---
title: "Benchmarks: AI Strategy"
source: "https://benchmarks.georgian.io/applied-ai/wave-3/benchmarks/ai-strategy"
intro: "Maturity, priorities, models, infrastructure, and ROI — the strategic baseline benchmarks for the full sample."
citation: "Georgian + NewtonX, Applied AI Benchmarks, Wave 3 (June 2026)"
---

# Benchmarks: AI Strategy

> Maturity, priorities, models, infrastructure, and ROI — the strategic baseline benchmarks for the full sample.

## AI maturity distribution

**Survey question:** How would you describe your organization’s current AI adoption maturity?

**Audience:** All respondents · n=501

**Wave 3**

**63%** — of B2B software executives surveyed are past the experimentation threshold — Jogger or Runner tier

**Data for AI maturity distribution, Wave 3**

| Category | % of organizations |
| --- | --- |
| Runner | 22% |
| Jogger | 41% |
| Walker | 30% |
| Crawler | 7% |

*All respondents · n=501 · Wave 3 · Runner n=110 / Jogger n=204 / Walker n=150 / Crawler n=37*

**Wave over Wave**

**+29 pts** — combined Runner + Jogger share across 18 months — from 34% in Wave 1 to 63% in Wave 3

**Data for Maturity tiers across waves**

| Category | Wave 1 | Wave 2 | Wave 3 |
| --- | --- | --- | --- |
| Runner | 12% | 14% | 22% |
| Jogger | 22% | 29% | 41% |
| Walker | 49% | 41% | 30% |
| Crawler | 16% | 15% | 7% |

*Direct W1 (n=601) → W2 (n=634) → W3 (n=501) · concept-based CWR matcher applied · all three waves comparable*

**By maturity**

*Not applicable — this benchmark is itself the segmentation variable. The Wave 3 distribution above is the reference base for every maturity cross-tab in the report.*

## AI strategic priorities

**Survey question:** Which of the following are among your organization’s top 5 strategic priorities?

**Audience:** All respondents · n=501

**Wave 3**

**83%** — selected AI for internal efficiency as a top-5 organizational priority — the most-selected priority in the survey

**Data for Top-5 strategic priorities, Wave 3**

| Category | % naming top-5 |
| --- | --- |
| AI for efficiency | 83% |
| Build AI internally | 65% |
| Market expansion | 47% |
| Deploy external AI | 43% |
| Brand development | 37% |
| Customer experience | 36% |
| Tech stack upgrade | 34% |
| Revenue diversification | 31% |
| M&A activity | 20% |
| Cybersecurity | 19% |
| Restructuring | 14% |
| Agile transformation | 12% |
| Data privacy | 12% |
| Employee engagement | 10% |
| Supply chain | 5% |
| Sustainability | 3% |

*All respondents · n=501 · Wave 3 · sorted by share naming each priority*

**Wave over Wave**

**−11 pts** — customer experience — the sharpest movement of any priority, as AI absorbs the CX agenda

**Data for Priority movement, Wave 2 vs Wave 3**

| Category | Wave 2 | Wave 3 |
| --- | --- | --- |
| AI for efficiency | 80% | 83% |
| Build AI internally | 63% | 65% |
| Deploy external AI | 42% | 43% |
| Market expansion | 41% | 47% |
| Customer experience | 47% | 36% |
| Revenue diversification | 39% | 31% |
| Brand development | 40% | 37% |
| Tech stack upgrade | 35% | 34% |
| Cybersecurity | 22% | 19% |

*Direct W2 (n=634) → W3 (n=501) · Wave 1 used a different encoding and is not directly comparable for this question*

**By maturity**

*Not meaningful at this level — this is a screened cohort, so 100% of respondents placed at least one AI priority in their top 5. The maturity story on priorities is covered in Benchmark 1.7 (AI product strategy) and Benchmark 2.14 (agentic investment reasons).*

## Most advanced AI model in production

**Survey question:** What is the most advanced AI model approach your organization uses in production?

**Audience:** All respondents · n=501

**Wave 3**

**42%** — use multi-step or multi-hop agentic API calls — the most common most-advanced approach in production

**Data for Most advanced AI model approach in production, Wave 3**

| Category | % of organizations |
| --- | --- |
| Multi-step / multi-hop calls (agentic) | 42% |
| Custom-trained small models (<3B) | 20% |
| Fine-tuned large models (>3B) | 16% |
| Single API calls | 11% |
| Custom-trained large models (>3B) | 7% |
| POCs only — none in production | 3% |
| No AI models | 1% |

*All respondents · n=501 · Wave 3*

**Wave over Wave**

*Not applicable — Wave 3 debut. No prior-wave equivalent for this question.*

**By maturity**

**7×** — Runners are roughly seven times more likely than Walkers to name a custom-trained or fine-tuned large model as their most advanced approach — 48% vs 7% combined

**Data for Most advanced approach by maturity tier**

| Category | Runner | Jogger | Walker | Crawler |
| --- | --- | --- | --- | --- |
| Multi-step / multi-hop | 29% | 46% | 52% | 24% |
| Fine-tuned large models | 30% | 18% | 4% | 6% |
| Custom-trained large models | 18% | 4% | 3% | 0% |
| Custom-trained small models | 17% | 26% | 19% | 12% |
| Single API calls | 5% | 6% | 17% | 29% |

*All respondents · n=501 · Runner n=110 / Jogger n=204 / Walker n=150 / Crawler n=37 · Wave 3*

## AI infrastructure stack

**Survey question:** Which AI infrastructure components does your organization currently use?

**Audience:** Technical Decision Makers · n=252

**Wave 3**

**48–53%** — adoption across queues, data orchestration, vector databases and LLM observability — the four components that define a minimum viable production AI stack

**Data for AI infrastructure components in use, Wave 3**

| Category | % adopted |
| --- | --- |
| Queues / background jobs | 53% |
| Data orchestration pipelines | 52% |
| Vector databases | 50% |
| LLM observability & evaluation | 48% |
| Serverless inference infra | 31% |
| Inference services | 31% |
| Durable workflow engines | 22% |
| Headless browser automation | 20% |
| Agentic auth | 19% |

*Technical Leadership · n=252 · Wave 3*

**Wave over Wave**

*Not applicable — Wave 3 debut. No prior-wave equivalent for this question.*

**By maturity**

**+34 pts** — Runner vs Walker gap on LLM observability and evaluation — 67% of Runners use it, 33% of Walkers do

**Data for Infrastructure adoption by maturity tier**

| Category | Runner | Jogger | Walker | Crawler |
| --- | --- | --- | --- | --- |
| LLM observability & evaluation | 67% | 53% | 33% | 12% |
| Queues / background jobs | 64% | 51% | 45% | 53% |
| Vector databases | 61% | 53% | 45% | 12% |
| Data orchestration pipelines | 55% | 57% | 48% | 35% |
| Inference services | 38% | 39% | 19% | 12% |
| Serverless inference infra | 38% | 34% | 27% | 12% |
| Headless browser automation | 32% | 19% | 12% | 18% |
| Agentic auth | 30% | 19% | 13% | 6% |
| Durable workflow engines | 26% | 20% | 25% | 6% |

*Technical Leadership · n=252 · Runner n=66 / Jogger n=94 / Walker n=75 / Crawler n=17 · Wave 3*

## AI model types in use

**Survey question:** Which of the following AI model types does your organization currently use?

**Audience:** Technical Decision Makers · n=252

**Wave 3**

**84%** — use large language models — now the baseline floor; where an organization sits above it defines the capability frontier

**Data for AI model types in use, Wave 3**

| Category | % in use |
| --- | --- |
| Large language models | 84% |
| Multimodal models | 51% |
| Large reasoning models | 44% |
| Open-source / open-weight | 41% |
| Voice models | 25% |
| Model routers | 17% |
| Diffusion LLMs | 16% |
| Vision foundation models | 14% |
| Vision-language-action models | 14% |
| Video foundation models | 9% |
| World models | 9% |

*Technical Leadership · n=252 · Wave 3*

**Wave over Wave**

*Not applicable — Wave 3 debut. No prior-wave equivalent for this question.*

**By maturity**

**+34 pts** — Runner vs Walker gap on large reasoning models — 65% vs 31% — the largest maturity gap in the full report

**Data for Model-type adoption by maturity tier**

| Category | Runner | Jogger | Walker | Crawler |
| --- | --- | --- | --- | --- |
| Large reasoning models | 65% | 47% | 31% | 18% |
| Multimodal models | 59% | 53% | 45% | 35% |
| Open-source / open-weight | 52% | 43% | 33% | 24% |
| Voice models | 38% | 27% | 15% | 6% |

*Technical Leadership · n=252 · Runner n=66 / Jogger n=94 / Walker n=75 / Crawler n=17 · Wave 3*

## AI UX strategy

**Survey question:** How does your organization primarily incorporate AI into its user experience?

**Audience:** Technical Decision Makers · n=252

**Wave 3**

**55%** — use contextual assistance — AI embedded behind fixed UI elements — the most common UX pattern in production

**Data for How AI is incorporated into the user experience, Wave 3**

| Category | % using |
| --- | --- |
| Contextual assistance | 55% |
| Background automation | 47% |
| Primary conversational interface | 46% |
| Adaptive / generative UI | 40% |
| Secondary conversational overlay | 30% |
| Voice | 22% |
| No formal UX strategy | 6% |

*Technical Leadership · n=252 · Wave 3*

**Wave over Wave**

*Not applicable — Wave 3 debut. No prior-wave equivalent for this question.*

**By maturity**

**+31 pts** — Runner vs Walker gap on adaptive / generative UI — 55% of Runners build AI that changes the interface dynamically, against 24% of Walkers

**Data for UX patterns with the largest maturity gaps**

| Category | Runner | Jogger | Walker | Crawler |
| --- | --- | --- | --- | --- |
| Adaptive / generative UI | 55% | 43% | 24% | 18% |
| Voice | 35% | 19% | 16% | 12% |
| Contextual assistance | 68% | 49% | 52% | 47% |

*Technical Leadership · n=252 · Runner n=66 / Jogger n=94 / Walker n=75 / Crawler n=17 · Wave 3*

## AI product strategy

**Survey question:** Which best describes your organization’s current AI product development strategy?

**Audience:** Technical Decision Makers · n=252

**Wave 3**

**27%** — are predominantly building net-new AI products — for the first time this leads enhancement-only, at 16%

**Data for AI product development strategy, Wave 3**

| — | Net-new AI products | Equal emphasis | Enhancing existing |
| --- | --- | --- | --- |
| Share of technical leaders | 27% | 57% | 16% |

*Technical Leadership · n=252 · Wave 3*

**Wave over Wave**

**+10 pts** — net-new AI product strategy from Wave 2 to Wave 3, while enhancement-only fell 12 points over the same period

**Data for Product strategy across waves**

| Category | Wave 1 | Wave 2 | Wave 3 |
| --- | --- | --- | --- |
| Predominantly net-new | 20% | 17% | 27% |
| Equal emphasis | 58% | 55% | 57% |
| Predominantly enhancing | 22% | 28% | 16% |

*Direct W1 (n=300) → W2 (n=326) → W3 (n=252) · identical question wording across all three waves*

**By maturity**

**2×** — Runners and Joggers are twice as likely as Walkers to be predominantly building net-new AI products — 33% vs 17%

**Data for Net-new vs enhancement focus by maturity tier**

| Category | Runner | Jogger | Walker | Crawler |
| --- | --- | --- | --- | --- |
| Predominantly net-new | 33% | 33% | 17% | 12% |
| Predominantly enhancing | 11% | 13% | 17% | 47% |

*Technical Leadership · n=252 · Runner n=66 / Jogger n=94 / Walker n=75 / Crawler n=17 · Wave 3*

## AI ROI measurement

**Survey question:** Which best describes how your organization measures the ROI of its AI investment?

**Audience:** Technical Decision Makers · n=252

**Wave 3**

**55%** — are unable to link AI investment ROI to hard revenue or cost savings

**Data for How AI ROI is measured, Wave 3**

| Category | % of respondents |
| --- | --- |
| Directly tied to new revenue | 23% |
| Directly tied to cost savings | 22% |
| Believes it yields benefits but not quantified | 24% |
| Tied to revenue or cost but not tracked | 15% |
| Unsure about ROI | 14% |

*Technical Leadership · n=252 · Wave 3*

**Wave over Wave**

**Flat** — actively tracked ROI held at 45–46% across 18 months — unchanged while AI maturity (Benchmark 1.1) nearly doubled over the same period

**Data for Actively tracked AI ROI across waves**

| Period | Actively tracked (revenue + cost combined) |
| --- | --- |
| Wave 1 (Nov 2024) | 45% |
| Wave 2 (Jun 2025) | 46% |
| Wave 3 (May 2026) | 45% |

*Direct W1 (n=300) → W2 (n=326) → W3 (n=252) · identical question wording across all three waves*

**By maturity**

**6 pts** — separate Runners from Crawlers on tying AI ROI to hard revenue or cost savings — maturity barely moves measurement discipline

**Data for ROI approach by maturity tier**

| Category | Runner | Jogger | Walker | Crawler |
| --- | --- | --- | --- | --- |
| Directly tied to new revenue | 27% | 26% | 20% | 12% |
| Directly tied to cost savings | 20% | 22% | 19% | 41% |
| Believes benefits, not quantified | 23% | 26% | 25% | 18% |
| Tied to revenue/cost, not tracked | 17% | 15% | 15% | 18% |
| Unsure about ROI | 11% | 12% | 21% | 12% |

*Technical Leadership · n=252 · Runner n=66 / Jogger n=94 / Walker n=75 / Crawler n=17 · Wave 3*

## AI sentiment (Tech)

**Survey question:** To what extent do you agree or disagree: “I would rather buy an AI agent that does a job than a SaaS tool that helps me do it.”

**Audience:** Technical Decision Makers · n=252

**Wave 3**

**81%** — of tech leaders feel energized by AI’s rapid advancement — but only 57% would replace a SaaS tool with an agent that does the job

**Data for AI perception statements — technical leaders, Wave 3**

| Category | % agree |
| --- | --- |
| I feel energized by the rapid advancement of AI | 81% |
| AI will revolutionize my function | 77% |
| AI tools have increased my team's expected work volume | 72% |
| There is too much hype around AI in my industry | 60% |
| I would rather buy an AI agent than a SaaS tool | 57% |
| Reliance on AI will stifle long-term innovation | 45% |
| AI will create more jobs than it displaces | 32% |
| AI will erode customer trust in our brand | 21% |

*Technical Leadership · n=252 · Wave 3*

**Wave over Wave**

*Not applicable — Wave 3 debut. No prior-wave equivalent for this statement.*

**By segment**

**−12 pts** — Runner × Enterprise lags Jogger × Enterprise on agent-over-SaaS preference — 59% vs 71%

**Data for Agent-over-SaaS preference by segment**

| Category | % agree |
| --- | --- |
| Jogger × Enterprise (n=51) | 71% |
| Runner × Enterprise (n=42) | 59% |
| Tech Enterprise (n=134) | 59% |
| Runner overall (n=66) | 58% |
| Tech overall (n=252) | 57% |
| Tech Growth-stage (n=118) | 54% |
| Jogger × Growth-stage (n=43) | 53% |
| Walker (n=75) | 48% |

*Technical Leadership · n=252 · Wave 3 · small cells directional*

*Source: Full agree/disagree distribution for the agent-vs-SaaS statement: strongly agree 18% · somewhat agree 39% · neutral 29% · somewhat disagree 10% · strongly disagree 4%.*

## AI perceptions (full sample)

**Survey question:** To what extent do you agree or disagree with each of the following statements about AI in your organization?

**Audience:** All respondents · n=501 (Tech n=252 + GTM n=249)

**Wave 3**

**80% & 78%** — agree AI will revolutionize their function (80%) and feel energized by its advancement (78%) — both statistically higher than any other sentiment measured

**Data for AI perception statements — full sample vs function, Wave 3**

| Category | Full sample | Tech | GTM |
| --- | --- | --- | --- |
| AI will revolutionize my function | 80% | 77% | 84% |
| Energized by rapid AI advancement | 78% | 81% | 76% |
| AI increased expected work volume | 66% | 72% | 60% |
| Too much hype around AI | 57% | 60% | 53% |
| Would rather buy an agent than SaaS | 53% | 57% | 49% |
| Reliance on AI will stifle innovation | 44% | 45% | 43% |
| AI will create more jobs than it displaces | 27% | 32% | 21% |
| AI will erode customer trust | 18% | 21% | 15% |

*All respondents · n=501 (Tech n=252 + GTM n=249) · Wave 3*

**Wave over Wave**

**+10 pts** — agreement that AI will revolutionize their function over 18 months — the strongest and most consistent sentiment trend in the benchmark

**Data for The two perception statements tracked across all three waves**

| Period | Will revolutionize my function | Too much hype |
| --- | --- | --- |
| Wave 1 (Nov 2024) | 70% | 56% |
| Wave 2 (Jun 2025) | 73% | 47% |
| Wave 3 (May 2026) | 80% | 57% |

*Direct W1 (n=601) → W2 (n=634) → W3 (n=501) · the remaining six statements are Wave 3 debuts and are not included here*

**By maturity**

**24 pts** — separate Runners from Crawlers on belief that AI will revolutionize their function — 89% vs 65%. The deeper the investment, the stronger the conviction

**Data for Perception statements by maturity tier**

| Category | Runner | Jogger | Walker | Crawler |
| --- | --- | --- | --- | --- |
| AI will revolutionize my function | 89% | 80% | 77% | 65% |
| Energized by rapid AI advancement | 84% | 81% | 74% | 68% |
| AI increased total work volume | 67% | 73% | 59% | 54% |
| Too much hype around AI | 54% | 57% | 57% | 62% |
| Would rather buy an agent than SaaS | 53% | 55% | 50% | 54% |
| Reliance on AI will stifle innovation | 37% | 49% | 43% | 46% |
| AI will create more jobs than it displaces | 34% | 25% | 27% | 11% |
| AI will erode customer trust | 15% | 19% | 19% | 19% |

*All respondents · n=501 · Runner n=110 / Jogger n=204 / Walker n=150 / Crawler n=37 · Wave 3*

## AI for internal efficiency vs customer-facing product

**Survey question:** Please indicate to what extent your organization uses AI for internal operational efficiency, and embeds AI within the product or service you sell to your customers.

**Audience:** All respondents · n=501

**96% and 88%** — use AI for internal efficiency (96%) and embed it in a customer-facing product (88%) — B2B AI integration is nearing full penetration on both fronts

**Data for Where AI is in use — internal vs customer-facing, by maturity tier**

| Category | Total | Runner | Jogger | Walker | Crawler |
| --- | --- | --- | --- | --- | --- |
| Internal operational efficiency | 96% | 99% | 97% | 95% | 81% |
| Embedded in customer-facing product | 88% | 97% | 95% | 81% | 48% |

*All respondents · n=501 · Wave 3*

## Use of AI by function

**Survey question:** In which of the following areas are you currently using AI? Select all that apply.

**Audience:** All respondents · n=501

**Wave 3**

**80%** — use AI for data analysis, collection and insight extraction — the most frequent area of AI integration

**Data for Functions currently using AI, Wave 3**

| Category | % using AI |
| --- | --- |
| Data — analysis, collection, insights | 80% |
| Sales and marketing | 79% |
| Engineering, product, R&D | 78% |
| Market research and insights | 68% |
| Customer service / success | 67% |
| IT and security | 58% |
| Operations and field operations | 49% |
| HR & recruitment | 47% |
| Legal and compliance | 32% |
| Finance | 31% |
| Supply chain and logistics | 16% |

*All respondents · n=501 · Wave 3 · multiple selections permitted*

**By maturity**

**+19 pts** — Runner vs Walker gap on data analysis, collection and insight extraction — 90% vs 71% — in our view the most important AI use case in B2B

**Data for AI use by function and maturity tier**

| Category | Runner | Jogger | Walker | Crawler |
| --- | --- | --- | --- | --- |
| Data — analysis, collection, insights | 90% | 85% | 71% | 65% |
| Engineering, product, R&D | 92% | 83% | 67% | 58% |
| Sales and marketing | 83% | 79% | 79% | 58% |
| Market research and insights | 72% | 72% | 59% | 61% |
| Customer service / success | 75% | 70% | 63% | 47% |
| IT and security | 65% | 72% | 44% | 25% |
| Operations and field operations | 55% | 51% | 43% | 30% |
| HR & recruitment | 54% | 49% | 42% | 30% |
| Legal and compliance | 39% | 38% | 21% | 16% |
| Finance | 37% | 34% | 27% | 12% |
| Supply chain and logistics | 23% | 16% | 13% | 10% |

*All respondents · n=501 · Runner n=110 / Jogger n=204 / Walker n=150 / Crawler n=37 · Wave 3*

## Motivation to apply AI, by function

**Survey question:** What’s your main motivation for applying AI to this area?

**Audience:** All respondents · n=501

**31%** — name increasing team productivity as their main motivation, averaged across 11 functions — a 13-point gap over the next most-cited reason, cost savings

**Data for Main motivation for applying AI — average across 11 functions**

| Metric | Value | Detail |
| --- | --- | --- |
| Increasing productivity of my team | 31% | High — Legal (42%), R&D (40%) · Low — Supply chain (15%) |
| Cost savings | 18% | High — Supply chain (40%), Finance (37%) · Low — Market research (4%) |
| Revenue growth / generation | 14% | High — Sales & marketing (49%) · Low — IT security & HR (1–2%) |
| Creating a competitive advantage | 13% | High — Market research (29%) · Low — Finance (2%) |
| Improving scalability | 8% | High — IT security (17%) · All remaining 4–11% |
| Improving customer experience / retention | 8% | High — Customer service (49%) · Low — IT security & HR (1–2%) |
| Executive mandate to explore or implement AI | 4% | Range 2–9% across functions |
| Upskilling and training my people | 3% | High — HR (14%) · All remaining 1–3% |
| Competitive pressure | 2% | Range 1–5% across functions |

*All respondents · n=501 · Wave 3*

## Top reason for agentic AI

**Survey question:** What are the main reasons your organization is investing in Agentic AI? % who selected each reason.

**Audience:** Technical Decision Makers · n=252

**53%** — of Runners name creating new revenue streams or business models as a reason for investing in agentic AI — 10 points above Walkers, and the clearest signal that Runners treat AI as a revenue lever rather than a cost lever

**Data for Reasons for investing in agentic AI, by maturity tier**

| Category | Total | Runner | Jogger | Walker | Crawler |
| --- | --- | --- | --- | --- | --- |
| Increase efficiency, automate workflows | 46% | 40% | 49% | 48% | 43% |
| Create new revenue streams / business models | 43% | 53% | 35% | 43% | 20% |
| Improve customer experience and engagement | 32% | 28% | 31% | 31% | 57% |
| Reduce costs, improve resource allocation | 30% | 28% | 27% | 33% | 42% |
| Enhance decision-making and strategic insights | 21% | 14% | 26% | 24% | 0% |

*Technical Leadership · n=252 · Wave 3 · % selecting each reason*
