DataInbox
    NL
    DataInbox as a teammate

    One teammate that keeps everyone up to date.

    DataInbox keeps current sources, rules, answers and evidence in one shared business information layer, so people and agents do not have to research the same question again.
    • The same current context for every teammate
    • Source, freshness and answer stay connected
    • AI models can change without losing business knowledge

    Example for a team of 50 colleagues

    €32,492

    addressable net value per year

    1,003

    repeat questions / month

    77%

    calculated overlap

    This example shows what less repeated work could mean. It is not a forecast or guaranteed saving. Compare team sizes below and explore the assumptions.

    The simple calculation

    How often does your team ask the same question?

    More colleagues and more questions increase the chance that the same information need returns. Fewer recurring question types create more overlap.

    Choose a team scenario

    Choose a team size to compare the possible impact of repeated questions. Every example uses the same fixed assumptions.

    What the projection shows

    1,299 questions per month

    Likely unique296
    Likely repeated1,003
    Addressable with DataInbox65%
    Value of recovered time€2,988 / month
    Avoided AI processing (minor)€15 / month
    DataInbox scenario- €295 / month

    Net scenario from recovered work

    €2,708 / month

    Recovered work already covers the DataInbox scenario before any tool consolidation.

    View assumptions and the DataInbox model

    The Inbox configuration uses an occupancy model: expected unique questions = question types × (1 - ((question types - 1) / question types)all questions). Repeats are all questions minus expected unique questions.

    This example assumes 65% of repeated questions are addressable, an hourly value of €55, five minutes per repeat and €295 per month. For smaller teams it shows how much existing webchat, automation, forms, knowledge or connector spend would need to be retired to break even. No licence saving is included automatically.

    About this example: calculated with the DataInbox Runtime on 24 Aug 2026. The team result uses fixed assumptions; model prices are checked separately against their own source and exchange-rate dates.

    The recommended strategy

    Make DataInbox part of the team.

    This strategy follows the documented DataInbox capabilities: receive sources and events, validate data, resolve current context, govern agents and write results back as traceable messages.

    Important distinction: DataInbox is the shared information and governance layer. It does not automatically replace every source system, team account or human decision.

    1. 1. Listens

      Receives changes from systems, sources, APIs and people.

    2. 2. Checks

      Validates structure, source, permissions, freshness and applicable rules.

    3. 3. Keeps everyone current

      Gives people and agents the same permitted, current business context.

    4. 4. Learns operationally

      Writes questions, decisions, results and exceptions back as traceable events.

    Content Inbox Expert in the Box CLI and runtime

    Architecture, not token tricks

    The cheapest AI operation is the one that doesn't need AI.

    See why events, state, rules and permissions should be processed deterministically, while LLMs are reserved for work that genuinely requires intelligence.

    Read the Machine Age Fund analysis
    Provider prices checked · 9 Oct 2026ECB rate date · 8 Oct 2026

    The same questions, changing model prices

    What have 1,000 questions cost over time?

    Compare the published cost of the same workload: 10,000 input tokens and 500 output tokens per question. See how model releases and price changes affect model costs.

    History checked · 9 Oct 2026
    OpenAIAnthropicGoogle
    Published model costs for one thousand questions over timeOpenAI, Anthropic and Google. Each point is a recorded release or price event, using ten thousand input tokens and five hundred output tokens per question. The price axis is logarithmic. Amounts and sources are also available in the event selector below the chart.€5€10€25€50€100€3502023202420252026

    Cost per 1,000 questions. The price axis is logarithmic: equal steps show equal ratios, keeping lower costs readable. Points are actual release or price events; lines connect different models and are not daily measurements.

    Source event: 7 Oct 2026ECB rate date: 7 Oct 2026
    Official provider sourceECB exchange-rate source

    Claude Sonnet 5.5 · per 1,000 questions

    €22.37

    Standard price per million tokens: $2 input · $0.1 cache read · $10 output.

    Only cache reads became cheaper. This historical workload uses no cache, so its USD cost is unchanged. The EUR view uses this event's exchange rate.

    €1 = $1.1177. USD cost divided by this rate = EUR cost.

    OpenAI

    First recorded

    €307.35

    GPT-4
    14 Mar 2023

    Latest recorded

    €22.02

    GPT-6.1 Sol
    29 Sept 2026

    Anthropic

    First recorded

    €83.72

    Claude 2
    11 Jul 2023

    Latest recorded

    €22.37

    Claude Sonnet 5.5
    7 Oct 2026

    Google

    First recorded

    €37.08

    Gemini 1.5 Pro
    23 May 2024

    Latest recorded

    €8.10

    Gemini 3.8 Flash
    2 Sept 2026

    Locally stored price history from 14 Mar 2023 to 7 Oct 2026, checked on 9 Oct 2026. EUR amounts use the ECB rate for each event, not today's rate. The calculation uses uncached input, standard processing and no batch discount. Equal token counts do not prove equal model quality. Tools, grounding, storage, regional processing and taxes are excluded.

    Combine answers, workflows and agent tasks.

    Choose how much of each you expect and see estimated model costs. Business value requires a pilot with real outcomes, quality and total costs.

    Explore my usage mix

    The team result above is an example with fixed assumptions; the usage mix shows estimated provider costs.