlocking growth and/or productivity benefits sooner (Mechanism 6). Incorporating peer components into commissioned businesses support services could facilitate this. Testing and evaluation are not just limited to implementation and take-up – it has an essential role in the design of support schemes too. A key determinant to a scheme’s success is designing a product that works for businesses, and is something they see a need for. A number of this review’s findings indicate that this product-market-fit will depend on a firm’s size (Mechanism 3), sector (Mechanism 2), and age (Mechanism 4). Schemes that are co-created with the relevant subgroups may therefore be particularly effective. For example, this review identified that growth and productivity gains from new technologies will be dependent on alignment with existing processes, which is a greater barrier for older firms (Mechanism 4).

Rigorous testing and evaluation practices can help finite business support resources achieve maximum impact in raising UK productivity levels. Successful ways to improve existing schemes can then be codified and disseminated as best practices.

5.3 Conclusion 3: De-shrouding business-to-business (B2B) markets can improve the quality of complementary investments

Complementary investments, such as firms receiving management training (Mechanism 1), or resourcing inhouse IT specialists and upgrading their IT infrastructure (Mechanism 2), are key to unlocking the benefits of technology. Being able to access these high-quality services is therefore crucial. This means it takes longer for new innovations to diffuse through the market, and^{[footnote 37]} There are, therefore, benefits to the creation and dissemination of transparent reputation systems (akin to Tripadvisor) that enable quick, reliable assessment of B2B firms providing IT and management training services, as well as technology suppliers and firms who assist with the redesign of organisational processes. These kinds of reputation systems can change the dynamics of markets by aligning the incentives of suppliers with those of customers. This in turn should save money, drive up the quality of products and services procured, and help to unlock the potential of new technologies.

1. HMG. “The UK’s International Technology Strategy.” (2023). ↩

2. Data, Interpreting Innovation. “Oslo manual.” Paris and Luxembourg: OECD/Euro-stat (2005). ↩

3. Ogink, Ruben HAJ, Martin C. Goossen, A. Georges L. Romme, and Henk Akkermans. “Mechanisms in open innovation: A review and synthesis of the literature.” Technovation 119 (2023): 102621. ↩

4. This includes studies with broader data sets, where the UK is one country in a larger sample. ↩

5. Brynjolfsson, Erik, Daniel Rock, and Chad Syverson. “The productivity J-curve: How intangibles complement general purpose technologies.” American Economic Journal: Macroeconomics 13, no. 1 (2021): 333-372. ↩

6. Dedrick, Jason, Vijay Gurbaxani, and Kenneth L. Kraemer. “Information technology and economic performance: A critical review of the empirical evidence.” ACM Computing Surveys (CSUR) 35, no. 1 (2003): 1-28. ↩

7. Giorcelli, M (2019), “The Long-Term Effects of Management and Technology Transfers”, American Economic Review 109(1): 1–33. ↩

8. Coyle, Diane, Kieran Lind, David Nguyen, and Manuel Tong. “Are digital-using UK firms more productive.” Economic Statistics Centre of Excellence Discussion Paper 6 (2022). ↩

9. Brynjolfsson, Erik, Wang Jin, and Kristina McElheran. “The power of prediction: predictive analytics, workplace complements, and business performance.” Business Economics 56 (2021): 217-239 ↩

10. Dixon, Jay, Bryan Hong, and Lynn Wu. “The robot revolution: Managerial and employment consequences for firms.” Management Science 67, no. 9 (2021): 5586-5605. ↩

11. Wu, Lynn, Lorin Hitt, and Bowen Lou. “Data analytics, innovation, and firm productivity.” Management Science 66, no. 5 (2020): 2017-2039. ↩

12. DeStefano, Timothy, Richard Kneller, and Jonathan Timmis. “The (fuzzy) digital divide: the effect of universal broadband on firm performance.” Journal of Economic Geography 23, no. 1 (2023): 139-177. ↩

13. DeStefano, Timothy, Richard Kneller, and Jonathan Timmis. “Broadband infrastructure, ICT use and firm performance: Evidence for UK firms.” Journal of Economic Behavior & Organization 155 (2018): 110-139. ↩

14. Ballestar, María Teresa, Ángel Díaz-Chao, Jorge Sainz, and Joan Torrent-Sellens. “Knowledge, robots and productivity in SMEs: Explaining the second digital wave.” Journal of Business Research 108 (2020): 119-131. ↩

15. Oliveira, Tiago, Manoj Thomas, and Mariana Espadanal. “Assessing the determinants of cloud computing adoption: An analysis of the manufacturing and services sectors.” Information & management 51, no. 5 (2014): 497-510. ↩

16. Dixon, Jay, Bryan Hong, and Lynn Wu. “The robot revolution: Managerial and employment consequences for firms.” Management Science 67, no. 9 (2021): 5586-5605. ↩

17. Akerman, Anders, Ingvil Gaarder, and Magne Mogstad. “The skill complementarity of broadband internet.” The Quarterly Journal of Economics 130, no. 4 (2015): 1781-1824. ↩

18. Acemoglu, Daron, Gary W. Anderson, David N. Beede, Cathy Buffington, Eric E. Childress, Emin Dinlersoz, Lucia S. Foster et al. Automation and the workforce: A firm-level view from the 2019 Annual Business Survey. No. w30659. National Bureau of Economic Research, 2022. ↩

19. Coyle, Diane, Kieran Lind, David Nguyen, and Manuel Tong. “Are digital-using UK firms more productive.” Economic Statistics Centre of Excellence Discussion Paper 6 (2022). ↩

20. Haller, Stefanie A., and Sean Lyons. “Broadband adoption and firm productivity: Evidence from Irish manufacturing firms.” Telecommunications Policy 39, no. 1 (2015): 1-13. ↩

21. Acemoglu, Daron, Gary W. Anderson, David N. Beede, Cathy Buffington, Eric E. Childress, Emin Dinlersoz, Lucia S. Foster et al. Automation and the workforce: A firm-level view from the 2019 Annual Business Survey. No. w30659. National Bureau of Economic Research, 2022. ↩

22. Oliveira, Tiago, Manoj Thomas, and Mariana Espadanal. “Assessing the determinants of cloud computing adoption: An analysis of the manufacturing and services sectors.” Information & management 51, no. 5 (2014): 497-510. ↩

23. Timilsina, Govinda, and Sunil Malla. “Do Investments in Clean Technologies Reduce Production Costs? Insights from the Literature.” (2021) ↩

24. DeStefano, Timothy, Richard Kneller, and Jonathan Timmis. “Broadband infrastructure, ICT use and firm performance: Evidence for UK firms.” Journal of Economic Behavior & Organization 155 (2018): 110-139. ↩

25. DeStefano, Timothy, Richard Kneller, and Jonathan Timmis. “Cloud computing and firm growth.” Review of Economics and Statistics (2023): 1-47. ↩

26. Acemoglu, Daron, Gary W. Anderson, David N. Beede, Cathy Buffington, Eric E. Childress, Emin Dinlersoz, Lucia S. Foster et al. Automation and the workforce: A firm-level view from the 2019 Annual Business Survey. No. w30659. National Bureau of Economic Research, 2022. ↩

27. Jin, Wang, and Kristina McElheran. Economies before scale: learning, survival, and performance of young plants in the age of cloud computing. Working paper, 3112901. Rotman School of Management., 2019. ↩

28. Acemoglu, Daron, Gary W. Anderson, David N. Beede, Cathy Buffington, Eric E. Childress, Emin Dinlersoz, Lucia S. Foster et al. Automation and the workforce: A firm-level view from the 2019 Annual Business Survey. No. w30659. National Bureau of Economic Research, 2022. ↩

29. Czarnitzki, Dirk, Gastón P. Fernández, and Christian Rammer. “Artificial intelligence and firm-level productivity.” Journal of Economic Behavior & Organization 211 (2023): 188-205. ↩

30. Stucki, Tobias. “Which firms benefit from investments in green energy technologies?–The effect of energy costs.” Research Policy 48, no. 3 (2019): 546-555. ↩

31. Cette, Gilbert, Sandra Nevoux, and Loriane Py. “The impact of ICTs and digitalization on productivity and labor share: evidence from French firms.” Economics of innovation and new technology 31, no. 8 (2022): 669-692. ↩

32. BEIS. “Business Basics: Nudging Firms to Improve Productivity. A rapid literature review of behavioural factors and best-practice business prompts.” BEIS Research Paper Number 2019/17 ↩

33. BEIS. News story: Final opportunity for businesses to access Help to Grow: Digital scheme. 2022. ↩

34. DBT. Transparency data Help to Grow: Management course enrolments and participant completions. 2023. ↩

35. BEIS. Evaluation of Help to Grow: Digital Final Evaluation Report. 2024. ↩

36. BEIS. Unpicking the Productivity Puzzle Business Basics Programme – final report. 2024. ↩

37. The Behavioural Insights Team. The Shrouded Economy. 2024. ↩