Challenges of Converting AI Spend to Value

OpenAI’s move into enterprise consulting reinforces a truth we see daily in healthcare AI delivery: models alone do not deliver value. The actual challenge lies in the final stretch: embedding AI into organization-specific workflows, managing exceptions, aligning with governance, and navigating change is where value is either realized or lost. A well-trained model might contain knowledge. But it cannot create impact without clear business logic, operational constraints, workflow orchestration, regulatory compliance, and stakeholder buy-in. This is exactly what we’re doing at KeyReply, we focus on bridging the gap between technical capability and real-world outcomes, especially in complex environments like hospitals, insurers, and public health systems. We integrate AI directly into operations in a way that works within existing systems, departments, and teams. The tangible value of AI comes from: (1) Aligning AI with revenue-driving and cost-saving priorities (2) Handling inconsistent or messy data (3) Addressing edge cases safely (4) Ensuring coordinated function across systems, departments, and user groups (5) Assigning clear accountability AI pilots typically fail not because the model lacks sophistication, but because the link between output and decision-making breaks down. Common issues include misaligned workflows, unclear ownership of processes and data, and fragmented infrastructure. Additionally, some systems are effectively non-interoperable, not because the technology can’t support it, but because the pricing makes it prohibitive. Larger vendors often charge exorbitant fees for API access or external integrations, making their platforms functionally standalone. Effective AI delivery depends on operational clarity and that is what drives measurable, sustainable results.

80% of enterprise AI projects are draining your budget with zero ROI. And it's not the technology that's failing: It's the hidden costs no one talks about. McKinsey's 2025 State of AI report reveals a startling truth: 80% of organizations see no tangible ROI impact from their AI investments. While your competitors focus on software licenses and computing costs, five hidden expenses are sabotaging your ROI: 1/ The talent gap: ↳ AI specialists command $175K-$350K annually. ↳ 67% of companies report severe AI talent shortages. ↳ 13% are now hiring AI compliance specialists. ↳ Only 6% have created AI ethics specialists. When your expensive new hire discovers you lack the infrastructure they need to succeed, they will leave within 9 months. 2/ The infrastructure trap: ↳ AI workloads require 5-8x more computing power than projected. ↳ Storage needs can increase 40-60% within 12 months. ↳ Network bandwidth demands can surge unexpectedly. What's budgeted as a $100K project suddenly demands $500K in infrastructure. 3/ The data preparation nightmare: ↳ Organizations underestimate data prep costs by 30-40%. ↳ 45-70% of AI project time is spent on data cleansing (trust me, I know). ↳ Poor data quality causes 30% of AI project failures (according to Gartner). Your AI model is only as good as your data. And most enterprise data isn't ready for AI consumption. 4/ The integration problem: ↳ Legacy system integration adds 25-40% to implementation costs. ↳ API development expenses are routinely overlooked. ↳ 64% of companies report significant workflow disruptions. No AI solution can exist in isolation. You have to integrate it with your existing tech stack, or it will create expensive silos. 5/ The governance burden: ↳ Risk management frameworks cost $50K-$150K to implement. ↳ New AI regulations emerge monthly across global markets. Without proper governance, your AI can become a liability, not an asset. The solution isn't abandoning AI. It's implementing it strategically with eyes wide open. Here's the 3-step framework we use at Avenir Technology to deliver measurable ROI: Step 1: Define real success metrics: ↳ Link AI initiatives directly to business KPIs. ↳ Build comprehensive cost models including hidden expenses. ↳ Establish clear go/no-go decision points. Step 2: Build the foundation first: ↳ Assess and upgrade infrastructure before deployment. ↳ Create data readiness scorecards for each AI use case. ↳ Invest in governance frameworks from day one. Step 3: Scale intelligently: ↳ Start with high-ROI, low-complexity use cases. ↳ Implement in phases with reassessment at each stage. Organizations following this framework see 3.2x higher ROI. Ready to implement AI that produces real ROI? Let's talk about how Avenir Technology can help. What AI implementation challenge are you facing? Share below. ♻️ Share this with someone who needs help implementing. ➕ Follow me, Ashley Nicholson, for more tech insights.

Most enterprise AI roadmaps have a use case for every function. A pilot in every department. It feels like coverage. It behaves like dilution. The organizations converting AI investment into enterprise value are doing the opposite: fewer strategic bets, deeper investment, clearer ownership, and explicit tradeoffs about where not to invest. 1/ Broad diffusion produces activity, not accountability Dozens of pilots across dozens of functions make it difficult to hold any single leader accountable to a business outcome because no individual initiative was designed to move one. → Ask which three business priorities your AI investment is funded to support → If the answer spans every department, the priorities aren’t focused enough 2/ Concentrated investment creates more enterprise leverage Concentrating capital, talent, data, and leadership attention in a smaller number of high-value domains creates greater potential for enterprise impact. → Fewer initiatives get deeper resources and stronger executive attention → Greater focus makes outcomes easier to measure and accountability clearer 3/ The ROI conversation has moved beyond productivity Futurum Group's 2026 survey found revenue growth and profitability combined rose to 21.7% as the primary AI ROI metric, while productivity declined as the leading success measure. → "We saved 4 hours a week" is an incomplete answer → Executives want to know which revenue, margin, or growth metric changed 4/ Most companies are still struggling to turn AI into financial impact Deloitte found that 66% of organizations report productivity and efficiency gains, but only 20% currently report AI-driven revenue growth. The gap between using AI and materially impacting the business remains wide and adding more use cases doesn’t necessarily close it. 5/ Economic leverage matters more than functional coverage The highest-value AI investments are the ones connected to the biggest economic levers: pricing, churn, conversion, revenue-cycle speed, margin improvement, or new products and services. → Rank initiatives by P&L impact, not the number of users they reach. → One focused AI capability can create more value than dozens of broadly adopted tools. 6/ Broad AI adoption and focused AI investment are different strategies Employees should experiment with copilots, agents, and AI-enabled workflows across functions. But experimentation and strategic capital allocation are not the same thing. → Encourage AI usage broadly across the organization → Concentrate AI investment where differentiation and financial impact are greatest 7/ AI strategy should be business strategy Every strategic AI investment should answer: → Which enterprise priority does this support? → Which financial lever will it move? → Who owns the outcome? → What are we choosing not to fund? If your AI roadmap can be removed from your business strategy without changing either one, they were never actually connected. Save for reference.

The uncomfortable math of AI transformation: 10% is the model. 90% is work your organization hasn't budgeted for. And that 90% is where almost every AI strategy quietly dies. It almost always starts ambitious at the top: sized opportunities, an approved budget, transformation targets, the competitive-advantage language everyone nods at in the boardroom. Then it has to fall through the reality gap into the actual work. Siloed and poor data. Legacy systems that won't expose what you need. Compliance and risk. Change resistance. And the silent killer: no clear owner. The strategy doesn't fail loudly. It quietly stops being anyone's job. This is the gap nobody budgets for, the distance between ambition and execution. The model was never the constraint. The operating model was. What survives the fall has a spine the boardroom version skips: strategy, then operating model, then workflow redesign, then integration and data, then adoption and enablement, and only then business value. Each step is a handoff where strategy can die, and the two that get cut first are workflow redesign and ownership. Which is exactly why pilots inflate and nothing has moved six weeks later. As far as I can tell, AI success is not a technology challenge. It is an execution problem, and the execution side does not improve on its own. PS: Every week in Human in the Loop, I break down where AI actually creates value and where it just creates demos: https://lnkd.in/dbf74Y9E

AI bill shock is coming. Even if it hasn't hit you yet. I was quoted in Paul Smith's AFR piece this week on corporate AI costs. Uber burned its annual AI budget in three months. Westpac's CEO has a token tracker on his home screen. I've spoken to a lot of executives about this. For some organisations, the AI bill is already a problem. For others, it isn't - yet. Either way, it's coming. Better to prepare now than react later. Here's how to get ahead of it: 1. Know your agents. Chat is not what blows the budget. Agents consume many multiples of the tokens a chatbot does. A recent Gartner report suggested 30x. In my experience, far more than that. So "we've built 1,000 agents" is not a goal. Know what your agents are doing. Know whether they're adding value. Quality over quantity. 2. Form an educated view on proprietary vs open-weights models. As I noted in the article, the best-value open-weights models happen to be Chinese, and there's no clear regulatory guidance on whether regulated firms can use them. Don't wait for the regulator. Have a documented position - even if that position is "not yet, and here's why." 3. Invest in AI literacy. Your people need to know when to use which model. You don't need Fable 5 to draft an email. And knowing how to manage a context window can cut token usage dramatically. Matching task complexity to model capability is a trainable skill - and one of the highest-ROI training investments you can make. 4. Get visibility of token usage across the organisation. It doesn't need to be fully automated - industry standards and observability platforms are still maturing. For now, aim for full visibility and a solid estimate. That means individual usage through LLM tools, API consumption, and the AI embedded in SaaS products like Salesforce and Snowflake. 5. Get the architecture right. Good AI tools don't replace good technology architecture. The right design patterns and frameworks materially reduce token spend - model routing is the one getting deserved attention right now. The bill is coming either way. It's important to be ready for it. Full article linked in comments.

AI spend is now one of the fastest-growing line items at every company, but here's the problem: it's invisible to traditional spend management. We experienced this firsthand at Brex, so we built Magpie: an internal tool that sources and normalizes AI spend across every vendor so we can break it down by model, customer, employee, and workflow. What Magpie has already unlocked: → ~85% cost reduction on an audit agent through prompt caching → Real customer cost-to-serve numbers that inform pricing and margin → Model routing governance we couldn't enforce before → A surprise: operational AI is much cheaper than we assumed So why is AI spend so hard to track? Because each one exposes usage data differently (some not at all). At Brex, tool-use costs weren't tracked alongside tokens, customer attribution was inconsistent, and a meaningful chunk of our daily AI spend was unattributable. Now with Magpie, we have reliable visibility. AI spend is more complex than anything finance teams have had to manage before, and it's exactly the kind of problem Brex is built to solve, shaping how we think about the future of our platform. More in the comments.

Financial services spend the most on AI and extract the least value from it. Financial services firms lead global AI spending, yet adoption remains low because operating models have not caught up with technical capability. Capital is being spent on models while workflows are still designed for manual reviews and slow approvals because governance has not evolved. The risk is not unused software but delayed decisions that slow revenue and increase compliance cost. Ignoring this keeps institutions operating at higher latency. AI systems now generate real-time signals in areas like fraud detection and customer targeting because data access and computing power have improved. Organizations struggle to act on these signals because approval structures and trust models were built for periodic reports and not continuous decisions. This creates a gap where insight exists but execution stalls. This means: value erodes before it reaches the customer or the balance sheet. AI adoption doesn’t succeed when added to unchanged workflows because people become the constraint instead of the technology. One practical way to begin is to choose a decision that currently takes days, redesign the approval path to work in minutes and avoid using AI where accountability cannot be clearly assigned. #AIInBanking #FinancialServices #AIAdoption #EnterpriseAI #OperatingModel #WorkflowDesign #DecisionLatency #AIGovernance #BusinessOutcomes #DigitalOperations #AIExecution #Leadership

The blowback from AI overspending and underdelivering will fall primarily on the product organization. SAP’s recent reorg is proof of how fast dramatic changes happen when AI products underperform. Its customers see SAP’s AI products as adding high cost, but minimal value. And it's far from the only one. Salesforce, ServiceNow, Microsoft, and many others face the same challenges. If you’re in a data or AI product manager’s role, here’s what you need to do before the end of July, even if you don't work at an enterprise tech company. 1. Drop the vanity metrics and attribute the value. If you can't trace savings or revenue back to a specific product decision, you have a demo, not a business case. Build the attribution model now, before someone asks for it in a room you're not in. 2. Audit every pilot and either ship it or end it. Pilot purgatory is a career liability. Every initiative still sitting in evaluating past 30 days is evidence you can't deliver value. Force a verdict on each one. 3. Re-price against value, not compute. SAP's mistake isn't the data infrastructure or AI. It’s billing customers in units they can't translate into ROI. If your product's pricing is measured in tokens, seats, or consumption instead of the outcome it produces, you're handing someone else the job of proving your product's ROI. That's supposed to be your job. 4. Translate the win into the language of the room you're in. Walk into each conversation with the same slide, and you've lost half the room before you begin. Build the cost, strategy, opportunity, customer value, roadmap, and throughput of your proof point and know which one to lead with to prove ROI in terms that the people you’re talking to care about. Tough Love: Right now, ROI is the only thing capable of standing between product teams and the next reorg headline.

Your AI agent demo cost $3 to run. Production costs $30,000 a month. Leadership thinks the difference is "scale." It isn't. It's nine line items that never show up in anyone's dashboard — and I've watched teams discover each one the expensive way. 𝗧𝗵𝗲 𝗼𝗻𝗲𝘀 𝘁𝗵𝗮𝘁 𝗵𝘂𝗿𝘁 𝗺𝗼𝘀𝘁 𝗖𝗼𝗻𝘁𝗲𝘅𝘁 𝗿𝗲-𝘁𝗿𝗮𝗻𝘀𝗺𝗶𝘀𝘀𝗶𝗼𝗻 Every turn resends the entire history. Turn 20 pays for turns 1–19 all over again. Your costs don't grow with usage — they grow with conversation length, quadratically. 𝗧𝗼𝗼𝗹 𝗿𝗲𝘀𝘂𝗹𝘁𝘀, 𝗿𝗲𝗽𝘂𝗿𝗰𝗵𝗮𝘀𝗲𝗱 That 30K-token API response your agent retrieved on step 3? You buy it again on every step after. Forty steps, forty purchases. 𝗧𝗵𝗲 𝘀𝗶𝗹𝗲𝗻𝘁 𝗹𝗼𝗼𝗽 An agent that needed 4 iterations ran 40, because nobody defined "done." Exit criteria aren't governance bureaucracy — they're a budget control. 𝗦𝘂𝗯𝗮𝗴𝗲𝗻𝘁 𝗺𝘂𝗹𝘁𝗶𝗽𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻 Five subagents means five copies of context, plus an orchestrator narrating between them. Sometimes one agent with good tools is 5x cheaper AND better. 𝗧𝗵𝗲 𝗹𝗲𝘃𝗲𝗿𝘀 𝘁𝗵𝗮𝘁 𝗮𝗰𝘁𝘂𝗮𝗹𝗹𝘆 𝗺𝗼𝘃𝗲 𝗶𝘁 → Prompt caching cuts cached input cost by roughly 90% — the single biggest lever most teams haven't pulled. → Model routing is second: your agent doesn't need a frontier model to format JSON. → Step caps and retry budgets close the tail risk. Full anatomy in the infographic — every line item, every lever. Save it for your next infra review. The uncomfortable part: none of this is a model problem. It's an engineering discipline problem. The teams whose returns are compounding aren't using better models — they're metering better. Where should cost ownership actually sit — with the engineers building agents, or with a platform team that meters everything centrally?

If you’re leading AI initiatives, here is a strategic cheat sheet to move from "𝗰𝗼𝗼𝗹 𝗱𝗲𝗺𝗼" to 𝗲𝗻𝘁𝗲𝗿𝗽𝗿𝗶𝘀𝗲 𝘃𝗮𝗹𝘂𝗲. Think Risk, ROI, and Scalability. This strategy moves you from "𝘄𝗲 𝗵𝗮𝘃𝗲 𝗮 𝗺𝗼𝗱𝗲𝗹" to "𝘄𝗲 𝗵𝗮𝘃𝗲 𝗮 𝗯𝘂𝘀𝗶𝗻𝗲𝘀𝘀 𝗮𝘀𝘀𝗲𝘁." 𝟭. 𝗧𝗵𝗲 "𝗪𝗵𝘆" 𝗚𝗮𝘁𝗲 (𝗣𝗿𝗲-𝗣𝗼𝗖) • Don’t build just because you can. Define the Business Problem first • Success: Is the potential value > 10x the estimated cost? • Decision: If the problem can be solved with Regex or SQL, kill the AI project now. 𝟮. 𝗧𝗵𝗲 𝗣𝗿𝗼𝗼𝗳 𝗼𝗳 𝗖𝗼𝗻𝗰𝗲𝗽𝘁 (𝗣𝗼𝗖) • Goal: Prove feasibility, not scalability. • Timebox: 4–6 weeks max. • Team: 1-2 AI Engineers + 1 Domain Expert (Data Scientist alone is not enough). • Metric: Technical feasibility (e.g., "Can the model actually predict X with >80% accuracy on historical data?") 𝟯. 𝗧𝗵𝗲 "𝗠𝗩𝗣" 𝗧𝗿𝗮𝗻𝘀𝗶𝘁𝗶𝗼𝗻 (𝗧𝗵𝗲 𝗩𝗮𝗹𝗹𝗲𝘆 𝗼𝗳 𝗗𝗲𝗮𝘁𝗵) • Shift from "Notebook" to "System." • Infrastructure: Move off local GPUs to a dev cloud environment. Containerize. • Data Pipeline: Replace manual CSV dumps with automated data ingestion. • Decision: Does the model work on new, unseen data? If accuracy drops >10%, halt and investigate "Data Drift." 𝟰. 𝗥𝗶𝘀𝗸 & 𝗚𝗼𝘃𝗲𝗿𝗻𝗮𝗻𝗰𝗲 (𝗧𝗵𝗲 "𝗟𝗮𝘄𝘆𝗲𝗿" 𝗣𝗵𝗮𝘀𝗲) • Compliance is not an afterthought. • Guardrails: Implement checks to prevent hallucination or toxic output (e.g., NeMo Guardrails, Guidance). • Risk Decision: What is the cost of a wrong answer? If high (e.g., medical advice), keep a "Human-in-the-Loop." 𝟱. 𝗣𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻 𝗔𝗿𝗰𝗵𝗶𝘁𝗲𝗰𝘁𝘂𝗿𝗲 • Scalability & Latency: Users won’t wait 10 seconds for a token. • Serving: Use optimized inference engines (vLLM, TGI, Triton) • Cost Control: Implement token limits and caching. "Pay-as-you-go" can bankrupt you overnight if an API loop goes rogue. 𝟲. 𝗘𝘃𝗮𝗹𝘂𝗮𝘁𝗶𝗼𝗻 • Automated Eval: Use "LLM-as-a-Judge" to score outputs against a golden dataset. • Feedback Loops: Build a mechanism for users to Thumbs Up/Down outcomes. Gold for fine-tuning later. 𝟳. 𝗢𝗽𝗲𝗿𝗮𝘁𝗶𝗼𝗻𝘀 (𝗟𝗟𝗠𝗢𝗽𝘀) • Day 2 is harder than Day 1. • Observability: Trace chains and monitor latency/cost per request (LangSmith, Arize). • Retraining: Models rot. Define when to retrain (e.g., "When accuracy drops below 85%" or "Monthly"). 𝗧𝗲𝗮𝗺 𝗘𝘃𝗼𝗹𝘂𝘁𝗶𝗼𝗻 • PoC Phase: AI Engineer + Subject Matter Expert. • MVP Phase: + Data Engineer + Backend Engineer. • Production Phase: + MLOps Engineer + Product Manager + Legal/Compliance. 𝗛𝗼𝘄 𝘁𝗼 𝗺𝗮𝗻𝗮𝗴𝗲 𝗔𝗜 𝗣𝗿𝗼𝗷𝗲𝗰𝘁𝘀 (𝗺𝘆 𝗮𝗱𝘃𝗶𝗰𝗲): → Treat AI as a Product, not a Research Project. → Fail fast: A failed PoC cost $10k; a failed Production rollout costs $1M+. → Cost Modeling: Estimate inference costs at peak scale before you write a line of production code. What decision gates do you use in your AI roadmap? Follow Priyanka for more cloud and AI tips and tools #ai #aiforbusiness #aileadership